As an anonymous user, you can only add new data. If you would like to also modify existing data, please create an account and indicate your languages on your user page.

International Beer Parlour/Archive20150131

From OmegaWiki
Jump to: navigation, search

Nuvola apps chat.png
Start a new discussion

[Here are archived discussion inactive from 2014-10-31 to 2015-01-31 (archiving date).]

Discussion pages
Strategic discussions
International Beer Parlour

- archives

List of Beer Parlours

International Linguists Beer Parlour

- Inflexions

Questions about words
Articles needing attention
Insect room (report bugs)

- archives

Functionality wanted

- archives


Colognian wordlist downloads?[edit]

Can we have lists of Colognian (ksh) words and definitions on the download page? Thanks. --19:13, 5 September 2013 (CEST)

Done. Kindly inform me if you encounter any problem with the word and definition list. Thanks -- 向榮 /Hiong3-eng5/ (talk) 07:04, 6 September 2013 (CEST)

Import of data[edit]

Hi, from an e-mail exchange with Kipcool I understand that uploading data that was edited offline is now possible. Please tell me if I am wrong. From past projects I have quite some data also in languages that by now are not present on OmegaWiki. What would be relevant for me is: how can I get the data in here? I would consider this way:

  • Creating 2 collections (one is our basic data for kids and the other specialistic data for translators/linguists)
  • Then an export of the English defined meanings + translation of the term in language X is needed
  • Completion of the data with my list here
  • Upload of the edited data

For the kid's data, eventually extending the OLPC-Collection could make sense, too Is this the way or how else could we go about it? Further we used to create files for Parley (KDE-Education) from our data. It would begreat to be able to do this in future as well. And an additional use: co-operation with Kevin Scannell on spell checkers for less resourced langauges. Also Apertium has always been in the background for us as a goal and so it will be also in future. Thanks for hints and ways how to get on. --Sabine (talk) 10:52, 19 September 2013 (CEST)

Hi, we currently have a special download page where a list of expressions and definitions can be downloaded and updated. Actually there is a third, which I call OmegaWiki Developer's csv, which I have used but is still has a buggy interface. I am still debugging the code. Once I am satisfied with it, I will submit it. It is a collection of csv for a language that links a set of definitions to syntrans (with annotations). And since it has defined meaning ids, these files can be linked to other language collections. I have used these csv to create OmegaWiki bilingual dictionary for mids CSVs. This can be used for the export part of defined meanings + translations. Once I finish it (I do not have too much time to spare though...) I will inform you. Hopefully since we have at least some ways of importing and exporting our data, OmegaWiki will become revive its activities for the benefit of those who need our data. --向榮 /Hiong3-eng5/ (talk) 17:01, 19 September 2013 (CEST)
It's good to know that. Tnak your for the info!
For now I am preparing the English expressions on subpages so that I can see what is already online and to create Collections. It's too much to explain now, but since my last activity on OmegaWiki quite some years passed.
I collected quite some terminology in databases and simple tables. Much of the terminology was used for KDE-Education (Parley) and I would like to have the possibility to achieve the same again with updated data and new data, of course. Other terminology has been collected for a project of the University of Bamberg, Germany and also that should go online and people should be able to re-use the data for different purposes. I am not a programmer though I do have some understanding of what is feasible.
It would be good to have this feature ready before International Mother Language Day in February since normally this is the day when we reach out to people and ask them to contribute adding translations to prepared terminology tables. Not having to deal with database export and import (MySQL queries) myself means I can concentrate on creating the contacts which actually is what I can do best: networking :-) --Sabine (talk) 18:06, 19 September 2013 (CEST)
Is there a place where one can read how to format data files to be uploaded later? --Purodha Blissenbach (talk) 15:33, 20 September 2013 (CEST)
Anything that is in the form of an excel table or a csv file is fine. We can then import it with the scripts that we already have.
If possible, there should be an extra column to indicate the DM number, so that we can directly import it to the right DM. But if it is not there, we'll find a way. --Kip (talk) 11:44, 20 September 2013 (CEST)
Can we have two more sections in the download page, duplicating the existing ones but including columns with id numbers in the downloads? --Purodha Blissenbach (talk) 15:40, 20 September 2013 (CEST)
I agree with Purodha, the best way to work with the data offline, prepare collections to be edited by people (like the projects we always did during International Mother Language Day). This means we do need lists starting with UUIDs: UUID - source term - source definition - target term - target definition
This would make it easy for almost anyone to contribute. --Sabine (talk) 17:56, 20 September 2013 (CEST)
We have a way to mass transfer of syntrans using our api. This would be done by creating a wiki page like what I use, you can check out the format here. This can be done but with a bot status that can be requested from Kip etal. If you do not want to use a bot, then you could request me to add them (I have a bot) or you could ask Kip (he has access to the server :) ). Basically the format should have 3 columns (expression/spelling, language_id and defined_meaning_id) for each line. an optional 4th column can also be added for identical meaning id (0 or 1). Remember that each columns are separated using a comma. Then the API can be called. Example
A help page has been set up here, but is incomplete. You could check out the documentation from the API itself. -- 向榮 /Hiong3-eng5/ (talk) 07:02, 21 September 2013 (CEST)
Thank you! This is exactly what I needed to get the tables I already do have here into the right format :) --Sabine (talk) 07:12, 21 September 2013 (CEST)
Thank you! Exactly the same with me. --Purodha Blissenbach (talk) 15:22, 21 September 2013 (CEST)
You're welcome :) -- 向榮 /Hiong3-eng5/ (talk) 17:57, 21 September 2013 (CEST)
identical meaning id (0 or 1) -- What do "unspecified, 0, 1" stand for? Please clarify Help:OmegaWiki_API#ow_add_syntrans - thank you. --Purodha Blissenbach (talk) 00:10, 24 September 2013 (CEST)
I will try to answer this the way I understood the question. Identical meaning id is needed when using ow_add_syntrans action one word at a time. Identical meaning id can be excluded when using wikipage (bulk transfer). But I should have warned you that doing so, identical meaning id is automatically set to 1 (meaning the word is identical). Note that identical meaning is a required element in adding syntrans by Wikilexical software. Hmm, should we also make it the same with individual action? should be automatically equivalent to ? -- 向榮 /Hiong3-eng5/ (talk) 06:49, 24 September 2013 (CEST)
Thank you! I have updated the API help page accordingly, so that you can find the values there, too. Feel free to amend for better understandability and clarity. --Purodha Blissenbach (talk) 07:55, 24 September 2013 (CEST)
Thanks for doing that. -- 向榮 /Hiong3-eng5/ (talk) 07:00, 25 September 2013 (CEST)

Half automated import?[edit]

Do we have a way, currently, to automatically load additional syntrans data into a page and look at it before we hit the "save" button manually (or possibly not hit it)? --Purodha Blissenbach (talk) 07:59, 24 September 2013 (CEST)

we have a semi-automatic import js script for importing translations from wiktionary, which does something like that ( User:Kipcool/common.js/wiktimport.js ) . If you know some javascript, you can probably adapt it to your needs. It basically clicks on buttons and fill the fields like a user would do, but let the user click on save. --Kip (talk) 11:33, 24 September 2013 (CEST)
Sounds interesting. I'll have a closer look at the weekend. At the moment I am preparing an uploadable list. The most tedious part is checking the definitions to really match, respectively, selecting which definition is right. --Purodha Blissenbach (talk) 16:46, 24 September 2013 (CEST)
Well as to upload of Data, I believe that we should consider creating terminology sets where people can work and once we have these start to get them done in various languages. I could imagine for example to say "export <UUID> <English source> <English definition> where <collection> is TRUE and the translation for <Language> is NULL" - in that way we can ask fairly easily for help. People are often freightened when having to use a UI, but when working in a spreadsheet the hurdle for many is much lower. --Sabine (talk) 17:15, 24 September 2013 (CEST)
Sabine, correct me if I am wrong. You want a list as you have described above which contains possible translation which lines can be deleted when wrong. Then convert the data in an uploadable format? 向榮 /Hiong3-eng5/ (talk) 22:13, 24 September 2013 (CEST)
As to export a .csv listing:
OmegaWiki-UUID-sourceterm - source Term in a given language - definition in a given language - target term in a given language
Example with source language English and target German
8320,"glasses","A pair of lenses in a frame that are worn in front of the eyes and are used to correct faulty vision or protect the eyes.","Brille"
This would allow for
  1. proof reading of present entries
  2. addition of translations to existing defined meanings when the target term is empty
Additionally it would be good to be able to export all defined meanings of a specific collection - and there it gets complicated when creating the query, I suppose.
This is an example of work we did for International Mother Language Day 2010 [1] - now I modified one of the various files so that I can add Defined Meaning UUIDs to the list for upload [2] - now it's a matter of creating sets of data where I can ask people to contribute. Consider that I already asked many of those that had been working in the spreadsheet to work online (OmegaWiki, Wiktionary etc.), but hardly anyone helped online, while editing a spreadsheet was easy to get. And converting the translated terms into an uploadable format is then fairly easy considering your example. --Sabine (talk) 07:39, 25 September 2013 (CEST)
Based on your needs, it seems you need to check the list of special pages. The one Expressions needing translation. 向榮 /Hiong3-eng5/ (talk) 13:08, 25 September 2013 (CEST)
This is more or less what I search for - just that I need it as .csv and with ID. But as much as I can see when looking at the source of the page, the needed elements are there, therefore the query is ok
<td class="spelling column-odd"><a href="/DefinedMeaning:blind (6807)">blind</a></td><td class="definition column-even">Unable to see.</td></tr><tr id="expression-6817-129087"><td class="language column-even" langid="85">English</td>
6807 - blind - Unable to see.
Therefore by transmitting the right parameters it should be possible to get it into a .csv
And by amending this query accordingly it should not be too difficult to get the terms "with translation" - and these are then needed to get exports for example for Dictionaries for Mids or Parley or glossaries for CAT-Tools etc.
This means that we need to use the "Collections" in a proper way and this needs quite some thought. I also had a look at the classes today, maybe a second approach is to translate the class "animals" instead of the collection. But well, one step at a time. Right now there are too many informations for me and I have to first get a plan on the way to go to involve people. --Sabine (talk) 15:05, 25 September 2013 (CEST)

Export of Data[edit]

I was just looking at Special:Ow data searchand got

<tr id="words-341452"><td class="short-text column-even"><a href="">Gschäfd</a></td><td class="language column-odd" langid="568">East Franconian</td><td class="column-even"><a href="">store</a></td><td class="definition column-odd"><span title="An establishment, either physical or virtual, that sells goods or services to the public.">An establishment, either physical or virtual, that sells goods or services ...</span></td></tr>

when saving as html. Such an export can be manipulated with bash scrips. But being this a MySQL query it should be possible to have it stored in CSV as well. When looking at this ... hmmm ... I need to have a look at the OmegaWiki database structure. Uhmmm .. didn't I say I don't want to work with DBs and just care about getting people to help??? --Sabine (talk) 18:16, 20 September 2013 (CEST)

Hmm, this page is not adapted for downloads. It shows only 100 results at a time, and the way to generate a page with 100 results and all results is different, which is why we have a separate page Special:Ow_downloads, but it does not have the same functionalities.
So, you would like a "download this list as CSV" button, even if it gets you only the 100 results shown? (this part is easily doable) --Kip (talk) 19:36, 21 September 2013 (CEST)
Well no - I'll explain this better next week. I am considering to install MySQL and phpmyadmin here on the windows machine - not sure if I have enough time though ... creating queries always takes some time for me, because I don't do it regularly and often have to look up how to reach what I want to reach. --Sabine (talk) 19:53, 21 September 2013 (CEST)

List of already present languages[edit]

I know there was a list somewhere ... I just cannot find it. Can anyone give me the link where I can find the list of langauges that is already present? Thanks :-) --Sabine (talk) 15:07, 21 September 2013 (CEST)

The Help:Index language section might help if updated. But if I were you, I'd use the special Statistics page instead. -- 向榮 /Hiong3-eng5/ (talk) 16:39, 21 September 2013 (CEST)
Thank you! Please be patient, I haven't been around for quite some time and often don't remember things or do not know new features. I cannot find Jola ISO 639-3 dyo since I do have a wordlist a friend of mine edited. Have to look a bit more during the coming week. --Sabine (talk) 17:02, 21 September 2013 (CEST)
Actually, I am happy that you have returned to OmegaWiki, you being a co-founder of OmegaWiki :) Welcome back! There's a Jola-Fonyi language available with one syntrans. I currently do not have access to my local OmegaWiki copy, so I can not tell you what its language id number. In case you need to add a language, Kip can help you, that is, if you are not able to do it yourself ( being a Bureaucrat here :) )
Special:ImportLangNames or Special:Languages might be useful when you need to add one. I am not sure, I do not have the rights to do so. Again, Welcome back! -- 向榮 /Hiong3-eng5/ (talk) 17:54, 21 September 2013 (CEST)
Thank you! Who knows why: I couldn't find it searching for Jola via the browser - or: I wrote it wrong and didn't note. It is indeed Jola-Fonyi and yes, I can add languages and know how to do this. Once I am up to date (after reading the blog etc.) I'll see to get some more people to help us. Often it is easier to have people working in tables. Let's see ... my translation is waiting over here :-) --Sabine (talk) 18:28, 21 September 2013 (CEST)
the language id of Jola-Fonyi is 401. 向榮 /Hiong3-eng5/ (talk) 02:50, 22 September 2013 (CEST)
Thank you! I suppose we will also talk about DfM, soon. Gert told me that you were thinking about some project there. --Sabine (talk) 16:05, 22 September 2013 (CEST)
good - I will find my way through. Btw. should I blog my thoughts on the OmegaWiki-Blog? Just wondering ... --Sabine (talk) 19:51, 24 September 2013 (CEST)
Yes, please blog :) --Kip (talk) 10:09, 25 September 2013 (CEST)
Yes, please do. Blogging always includes the chance to attract interested people, if nothing else. :-) --10:35, 25 September 2013 (CEST)

Next stupid question: pronuciation (sound)[edit]

We link to commons for pictures - do we also for pronunciation? If not: could this be handled? It just came to mind ... so why not ask? --Sabine (talk) 19:50, 24 September 2013 (CEST)

I think that is in a wish list somewhere :-) how to incorporate it to the view pages is not yet implemented though :-( We don't have any programmers with lots of time to implement this. 向榮 /Hiong3-eng5/ (talk) 22:21, 24 September 2013 (CEST)
Ok, thanks. This question was just to understand where we are re. "sound". It's not the most relevant part for what we can do, because most people won't help easily with recording.
I for one would not mind having hyperlinks to sound recording as annotations in basically the same way as links to images, even if they do no show in 'view' pages at all, or only display a klickable link there.
We do have collections of tens of thousands of sound recordings for some languages at Commons, e.g. for Polish, English, and Dutch, so we could attempt automated or semiautomated imports once we have a way to store hyperlinks to recordings.
Also, in the vocabulary trainers of Shtooka, there are large collections of sound recordings of words of several languages which are, so I believe, usable for OmegaWiki at least licensewise. I do not know about linking to them and online storage for them, however. --Purodha Blissenbach (talk) 10:30, 25 September 2013 (CEST)
No, wait, we do have sound file annotations. See Expression:dictionary for an example with both US and UK sound files.
We just miss the function to hear the sound directly (should be done with javascript, like for the images). If you click it takes you to commons. --Kip (talk) 10:12, 25 September 2013 (CEST)
Well having the link is enough to associate things. Let me see if I can get my son working on such tings ... he's old enough and bored by ordinary computer games. Another thing I find very relevant is that we can also link to - often the pictures there are more fun for kids and therefore could be associated mainly to OLPC-dictionary (well I'd call it Children's Dictionary). --Sabine (talk) 10:51, 25 September 2013 (CEST)

Having links to sound recordings on Commons on our view pages, I would like to:

  1. Show the URLs of the description pages in Wikimedia Commons, linking to them.
  2. Show the associated descriptions fields from the annotations in OmegaWiki.
  3. Provide a button dowmloading the sound file without playing it in the browser.
  4. Provide a button for immediate play, subject to browser support.

For the latter two, different URLs from the stored ones are needed (ścislaŭ.ogg versus e.g.) i.e. different servers and differing paths. This is a bit tedious, even though they cen be computed from one another. The conversion algorithm depends on wiki settings and might be altered over time. If it not already does so, the MediaWiki API can be extended to provide the conversion.

Showing the URL might be less desired. Then the Link should be labelled something like "Sound file meta data". Since downloads are also possible from the Commons description page, download links may be deemed superfluous. Most user might not know that there is an option for dowloading in the description page. Thus, in this case, the link should be labelled "Download and Metadata page" or similar.


  • Why do we store the part http[s]:// of the URLs, if we label the field Pronunciation file from "Commons"? It is constant, and it is in the way, when the server name or paths happen to be altered in the future.
  • We can enter arbitrary URLs in the field. Why not change the label and allow links to other sites?
    • If so, we would preferrably want to have two links or at least two URLs; one to the file, one to a meta data page about the file.

Greetings --Purodha Blissenbach (talk) 17:28, 28 September 2013 (CEST)

External OmegaWiki software[edit]

I am planning to set up a sourceforge project for OmegaWiki. But am not sure how to call it. Can anyone suggest a name for the project? I plan that this would includes documents, CSVs, scripts, database, whatever that can help OmegaWiki's scope of audience wider.

I originally plan to use Owlet (a small OWL, as in OmegaWikiLexical), but there are two project with that name.

So maybe we could use...

  • OmegaWikiLexical External Toolbox. (O.W.L.E.T)
  • OmegaWiki Z. (a play on OmegaWiki and Wiktionary Z, Z to point out that we are reaching out for the end users or something)
  • OmegaWiki. But this might be confused with the main OmegaWiki Project.

What do you think? -- 向榮 /Hiong3-eng5/ (talk) 10:48, 25 September 2013 (CEST)

Your first name is fine, I'd say. Acronyms are easy to remember. Just to give an additional example from another project: for OmegaT time ago someone used simply "Omegat+" or "OmegaT Plus" to indicated that things are addidional to the CAT-Tool. It could be a way, so over time you can include various things. --Sabine (talk) 10:55, 25 September 2013 (CEST)
Nice idea :)
Also note the existence of jowkl which is made by people I don't know. --Kip (talk) 18:54, 25 September 2013 (CEST)


Check out DictionaryForMIDs! I have submitted some bilingual and monolingual .csv to them and they liked it. One of our goal is that our data be useful and this is one way to do it. So when they uploaded our OmegaWiki DfMs, OmegaWiki will reach their users (me including), anyone who have java jar enabled or Android mobile phones. Actually they even have a PC version. How cool is that!

They also requested multilingual .csv and I will work on it. Hope this encourage our tribe to improve our project.  :) -- 向榮 /Hiong3-eng5/ (talk) 06:47, 26 September 2013 (CEST)

Awesome! Thanks for your work on that :) It's something I have been wanting to do for a long time, but you did it in a short time ;-) --Kip (talk) 08:41, 26 September 2013 (CEST)
Kipcool ... I tried to save at the same time :-D
Yes, I do know the project since my aim with any dictionary project always was "reusability". This was a mayor hurdle here in the past.
Projects where our data can be used:
These are the usages I mainly foresaw in the past and that now seem to be much nearer than before. From my data I started to build files for Parley and other applications using .sh scripts and python scripts (be aware that I actually do not know how to code, I had basics in programming and database management many, many years ago and therefore having one working script and adapting it is not too difficult for me, while writing myself is almost impossible for several reasons: then again, I am better to work in communications and marketing than doing the ground work, but it is always good to know exactly what I am talking about.--Sabine (talk) 08:50, 26 September 2013 (CEST)

Another way for our data can be used is in those websites who uses other free online dictionaries and show definitions from these various dictionaries in one page. Forgive me, I do not know how to call these sites. I do not know which one are these, maybe if they see our API, maybe they can add OmegaWiki on their list. Do you know any? -- 向榮 /Hiong3-eng5/ (talk) 09:13, 26 September 2013 (CEST)

There is a number of websites arelady using OmegaWiki with most of them not giving attribution. I know that universities do extract data from here. Mostly bilingual tables are extracted, without definitions and then mixed with their existing data. Therefore it is impossible to really understand where stuff is used. The only way to understand this is doing like the big dictionary producers do: they have a number of errors in their dictionaries and these are not there by chance, but to proof somebody is copying those dictionaries. We never did care about this here. Of course it would be good to co-operate and step by step we will very likely get to them.--Sabine (talk) 09:56, 26 September 2013 (CEST)

More Android Apps:

First of all thanks to 向榮 /Hiong3-eng5/ for the list that needs translation in German. Considering DictsForMids, I learnt the following: not all data we have is good for the dictionary, therefore we should consider to create a collection where we add the terminology that should go into DfM. Probably all that is in the Children's Dictionary and in Destinazione Italia should go in there. --Sabine (talk) 17:07, 3 October 2013 (CEST)
I created a Collection called "General Dictionary" which should be the one for DfM. Now it's a question of how to go about to add terminology. Probably starting with a lemma list by topic makes sense, because things we deal with each day are those that need to go in there first. By using topics we can also go for "translation projects" where people can help to build corpus. --Sabine (talk) 10:57, 8 October 2013 (CEST)

Adding very specific terminology ... a problem from the first days on ...[edit]

When I work on translations, normally I always have a group of terms OmegaWiki does not have. Now consider us translators having to work "bulk" with words. The old and still actual problem is: even if I would like to add these glossary parts, it takes a lot of time to write the definition. Very often, when dealing with screws, bolts, nuts etc. it is simply too much. Normally I would add the term, no definition, translation and if time allows a link to a picture. Time ago, and eventually you came through such entries, I added some with [definition needed] in the definition field. At that time pictures were not possible. Now I could imagine a picture or a technical drawing to define "any expression". At least that would be much faster. It's one of those thoughts that keep me thinking and grumbling :-) --Sabine (talk) 10:07, 26 September 2013 (CEST)

See also this talk page contribution. --Purodha Blissenbach (talk) 17:41, 29 September 2013 (CEST)
I noted that there is already quite some terminology that does not have really good definitions, but where by the use of the pictures one can understand what exactly is meant. Therefore I'd go exactly this way. Of course: we can always add good and long definitions, but IMHO in quite some cases it's not a must.--Sabine (talk) 10:59, 8 October 2013 (CEST)


Since we already have an annotation area. Should we not move the British and American Expressions to reflect the area and then delete the Brit and Am expressions? When no expressions are left, lets remove these languages?

These languages are spelling variants not scriptal ones. And besides, no one seems to be adding words on these.

What do you think? 向榮 /Hiong3-eng5/ (talk) 03:57, 29 September 2013 (CEST)

I'd say this makes perfect sense and it is what we need for many less resourced languages as well: my own language, East Franconian, has many variations and therefore such annotations are needed. --Sabine (talk) 17:43, 3 October 2013 (CEST)
That was also my plan when I created the area annotation. Having separate languages for English, English(British) and English(American) only creates additional problems and complexities for the software. It is also an inexact separation, because Australia might use English(British) but sometimes has specific words.
Hiong3.eng5 : do you think the change can be done with the API? --Kip (talk) 18:44, 3 October 2013 (CEST)

Unspecific or untopical words[edit]

Languages have words that are very frequent. Mostly these words have either very broad meanings or are structural or function words having hardly any meaning of their own. Assigning a topic to those words does not really make sense most of the time unless we choose a topic such "none" or "no topic" as such. Currently, this appears ike a good idea to me. --Purodha Blissenbach (talk) 14:54, 8 October 2013 (CEST)

What about "general"? So at least we have such terminology on a list and it can be translated (whenever possible, because some simply cannot be translated - like our Franconian "fei":-)

How to add a topic to the DM?[edit]

I was searching, but cannot find it. On the help page only the various topics are listed. Thanks :-) --Sabine (talk) 17:00, 8 October 2013 (CEST)

I added a screenshot in the page Help:Topic
"topic" appears as "subject" with the English interface, which is maybe why you didn't find it. --Kip (talk) 19:24, 8 October 2013 (CEST)
ok, thank you! I saw that and I find it. Probably some more will be needed even if "countries" for example are going to be redundant. But: if we go for an export by subject to have data edited by people offline, like I did time ago with Destinazione Italia, then we need them in one structure or two different queries .... hmmmmmmmmmm --Sabine (talk) 19:45, 8 October 2013 (CEST)

There are 3 major ways in grouping words, by collection, class and topic. Countries seems to be under class. 向榮 /Hiong3-eng5/ (talk) 23:38, 8 October 2013 (CEST)

Yes, I know, there are classes also for animals, but - and this does not match with what is on top of the Topic help page - let's take a sparrow - a sparrow is a bird and is a wild animal, while a hen is a bird and a domestic animal. When it comes to topics: one word can be part of more topics. Classes do have a "tree" that is: animal, bird, wild bird - animal, bird, domestic bird - that tree application is irrelevant for people learning a language: they want lists of "all birds", "all domestic animals" (including hen, cow, dog etc.), all whatever. Therefore classes are for classification while topics (subjects) are for bulding groups and collections are for building specific corpusses.
This is indeed one of the OW problems - flexibility does not seem to be there, but IMHO this is not true: flexibility is there, because subjects/topics can solve many issues.
People will not be able to define if they have to search for a class or a topic - us who add them do know (or do not know), but imagine someone just wanting to use our data: it's not self explaining enough IMHO. I would make a distinction like:
  • classes = scientific classification of something
  • topics (subjects) = grouping of terminology for "human beings"
  • collections = groups of terminology of specific origin or for specific scopes: OPLC dictionary, GEMET, Destinazione Italia etc.
--Sabine (talk) 08:12, 9 October 2013 (CEST)
We also have a "hypernym" relationship which could be used to build trees, like it is done at WordNet...
I agree that classes / hypernyms / topics / collections are somewhat overlapping. They each correspond to a different functionality of the software, but for a user who don't especially care about the mysteries of the software, the distinction is not so clear.
"topic" was initially meant to indicate the domain to which a words apply at a very general level. So, not "city", "river", "country", but only a general word "geography". This very general classification was already used by the first encyclopédie of Diderot ;-) We could decide to make a finer classification with topics, if it is something that we want? but then it should not be displayed as a long flat list of possibilities (as it is now).
classes is (initially) only meant for adding class-specific annotation (Class "country" enables you to add an annotation "capital" ; class "animal" enables the annotation "scientific name" - imho a nice feature that is unfortunately missing in Wikidata). Of course, it is also technically possible to add classes that have no class-specific annotations.
So, in my opinion, for a list of "all birds", we could use either topic or classes. We just have to agree on which one, and adapt the help pages accordingly. --Kip (talk) 21:07, 9 October 2013 (CEST)

Semantic annotations only allows links to Commons?[edit]

There are various free projects around, such as where one can go and get public domain cliparts, for example for children's books. Furthermore there are many websites with specific descriptions, for example when it comes to machinery. It would make sense to be able to link to them. I know we are nto a search engine, but being a translator often it drives me crazy until I understand that term A in Italian is term X in German, just because it is hard to find pictures to understand if they are really the same. We could also link to Wiktionary - why not? Logos Dictionary - why not? I mean it's just additional information we give ... nobody will add links to all terms, but having them here and there can be really helpful to many Just another 2 cts ... :-) --Sabine (talk) 19:33, 8 October 2013 (CEST)

It is always a bit problematic to decide when a link is relevant, and when a link is just noise. One possibility is to link to as many resources as possible. The drawback is when a page then contains a huge list of links to other websites, which does not look especially good.
Maybe we can add a generic annotation "link to external resources", and take care of not adding too many of them, but only links that actually add information (as it is done in Wikipedia under the section "external links").
Regarding images, we have only images from Commons, because for each website, a specific code needs to be written to extract its image. I'll see what we can do about openclipart, I also see it as a great resource. --Kip (talk) 21:16, 9 October 2013 (CEST)

Facebook fanpage[edit]

Who is the admin of the fanpage on Facebook? Because we should use it ... --Sabine (talk) 16:03, 9 October 2013 (CEST)

It is me :) I have a script that automatically sends the words of the day there, when there is one.
You probably know - and use - Facebook better than me, so I added you as an admin there. --Kip (talk) 21:36, 9 October 2013 (CEST)

Social Media Marketing[edit]

Like some probably know, Social Media Marketing is part of my work. I am right now trying to understand which channels we do have and how I can use them best to get our contents out

  • Facebook (I already have access there, so I will program a word a day)
  • Google+ (I found there is a page)
  • Twitter
  • ... more?

I'll go ahead listing my hashtags now - cu then :-) --Sabine (talk) 20:26, 10 October 2013 (CEST)

I gave you access to Google+ as well.
The Twitter account is not mine, but belongs to User:Patio. Dunno if he can give you access, just ask him ;) --Kip (talk) 20:49, 10 October 2013 (CEST)
facebook: we will have one message a week for now until the end of 2014 - as soon as I have finished preparing the terminology to upload and it is online, I will go ahead preparing publications, wishing to reach one word a day in English. Probably this can be automated somewhat, I already have an idea on how to use a data export to get it on the way, but relevant things first. For now we are somewhat covered. As to G+ I believe we will create an additional page where we mainly work with hashtags. People will come in over time.
If someone wishes to create a word of the day list for another language: please contact me so that I can tell you which format to use.--Sabine (talk) 13:18, 12 October 2013 (CEST)

Possibility to add noun/verb etc. to East Franconian (Main-Franconian called here)[edit]

I cannot add these myself - could someone (kipcool?) please do that for me? We have the same forms as German. If something is missing, I will let you know. Btw. Main-Franconian is less used than East Franconian here in Germany. --Sabine (talk) 11:55, 12 October 2013 (CEST)

I also added the annotation "region => Nürnberg / Würzburg / Sonneberg" for words that are specific to these regions.
You can edit the annotations in the page DefinedMeaning:Mainfränkisch_(720147) (bureaucrat only) in the section "class attributes" near the bottom of the page. Or I can also edit it for you, no problem.
(In the old times, you might remember that every annotation was in the page "lexical item", but since then I've moved the language-specific annotations to the respective language pages to avoid having a huge "lexical item" page). --Kip (talk) 14:05, 12 October 2013 (CEST)
Ah well, I will learn how to do that ... Thank you! --Sabine (talk) 18:40, 12 October 2013 (CEST)

Two keys to one collection[edit]

The collection ISO 15924 has both a set of 3-digit numeric keys and a set of 4-character alpha key, as you can read e.g. on the website of the UNICODE consortium:

Since the UNICODE consortium is in charge for the administration and assignment of these codes, these page can be viewed as authoritative.

The Problem
We have one collection. Adding another key after one already exists is not possible. OmegaWiki silenty ignores the 2nd key when you attempt to add the second key. Thus we either
  • need to allow any number of keys, or
  • need to have an extra collection for each key type.

Both are somewhat questionable depending on ones standpoint of view. I prefer the 'extra collection' solution, with preferrably polysemic definitiions, such as:

  • Expression - Definition - Collection?
  • ISO 15924 - … - (no)
  • ISO 15924 - …, portion organized via 3-digit numeric keys - (Yes)
  • ISO 15924 - …, portion organized per 4-character alpha keys - (Yes)

--18:14, 12 October 2013 (CEST)


Since currently only 3 of 163 collection members seem to have been entered in OmegaWiki, it might be wise to import the lists linked above. How would that be done? Likely, there are these steps to be repeated for each list entry:

  1. Create a new DM, in one language (unless a DM already exists)
  2. Add syntrans for the other language,
  3. Put the DM into the 1st collectiion using its 3-digit numeric key,
  4. Put the DM into the 2nd collectiion using its 4-character alpha key.

Do we have import tools capable of doing that? --18:14, 12 October 2013 (CEST)

I have created an API function that can add dm to collections recently. It has the ability to batch process a wikipage on csv mode using dmid of word to be added, dm id of the collection and the internal membership id(which àan be blank). You have the option to test it first (no real change on the database, to avoid entering errors). we will inform you when it can be used and if help page is updated. 向榮 /Hiong3-eng5/ (talk) 21:32, 12 October 2013 (CEST)
Kindly check this Help page, Help:OmegaWiki_API#ow_add_to_collection. Thanks -- 向榮 /Hiong3-eng5/ (talk) 23:56, 12 October 2013 (CEST)
Thank you, that sounds good! --Purodha Blissenbach (talk) 03:24, 13 October 2013 (CEST)
This is really relevant to me - if I understand this correctly I just create a separate wiki page, such as subpage of my user page and there I do add one entry per line, separated by commas - and the same is true for other "adds". Have to check also other possibilities, because eventually we can get really relevant data this way (people can edit in tables and we can then upload. This really makes a huge difference. --Sabine (talk) 13:47, 13 October 2013 (CEST)
Yes and no. Yes, the add to collection and add syntrans API will work as you described. Only add to collection have a test mode. Others do not... for now. I plan to add the batch processing and test mode functionalities. Too much to do still, so less time! Other annotations are still missing, my Min Nan stats are still thin. I will try my best to juggle this. Your worksheet approach really is inspiring though! And btw, the wikipage idea came from Kip --向榮 /Hiong3-eng5/ (talk) 14:33, 13 October 2013 (CEST)
Well, one step at a time things will work. It is relevant for me to know such features, that will be there in future, because it helps me to avoid many single edits now which take loads of time and to concentrate on those things that need to be done manually. Thank you for the great work done!--Sabine (talk) 16:59, 13 October 2013 (CEST)

How to amend an incorrect link to Wikipedia?[edit]

DefinedMeaning:Kölsch_(6315) has an incorrect link to a Wikipedia disambiguation page. When I tried to alter it, I was unable to find the annotation where it is stored. --Purodha Blissenbach (talk) 02:58, 13 October 2013 (CEST)

The link to Wikipedia is automatically created from the Wikidata ID. What you need to do is correct the WikidataID (under "Semantic annotation" => "plain text"). --Kip (talk) 16:08, 13 October 2013 (CEST)
This is not going to work - I believe. Wikipedia pages linked to each other via Q-codes are quite often not about the same topic, leave alone the same DM. So, it may be helpful here and there to find Wikipedia pages via WikiData, and of course WikiData Q-Code links are good candidates. But need manual appproval on a per language basis and constant surveillance. Wikipedia pages are moved, or their polysemous meanings reassigned, they are often turned into disambigution pages as coverage grows, or even deleted.--Purodha Blissenbach (talk) 19:46, 13 October 2013 (CEST)

How to make wiki page editing usable again?[edit]

When editing wiki pages, the textareas are now having a small proportional font. I want to go back to a readable monospaced font. how can I do that? I tried many things already which did not work. --Purodha Blissenbach (talk) 03:03, 13 October 2013 (CEST)

Try this in your common.css, it works for me (and refresh cache). --Kip (talk) 16:19, 13 October 2013 (CEST)
textarea {
  font-family: courier,monospace;
  font-size: 200%;
I tried many such things - they do not work. Yes, my textareas come in green (for the sake of confirmation) and apear monospaced for a blink of an eye when first loaded after a browser restart, then get proportionalized while page loading is finishes. I believe in a script doing the 'evil' but since I dislike scripts anyways, I have no idea how to find out what and if so. --Purodha Blissenbach (talk) 19:46, 13 October 2013 (CEST)

Topics and Classes[edit]

Like so often I think about making things easier. Well we do have topics (subjects) that are classes. This is very hard to find for people. Now we do have two possibilities: or we create overlapping topics or we create a comprehensive list including topics and classes, where behind the name of the "subject" we add (topic) or (class) so that everyone knows what to search where. I mean, I already know OW from the past and have my problems to find my way through ... --Sabine (talk) 14:02, 13 October 2013 (CEST)

overlapping topics means giving the same annotation twice (duplicate work: bad ;) )
I'd go for the comprehensive list including topics and classes, and maybe later, in view of that list, reconsider whether topics and classes should stay separate or be unified.
Help:List of topics and classes --Kip (talk) 16:25, 13 October 2013 (CEST)
That looks great - thank you! --Sabine (talk) 17:29, 13 October 2013 (CEST)
I finished the list. I remembered a difference between classes and topics (at least one of the arguments we gave when we created topics):
We have currently for example a class "religion" and a topic "religion". They are not the same.
The religion-topic is (supposed to be) for any word that is somehow related to religion.
The religion-class is only for a word that is a religion.
So "Expression:Buddhism" should belong to both religion-topic and religion-class. However, "Expression:Bible" or "Expression:apostle" belongs to religion-topic but not to class (Buddhism is a religion, Bible is not a religion).
So to sum it up, topic = "list of words related to religion", class = "list of religions". Does this distinction makes sense?
Unfortunately, "topic" was created only recently. Before that, classes were used instead. A certain number of classes should therefore be moved to topic according to the above distinction. However, the guy who convinced us to introduce topics has left and nobody had the energy/time to do the changes. --Kip (talk) 18:54, 13 October 2013 (CEST)
Well, when we get to the stage, that we outsource the "translations part", then here we do have time to care about classes and topics. Our vocabulary list is divided into subjects and it's always a list of things that have to do with that specific topic. Like: a feather is part of the topic birds, but not of the class ... that needs really quite some thought and consideration, but we do need the topics (subjects). --Sabine (talk) 19:09, 13 October 2013 (CEST)
Is not 'member of class X' then almost identical to the 'is hyponyn of' X' relation? --Purodha Blissenbach (talk) 19:46, 13 October 2013 (CEST)
I think we don't need classes indeed. With classes, we would need subclasses, sub-subclasses etc.. For instance, from "living being" to "homo sapiens", there are 37 levels of classes according to Wikispecies. And these classes are useful for translation. For instance, if "The CCTVs show the men going out. Then they walk to the station." is to be translated, the fact that "men" are vertebrates proves they can walk (if no exception occurs, such as "disable" men, "dying" men or "died" men), while CCTVs can't walk, which proves "they" refers to "the men", and has to be translated in French by "ils", not "elles". And for this, hypernyms play the role of narrow class system and are very useful. Another useful thing would be a list of subjects, direct complements, indirect complements, adverbs a verb can accept, using such a cascading system: useless to say a cat can "walk" if you already stated that all vertebrates can "walk" and a cat is a vertebrate. The same way, we (or Apertium) will need a list of adjectives a noun can accept etc.. So please fill the "hypernym" field, and fill it as narrowly as possible. For instance don't choose "vertebrate" as hypernym of "homo sapiens": choose "homo" (talk) 04:19, 14 October 2013 (CEST)
I believe that topics may be quite helpful in machine-guessing word relations for translations in unannotated texts. --Purodha Blissenbach (talk) 19:46, 13 October 2013 (CEST)
Well, topics and classes serve two different scopes. They will always be redundant in a certain way, but with Topics you cannot get the tree structure, while with Classes you do. Depending from what our users expect, one or the other will be perfect for them. So I don't see a reason for deleting either of them, but there is a need to eventually define them properly. We will see this in future. Once we have more people working on translations and adding data there will be time to dig deeper.--Sabine (talk) 18:41, 14 October 2013 (CEST)
Fully agreed. There might be a need to harmonize 'class' and 'topic' and 'is hyp(er)ony of' the sense that when they appear to contradict one another, we should suspect an error, suggest to split definitions, or similar. We shall likely have an automated way to detect that in the future. --21:14, 14 October 2013 (CEST)
The question now is not the redundancy between topics and classes: Sabine explained clearly the difference. The question is the redundancy between hypernym/hyponyms and classes. See just above Purodha Blissenbach's question and my (talk) 22:13, 14 October 2013 (CEST)
I think hyp(er)onyn should be used for very specific relations while class be used in a less strict way. Meaning, class would be somewhere int he middle of topic/subject and hyp(er)onyn. So when a list would not fit the strict standard of hyp(er)onyn, use class. I mean if I want a list of animals regardless of whether they are vertebrates, amphibians, or what species or whatever they are categorized, hyp(er)onyn will not fit that. -- 向榮 /Hiong3-eng5/ (talk) 09:07, 15 October 2013 (CEST)
They will. We just need a small recursive function. Don't waste our time and other users' time by inputing manually the same information several times. This is exactly what Omegawiki is about. (talk) 11:16, 15 October 2013 (CEST)
To wrap that up a bit - the relation 'in hyp(er)onym of' is to be used most restrictively in a sense of 'in hyp(er)onym of and there is nothing in between giving an additional level of hy(er)onymity' - Sound good, but has of course two potential weaknesses: • people may not be aware of this rule, a word they would consider the most direct hyp8er)onym is missing from OmegaWiki or not known to contributors. On the other hand, following hyperonymity chains 'upwards' may be fun as well as a "select more specific word" page for, say, preachers, writers or translaters wantig to enhance their style. :-)
Nevertheless, we shall need to redact hyp(er)onymity relations so as to make them function in the intended way. Of course, that can be semiautomatized, and may be little work after a while. I question it to be a good crowdsourcing trait based on the assumption that many would not understand it when entering data. I am not considering this an argument against, though. --13:34, 15 October 2013 (CEST)

Collections of differences between Classes and Topics (Subjects)[edit]

I just seem to note that Classes seem only to deal with nouns. And it's IMHO the only thing that makes sense. Like Africa is a continent of the World, Germany is a country of Europe, Paris is a city of France etc. Do I see this correctly? --Sabine (talk) 12:28, 15 October 2013 (CEST)

Yep, Smiley.svg and Stras(s)b(o)urg is a city of France and/or a city of Germany, depending on whom you ask, at which time you ask and under which circumstances you ask. --13:34, 15 October 2013 (CEST)
We don't have this problem with hypernyms/hyponyms, which also work with verbs: "to shout" is a hyponym of "to express oneself", for adjectives: "mathematical" is a hypernym of "geometrical", and for adverbs: "mathematically" is a hymernym of "algebraically". (talk) 15:21, 15 October 2013 (CEST)

Classification using existing schemes and thesauri[edit]

There are hundreds of well established and much used thesauri and classification schemes in the world. Very many are available electronically and can be used more or less royalty-free or under very permissive licenses. I sincerely doubt that we should try to reinvent these wheels, with dozens of good reasons. Of course, we should use them when possible (which may not be now) and find ways to import (and reimport) their vocabularies and structures with credits and back references. I know, GEMET turned out to be an import having to be followed by some housekeeping, but hey, we can learn! --Purodha Blissenbach (talk) 13:54, 15 October 2013 (CEST)

Swadesh List: 'left' and 'right'[edit]

We have the substantive meanings of right and left as members of the collection "Swadesh list". I believe that Swadesh meant the adjective or adverbial meanings. Can anyone confirm or disapprove this? --Purodha Blissenbach (talk) 20:36, 13 October 2013 (CEST)

Many of the translations do not refer to substantives in their respective languages. I meanwhile believe that the headline "substantive" in either page is wrong and should be altered. Where does it originate from?
I do not know at the moment. May they be there, because I approached the pages via the English Expressions? I'll see to get rid of them unless I find some better way to go. At least some of the definitions that I understand, do not lend themselves well to an English-substantive meaning, imho. --Purodha Blissenbach (talk) 21:18, 14 October 2013 (CEST)

Finally I came to the conclusion that English the DMs right and left should be annotated as having three word classes, or parts of speech: substantive (=noun), adjective, and adverb. The problem is: Omegawiki does not accept another such annotation once one is already there. This is imho wrong. When defintions are identical, we should not be forced into creating separate meanings for each POS. Opinions? --Purodha Blissenbach (talk) 13:55, 19 October 2013 (CEST)

Substantive, adjectives and adverbs are different concepts and should be in different meanings. One reason is that the translations are different in other languages (German: "Rechte" and "rechts", French: "droite", "à droite").
As to "Omegawiki does not accept another such annotation once one is already there", actually yes it does. You can e.g. indicate that a word is used in several regions (region annotation and pos are of the same type). But please don't do that for POS at least until we reach an agreement.
The main issue here in fact is that the definitions for noun, adverb and adjective should not be the same (cf. English Wiktionary, they have nice distinct definitions).
I agree that Swadesh meant adjective or adverb (I would guess for adverb, but no proof). --Kip (talk) 18:21, 19 October 2013 (CEST)
  • I very sincery doubt that substantive, adjectives and adverbs are different concepts when there is only one word in a language to express them and natives do not differentiate, cannot tell a difference in their own language, and their grammar does not have said concepts either. I remember that an Asian told me that there was no way to distinguish between beauty, beautyful, and beautyfully in whatever language he was speaking about, and he was unable to tell his less educated people why the Engish dictionary had three entries for the same meaning. He resorted to it is for grammatical reasons but even that is hard to grasp: at least beautyfully is derived from beautyful in a complete standardized grammaticalized way, thus even from an English grammaticarians standpoint, it could be viewed as a simple grammatical form of beautiful.
  • In German, with the exception of a very limited list of some 25 exceptions, adverbs and adjectives are identical in their lemmatized form. Adverbial usage can be told from the use as adjective syntactically and grammatically, but as long as we do not have inflected forms, we need only one DefinedMeaning for them in German. Whether or not DMs should be differentiated depending on grammatical forms has, to the best of my knowledge, not been decided upon.
  • Translations differing depending on usage, such as the ones you mention, do not warrant another DM. This has to hold even if in the target language you commonly find these different usage forms lemmatized to different entries in their respective dictionaries.
  • Sure, I tried to add a 2nd POS so as to find out whether that would work, but even before that I was sure that doing so or not would have to be discussed prior to using the possiblity.
  • Well, the issue of how to deal with Swadesh's unknown exact intentions has been solved for the time being by me ;-) adding several DMs of some (few) of the words of his list with one index. That works, and I shall elaborate more on this when asking questions about automatically importing some 200 existing Swadesh lists. Adding multiple DMs when we can assume that Swades did not mean to differentiate between very closely reated DMs is akin to automtically adding synonymes which is technally unavoidable. Also, looking at the Swadesh lists for many languages, you find occasional duplicate entries in some positions. Sometimes they may be synonymous, sometimes you have apparent differentiations at the DM level of comparatively imprecise English meanings, such as we = 'me and others including you' versus we = 'me and others excluding you' and similar. --Purodha Blissenbach (talk) 19:44, 20 October 2013 (CEST)

Special ExportTSV[edit]

For Sabine I (discovered and) resurrected the page Special:ExportTSV.

It allows to export a list of words and translations for a specific collection (and now also class and topic :) ) to a TSV (tab-separated-value) text file, which can be then easily converted to an Excel table.

I think it can be actually pretty useful to anyone, so I've just given access to all users.

You can use it to see what words are in a specific topic (there is no other way yet). You can also edit the table offline with Excel, and add translations, which can then be imported with the corresponding Special:ImportTSV (only accessible to bureaucrats and other users with "import" rights). --Kip (talk) 19:48, 15 October 2013 (CEST)

Language numbers for languages[edit]

How can I find out the language IDs? We do already have some entries in dyo and there I would like to add data I have here :-) So which Language ID should I use?--Sabine (talk) 17:40, 17 October 2013 (CEST)

You can find them one by one by asking for a list words of a language in a collection looking at the URL you are redirected to. Clumsy but working. --Purodha Blissenbach (talk) 22:40, 17 October 2013 (CEST)
Jola-Fonyi (dyo) is 401 -- 向榮 /Hiong3-eng5/ (talk) 09:43, 18 October 2013 (CEST)
The id also appears in the html source. However it would be nice to have all of them readily available on a page - also for me. Any idea where I could do that? An existing page or a new page? --Kip (talk) 10:41, 18 October 2013 (CEST)
Thanks for the code :-)
A separate page in the help section, connected to the API description? Because that's where I would search for it. --Sabine (talk) 11:58, 18 October 2013 (CEST)
I changed the title. I now have here Jola-Kasa [3] - and it doesn't seem to have entries by now. I will add the page adding ??? where the language ID normally goes - so a fast search/replace will do.--Sabine (talk) 12:22, 18 October 2013 (CEST)

Word of the Day publications[edit]

I am preparing the Word of the Day right now for the following languages: English, French, German, Franconian, Italian, Neapolitan. Since I decided to work with collections and classes/topics here I'd like to get efficiency to a maximum and export the tables for all languages needed. Therefore I wish to ask you, which other languages I should add. I have some more in mind, but knowing that some of you do really hard work on other ends of OmegaWiki, I don't feel like asking to do even more. Just let me know if you wish a language to be added and translate these few messages needed for facebook, twitter and google+ (for now). I will export the data by tomorrow evening, so until then I need to know the languages. The messages are needed by the end of the coming week then.--Sabine (talk) 14:05, 18 October 2013 (CEST)

Dear Sabine,
Of course you're free to call it word of the day from now on, but on our home page we speak of expression since what's one word in a language may be more words in another. Dutch and German are languages where we can make long words while English many times needs a few spaces. The word 'expression' describes all methods in my opinion.  Klaas|Z4␟V:  13:09, 19 October 2013 (CEST)
Let us rather stick with 'expression' --Purodha Blissenbach (talk) 16:13, 19 October 2013 (CEST)
Well #wordoftheday or #wotd is commonly used by all bigger dictioanries. I do not link to expressions, but to defined meanings and I will stick to what is used on the market. Nobody would ever search for "Expression of the day". --Sabine (talk) 16:55, 19 October 2013 (CEST)
I agree with Sabine. What we call it internally is more exact, but we have to use the terminology that people would google. --Kip (talk) 18:02, 19 October 2013 (CEST)

You guys are right. Better be less exact and be found by Google c.s. than more exact and stay alone. More chance to get more followers, friends and people in circles.  Klaas|Z4␟V:  10:17, 20 October 2013 (CEST)

DM IDs crossing the 1.5 Million barrier.[edit]

Our DM Numbers are now beyond 1.5 Million.

I could not spot the 1500000th DM. Maybe it does not even exist. It looks like DM IDs are not increased in installments of one. I can only point to DefinedMeanings 1499988, and 1500010 from the RecentChanges. Anyone interested to blog or twitter about it? Is it worth to?--Purodha Blissenbach (talk) 16:12, 19 October 2013 (CEST)

No, the system does not work that way. Kip correct me if I am wrong here, but the numbering system, whether expression id, defined meaning id, syntrans id etc are based on the transaction id, therefore unique. meaning transaction number 1500000 probably is not a DM. Technically, one can use the total number of DM ids to get that, either with or without the removed DM. -- 向榮 /Hiong3-eng5/ (talk) 16:27, 19 October 2013 (CEST)
That explanation is making sense to me. Thus, the figure 1.5 Million transactions is not too impressive, and may not worth being publicized. If main page figures are fairly recent, we have some 46 Thousand DMs of unique concepts. Not bad, but also not too impressive given that we support 471 languages. --Purodha Blissenbach (talk) 16:38, 19 October 2013 (CEST)
I don't think that it is based on the transaction_id (a transaction is an edit, two syntrans can be created with the same transaction), but you are correct that the expression id, defined meaning id, syntrans id etc. are counted together. There is a table in the database, the objects table that will tell you if a given ID is a DM, a syntrans, etc.. It is in this table that the +1 increment on the id is performed.
With a small query, we can see how many IDs of each table we have (cf. User:Kipcool/Number of objects ). --Kip (talk) 17:57, 19 October 2013 (CEST)
Yep, my bad, I did told you Kip will correct me if I am wrong :)
The query produced the total IDs of each table, including those removed from their respective tables. Is there a way to exclude the number of removed ids? -- 向榮 /Hiong3-eng5/ (talk) 20:40, 19 October 2013 (CEST)
Then you have to go to each table and use remove_transaction_id. But you can already have the number of DMs, definitions, expressions and syntranses directly from the statistics page Special:Ow_statistics. --Kip (talk) 12:51, 20 October 2013 (CEST)

Expression: does not appear[edit]

I changed the page title so that now the "Expression:" prefix now does not appear in view mode. So now, in Expression:expression, you will see on top "expression" instead of "Expression:expression".

The main reason is to see if the Google referencing works better that way.

Now I am wondering if maybe we should have the language name there as well, instead of having a second line indicating the language. --Kip (talk) 20:29, 19 October 2013 (CEST)

I had look at the HTML page source that Google sees for expression pages - frankly, it is pretty messy, and buggy from a technical perspecive. Just two quick hints:
  • First of all, the language attributes of the HTML title and h1 should be correct. They're not showing the language code of the expression but the uselang or the users preferred language or (likely, but I did not check that) the language as requested via browser preferences. This needs to be fixed.
  • Then, there is a pink field with rounded left and right edges (at least in my browser and skin) containing "language: language", translated, or partially translated, nothing is properly tagged for its language. It is made up by a HTML span. Change the span to h2 so as to make the language name more prominent for Google. That might need a CSS adaption, too.
Since I am on my way to have a working test installation of OmegaWiki, I shall be technically able to submit proposed code changes, hopefully soon, and I shall do so. --Purodha Blissenbach (talk) 17:42, 20 October 2013 (CEST)
Please do. --向榮 /Hiong3-eng5/ (talk) 10:28, 27 October 2013 (CET)
Unfortunalely, it is likely to take some time before I have a working environment which would allow me to even run little tests. See Insect room (towards the end of the liked section) for details. --Purodha Blissenbach (talk) 10:16, 3 November 2013 (CET)

long - having really trouble here[edit]

Could someone please check these DMs as well? Some seem to be deliberately wrong. --Sabine (talk) 11:58, 26 October 2013 (CEST)

and another one expression:rich - redundancy is simply too much - sorry but I would go and simply delete. --Sabine (talk) 12:21, 26 October 2013 (CEST)
Could you be more precise? What exactly seems wrong to you? All the DMs, as expressed in English, seem right to me. Exactly what would you delete, and why? (talk) 12:28, 26 October 2013 (CEST)
Most are redundant. If something is completely wrong and I spot it, it gets deleted.
I am stopping here to add additional DMs to the General Dictionary collection coming from expressions I call to attribute the translations I have here - we have loads of redundant definitions and I will avoid them by using collections and keeping those clean. Give redundant data to people to work on it offline and they will stop working (that's and experience we made - and we had only three double voices in the table because of a group of us adding data).--Sabine (talk) 12:38, 26 October 2013 (CEST)
Omegawiki is a collaborative work, and if you want to delete a full DM, you have firstly to say exactly what you would delete, secondly to justify your idea, thirdly to wait for answers a reasonable amount of time. In the 2 examples you gave, there might be synonyms, but I'm not sure at all. (talk) 12:47, 26 October 2013 (CEST)
lol ... you want to tell me what OW is? ok ...--Sabine (talk) 12:58, 26 October 2013 (CEST)
  • It is a question of how precise we want to be - or need to be. Need may arise from translations that include one and exclude another of the definitions that we consider redundant or overlapping at first sight. --Purodha Blissenbach (talk) 20:54, 26 October 2013 (CEST)

In Oxford dictionary, there are 28 meanings or sub-meanings of "long", not taking phrases into account. In the Dictionaire de l'Académie française, there are 14. In Le Robert & Collins super senior French-English dictionary, there are 7. These 3 dictionaries give many more examples that defined meanings. The result of this precision is that these dictionaries are among the most famous in the world in there respective fields. (talk) 03:40, 27 October 2013 (CET)

Pons (1991, based on Collins, 1983) gives 11 definitions of "long" plus lots of supposedly independent composites such as long johns, long shoreman, etc.. Since some of their definitions of "long" have subdivisions - either they have the word "or" inside, or they are translated to words doubtlessly having to have their own DefinedMeanings each in the target language - I count 25 DefinedMeaning candidates, excluding idiomatic uses such as "so long". Of course this does not invalidate Sabines argument that there were overlapping or redundant DefinedMeanings for "long" in OmegaWiki --Purodha Blissenbach (talk) 09:30, 27 October 2013 (CET)

Alternative definition[edit]

Hi, if I remember well there was a field "alternative definition" - I cannot find it - I found a discussion about it in the past. It could be used for the simple definitions for the Children's dictionary, because some of the definitions we have are difficult enough for an adult to understand.--Sabine (talk) 12:07, 26 October 2013 (CEST)

Anyways, when it comes to Children's dictionary, we should add images most of the time and, if possible, painted images. --Sabine (talk) 12:09, 26 October 2013 (CEST)
Unfortunately, Kip deleted this (talk) 09:34, 27 October 2013 (CET)
Actually, in the source code, the call to show this alternative definition is temporarily commented since this feature's purpose is unclear and no one is using it. It can be used once uncommented. -- 向榮 /Hiong3-eng5/ (talk) 10:24, 27 October 2013 (CET)
I thought something like that - it's good that it is uncommented. When time is due, this can be considered. But first: many other points already require attention. When time is due things will be done and on their way. It has always been like that and it will be like that also in future. Said that: time to concentrate on adding IDs to my table :-) --Sabine (talk) 15:50, 27 October 2013 (CET)
Example where two definitions could be merged: picture and image.  Klaas|Z4␟V:  09:16, 20 November 2013 (CET)

Spelling variants[edit]

Long time ago, Gerard Meissen told me somethig about spellng variants. That is, for instance, the German expression Muße being spelled Musse in typing environments, where the letter 'ß' was not available, such as on Swiss typewriters. Since spelling variants share their definitions, translations, and all kinds of annotation not explicitly refering to spelling, having them is a time saver and avoids unecessary duplicate annotations and errors arising from duplication. Currently, I am entering spelling variants as synonymes, but imho that is not a future-safe way to go. If we had spelling variants at all, which I do no remember clearly, where are they gone? --Purodha Blissenbach (talk) 16:47, 27 October 2013 (CET)

Policy to create separate definedMeanings or not[edit]

After 7 months, should I consider all contributors interested in the "Policy to create separate definedMeanings or not" have given your opinion there, or should we wait a bit more? I suggest before you answer you take time to read the previous discussions " Need for a policy to produce DMs " and "Don't verbs and corresponding nouns of action express a same meaning?" (talk) 04:15, 27 October 2013 (CET)

The question "Don't verbs and corresponding nouns of action express a same meaning?" is a kind of a pseudo question, since it does not have a general answer. In other words, it depends on language. Whether it exists or not - for some languages, it does not make sense, since they do not have "nouns of action" or "verbs" - is a matter of language typology. You can answer it only for other languages, at best. And even then, answers may differ depending on the kind of grammar you are using. Whether or not a "noun of action" and a "verb of action" are seen as derivatives of one another may be a matter of standpoint, that is, the grammatical school you refer to. The question has better to be asked from another end: "Can different parts of speech share a DefinedMeaning?"
To give an example: There are many Native American Indian languages that lack adjectives having verbs conveying meanings represented by both adjectives and adverbs in English. English translationas of the bare verbs are customarily short phrases, such as "to be small", "to stink dog-like", "to use literally", and so on. Translations in contexts would usually use adjectives, attributive phrases, or adverbial phrases, such as "I saw a small dog", "She met an animal stinking dog-like", "He literally said ..."
--Purodha Blissenbach (talk) 09:06, 27 October 2013 (CET)
You expressed perfectly one of the questions to be answered in "Policy to create separate definedMeanings or not". Don't hesitate to add there cases I forgot, like "verb of state and adjective" or, if you think the question should better be generally answered, "Should different word classes share a DefinedMeaning?". (talk) 09:28, 27 October 2013 (CET)

Five developers[edit]

Good news on possible adoption by Wikimedia Foundation: RFC @Meta. There are 5 volunteers since early November and over 70 approvals, few oppose. We have a go soon?  Klaas|Z4␟V:  14:24, 19 November 2013 (CET)

Part of speech[edit]

I tried to update a translation of the page Help:Part of speech and while doing so, I tried to do eveything that is described there. The description is wrong or outdated. While I remember that I once added parts of speech via DefinedMeaning:lexical item (402295) somehow, it does not work the way described in the help page. --Purodha Blissenbach (talk) 09:05, 25 November 2013 (CET)

If you have delete rights you only have to check the box in the column and row where the trash can icon is on the top and eventually add a new one like I did for the archaic and anglicistic description for Dutch in the same example you mentioned. BTW, you're right that especially our help pages in some (if not all) languages need some refresing, a helluva job. We need "help on help" and more improvements to become more "reader friendly" . Kind regards from a cold Tuscany,  Klaas|Z4␟V:  09:30, 26 November 2013 (CET)
I do not want to delete anything so as not to accidentally delete data worth two or three or four days of labour that are already there and appear to be usable. I only want to make minor amendmends and additions, and get the Help page right. --Purodha Blissenbach (talk) 20:52, 26 November 2013 (CET)


I have replaced the templates used at Template:Statistics with our very own wiki tags :) It will automatically update the statistics every 24 hours. Any problem with this tags is my fault, so inform me :) Thanks for all who updates those templates that kept us updated with the stats. -- 向榮 /Hiong3-eng5/ (talk) 20:08, 2 December 2013 (CET)

Would it be possible to also update Special:DoubleRedirects and Special:BrokenRedirects and maybe other such pages somewhat more frequent? The current versions are of 2011. --Purodha Blissenbach (talk) 20:53, 2 December 2013 (CET)
I just updated the pages (note to self )
it has to be done manually because we use MiserMode ($wgMiserMode ). So, just request an update when you need one. --Kip (talk) 19:34, 3 December 2013 (CET)
Okay, thanks a lot. --Purodha Blissenbach (talk) 21:28, 3 December 2013 (CET)
Maybe, run the udate script monthly or bimonthly via cron? --Purodha Blissenbach (talk) 21:33, 3 December 2013 (CET)

Questionable pages.[edit]

--Purodha Blissenbach (talk) 12:47, 3 December 2013 (CET)

I deleted a bunch of these.
The page Sicilia is supposed to be a country portal, similarly to Portal:DE for Germany. Though it is obviously not in the right format.
I am too tired now to give an opinion on the rest. --Kip (talk) 19:19, 3 December 2013 (CET)

Deleting old stuff is too easy. Perhaps a special namespace called 'Archive: or History::, eventually as subpages of an appropriate existing namespace if that's easier. Just a weird idea,  Klaas|Z4␟V:  11:58, 4 December 2013 (CET)

Relating DefinedMeanings to one another - how to?[edit]

We have DefinedMeaning:Sutsilvan-N (607802) and DefinedMeaning:Sutsilvan-S (607808) both lackiing a definition, thus it is unclear what they precisely stand for. Yet they have parent/child relations to other DMs wihout Definitione which look even like having spelling errors.

Despite that, I am convinced that DefinedMeaning:Sutsilvan (1510671) should relate to the -N and -S varieties as common parent and consequently, the others be children of it. How would I enter these relations?

The only relations that are offered to me are:

  1. antonym
  2. entailment
  3. holonym
  4. hypernym
  5. hyponym
  6. meronym
  7. troponym

which is rather useless. A relation has to be of the kind "<current entry> relates as a <something> to <entry to be choosen>" or vice versa, but the words "antonym" or "hyponym" only name the <something> part hiding the directions of their assymetric relations. Besides that, "parent of" and "child of" relations are missing and I did not detect them elsewhere. --Purodha Blissenbach (talk) 10:30, 4 December 2013 (CET)

"parent of" and "child of" relations only become available when you put the DM into the "language" class. --Kip (talk) 22:30, 4 December 2013 (CET)


Read more Thoughts? ideas? Gerard, Kip, Sabine, anyone?  Klaas|Z4␟V:  13:59, 6 December 2013 (CET)

Looks good to me. New codes are always welcome.
It seems the only language that affects us is Dhuwal DefinedMeaning:Dhuwal_(706009) which we should probably change to "Modern Dhuwal (dhh)". --Kip (talk) 20:15, 6 December 2013 (CET)

Languages wanted for editing[edit]

I have wordlists which I am preparing for upload for these languages:

--Purodha Blissenbach (talk) 00:18, 8 December 2013 (CET)

--Purodha Blissenbach (talk) 15:10, 17 December 2013 (CET)

ok for adding them,
but I need a (even dummy) iso code for all, and the codes have to be max 10 characters (because the sql table is limited to 10, dunno why, so don't ask ;) ) --Kip (talk) 20:03, 17 December 2015 (CET)
See suggsstion T121786. But we can have that data as well as a collection index.
I have completed the left column. Please take it and truncate it, if necessary. --Purodha Blissenbach (talk) 20:53, 17 December 2015 (CET)

There are more wanted languages:

--Purodha Blissenbach (talk) 11:28, 19 December 2015 (CET)

Ok, added all :)
  • the iso code got truncated automatically to the first 10 characters (sl-rozaj-b for sl-rozaj-biske), I checked in the database.
  • I used the English name when adding, as we usually do. Some of them are long (like "thieves cant of Stotzheim") we will see if this is practical. They can be changed (either through the database, or through the automatic language names translation of OmegaWiki)
  • To get automatic translations from OW, the DM need to be associated with the iso code, by adding it to iso 639-3 collection. Probably the iso given there needs to be truncated as well, unless there is (or unless we add) a truncation in the corresponding php code of the special page that refresh the translations (Special:ImportLangNames).
--Kip (talk) 19:52, 21 December 2015 (CET)

Removed duplicate expressions[edit]

I would like to announce that we had some duplicate expressions, as noted by Cysouw. I have removed them. In case anyone is interested in checking them out, these can be found at User:Hiong3.eng5/Duplicate Expression. Btw, in this link, there are some expressions needing merging or corrections due to conflicting informations. Again, thanks Cysouw :) -- 向榮 /Hiong3-eng5/ (talk) 15:31, 8 December 2013 (CET)

Serbocroatian language should be editable[edit]

There are so many links to Serbocroatian in Wikipedia pages that we will every now and then stumble over opportunities to enter translations for Serbocroatian. --Purodha Blissenbach (talk) 01:59, 9 December 2013 (CET)

There is a problem with Serbocroatian, because it is a macrolanguage [4], and we have already the languages Serbian, Croatian, Bosnian and some others.
Native speakers should decide what to do in that case. --Kip (talk) 19:53, 17 December 2015 (CET)
Understood. If i happen to have a translation which i got from a source where it is labelled "Serbocroatian", i cannot decide which variant it belongs to. Never mind. --Purodha Blissenbach (talk) 21:26, 17 December 2015 (CET)
done OK Done. --Purodha Blissenbach (talk) 21:26, 17 December 2015 (CET)

How to extend the option list of an annotation type?[edit]

Hi, under ► Lexical annotations ▼ Option values, I want to add the choice scientific - how do I do that? Purodha Blissenbach (talk) 14:00, 7 January 2014 (CET)

Go to DefinedMeaning:lexical_item_(402295), edit class attributes (should be available for admins now). This is for general annotations.
For language specific annotations, go to the language DM page and edit its class attributes. --Kip (talk) 19:56, 17 December 2015 (CET)

Questionable page[edit]

Does it make sense, to keep page Meta:Languages? 18:59, 9 January 2014 (CET)


How do I enter Low German / Modern Lower Saxon (nds) translations?[edit]

I dot find the correct name in the strange mix of language names in various languages that is presented to me. Purodha Blissenbach (talk) 00:07, 10 January 2014 (CET)

Found the page: Expression:Low_Saxon.
Please change the translations if they seem not correct to you, and then we have to reload the language names translations (it is a special page, but I don't remember to which user groups it is available). --Kip (talk) 20:01, 17 December 2015 (CET)
done OK Done. --Purodha Blissenbach (talk) 21:23, 17 December 2015 (CET)

Bot flag to fill Armenian words.[edit]

Hi! I have creted Armenian frequency list (based on wikisource) and written a simple interactive console interface to add absent expressions. Nothing special, I do everything manually, just more useful interface iterating words from frequency list and showing DefinedMeanings to choose (editing through web interface is annoying). So, I want to ask, if I can get bot flag for User:FHenBot? If it is possible to limit bot's permissions, "insert-only for expressions" will be OK! --FHen (talk) 14:41, 11 January 2014 (CET)

Hi! I have added bot rights + edit rights to FHenBot.
I guess you plan to use Help:OmegaWiki API. It has not been used much yet. If any problem or question arises, User:Hiong3.eng5 is the API expert :) - but I can probably help as well.
As far as I know, there is not yet a function in the API to remove data. So that should not be a problem. --Kip (talk) 12:20, 12 January 2014 (CET)

Required lexical annotations for Middle High German.[edit]

There are dozens of lexical annotatons missing from Middle High German. Only three exist, one of them even a global one unspecific to Middle High German. Without correct annotations, entering Middle High German.translations is often impossible, useless or strongly misleading. How can these Annotations be made possible? Who can do that? --Purodha Blissenbach (talk) 20:08, 22 January 2014 (CET)

nth child[edit]

Do you think that this kind of expressions are useful? How many of them are there or similar vandalism?  Klaas|Z4␟V:  13:51, 30 January 2014 (CET)

This is not vandalism. If you look closely, you will find lexicalized expressions in Malay e.g. that do not refer to numerals. So we cannot avoid having them, even though they may appear boring and unneccessary in English. If we proceed to many austronesian languages, we shall find even more of those kinship kinds of lexemes which often even refer to the speaker and the listener as well. So "my third-elder brother" may have to be translated differently, depending on to whom I speak. Greetings -- Purodha Blissenbach (talk) 15:13, 16 February 2014 (CET)

Update to statistic figures on the main page[edit]

We did not have any updates to statistic figures on the main page for several days. Broken automatism? --Purodha Blissenbach (talk) 02:41, 8 February 2014 (CET)

Hi, if there were any problems then, I think it has been fixed. --向榮 /Hiong3-eng5/ (talk) 05:46, 11 February 2014 (CET)
Done Done. Yes. Thank you. --Purodha Blissenbach (talk) 09:16, 11 February 2014 (CET)

OmegaWiki API 1.1[edit]

Yeah, yeah, our API 1.0 is slim, but I would just like to ask if any one here can help me plan 1.1, what you think should we need and what parameters are needed. Our API should produce only essential outputs using JSON. I am planning also, if I have the time, to create a class and modules for javascript, php and perl to access our API. Actually, I have created one for javascript, though if anyone knows these language (which I am not really professional with these) can help. Maybe I will set up a test webpage that will access our API for demonstration. My roadmap is located at User:Hiong3.eng5/OmegaWiki API Roadmap‎ (Under Construction :) ). Thanks -- 向榮 /Hiong3-eng5/ (talk) 10:02, 20 February 2014 (CET)

btw, OmegaWiki API 1.0 Help can be seen on Help:OmegaWiki_API and accessed here -- 向榮 /Hiong3-eng5/ (talk) 10:07, 20 February 2014 (CET)

I'd love to help, Hiong, but have no botflag. Can you give it to me?  Klaas|Z4␟V:  11:34, 20 February 2014 (CET)
Botflag is necessary only if you want to add new data. If you just want to access data, you don't need any permission.
In any case, if you need a botflag I can give you one. Just tell me the name of your bot account :) --Kip (talk) 19:03, 20 February 2014 (CET)
Generally, I would like to help, but currently, I cannot even add the WikiLexicalData extension successfully to a recent MediaWiki installation owed to oh so many bugs & coding style issues. So I am out of luck when code testing is concerned. --Purodha Blissenbach (talk) 03:16, 21 February 2014 (CET)
I am not very knowledgeable about the API structures in general. Why JSON? Does not the API code in general already automatically support an array of possible output formats to choose from? --Purodha Blissenbach (talk) 03:19, 21 February 2014 (CET)
Actually, I also have a problem with my local ow copy... but since I do not have much time to spare, I think it does not mean that we cannot discuss how the API should be built. yes, we do not need to bother about the format, since MediaWiki takes care of that automatically. Why JSON? Since, if I remember correctly, I have read somewhere that they plan to remove other formats except JSON and I think many web programmers believe that JSON is the way to go. JS, PHP and PERL seems to have ways to convert from this format to their respective arrays/objects. --向榮 /Hiong3-eng5/ (talk) 16:28, 21 February 2014 (CET)
In case you want to try it an html I created to test our api, you can download it at Sourceforge. Unzip and click the index.html file. Tell me what you think. Thanks. --向榮 /Hiong3-eng5/ (talk) 06:29, 22 February 2014 (CET)
Sourceforge sent a notification, I downloaded the owlet and tried it out. Klicking one ot the images, or hitting return always yields data about 咱, no matter what I type in the input field. Is that the desired behaviour? --Purodha Blissenbach (talk) 11:47, 22 February 2014 (CET)
yes. the form is not functional yet. my point here is that this is possible. imagine, this html is powered by omegawiki web api, made simpler to program using the omegawiki javascript classes. you could create a website for your language that does not need a database, or create a web program to assist you with adding data to omegawiki. The omegawiki class used here can be found at the script folder. --向榮 /Hiong3-eng5/ (talk) 20:20, 22 February 2014 (CET)
The form became functional with the new release. Now I see a list of language headers, each followed by a definition in the default language of my browser with English fallback, followed by a list of synonymes, followed by a list of translations in various languages. Looks good! There is a JavaScript warning before each of the sections is displayed which looks like being there intentionally so as to show the original data received from OmegaWiki. Good! --Purodha Blissenbach (talk) 15:32, 26 February 2014 (CET)
Yes, but the warnings will be removed later, after I debug this. I just hope that this might produce interest in our community and maybe become somewhat like the pywikipedia, useful tools that can be used to extend our reach. If anyone knows how to use indexDB, it would be helpful. Mediawiki suggest that it is good to cache search in order not to overwhelm the server used to receive the API data. since the introduction of indexDB, it is now possible, I think, to cache the search, maybe in 30 minutes or so... If anyone knows how to do this, it would help a lot. I am just learning js and I am getting some sticking point, as if my mind can not absorb anymore info. Any help in this area will be much appreciated. btw, I further updates will be seen here and the javascript here. I will update this tomorrow. The current javascript is already outdated, the latest uses methods instead of plain functions. Thanks. --向榮 /Hiong3-eng5/ (talk) 16:12, 26 February 2014 (CET)
I canot helpl with IndexDB. It s new to me, too. Just ask, if you have questions, I might be able to help with other things. --Purodha Blissenbach (talk) 01:34, 27 February 2014 (CET)


I noticed that there are incompatibilities with formatting JSON and XML. Currently, the array produced by ow_express is:

  'expression' -> "咱",
  'ow_define_1' -> array(
     dmid -> 5455,
     langid -> 387,

where both formats are compatible. Programming wise, it seems better to create the array as ...

  'expression' -> "咱",
  5455 -> array(
     dmid -> 5455,
     langid -> 387,

Yet this produces an error with the XML, while JSON converts this without an error. It seems that XML won't accept a numeric key. Meaning,

<api><ow_express expression="咱">
  <ow_define_1 dmid="5455" langid="387"/></ow_define_1>

is valid XML while

<api><ow_express expression="咱">
  <5455 dmid="5455" langid="387"/></5455>

is invalid. JSON has no problem with numeric keys. Kip and I had encountered this problem last year, but I just notice that this XML invalidating of numeric key was the problem. So, unless anyone wants to add XML support to our API, I will design our API 1.1 with JSON support only.

Since I do not know if anyone is using our API or not, API 1.0 will still produce the same output as before. The new 1.1 API will be called using the ver parameter. So to use API 1.1, one will use the api call ...咱&format=json&ver=1.1 (not yet implemented)

--向榮 /Hiong3-eng5/ (talk) 18:26, 1 March 2014 (CET)

Don't limit the API to specific formats. Keep the OW/WikiLexicalData API compatible with everythig the standard MediaWiki API does. Just my 2 ct. --Purodha Blissenbach (talk) 23:40, 1 March 2014 (CET)
Well, there is a compromise, which I used with ow_syntrans, which is to add a period(.) to the numeric key to make it a string. so
<api><ow_express expression="咱">
  <5455. dmid="5455" langid="387"/></5455.>
is valid. I guess I will add the prefix dmid instead like below
<api><ow_express expression="咱">
  <dmid5455 dmid="5455" langid="387"/></dmid5455>
--向榮 /Hiong3-eng5/ (talk) 02:59, 2 March 2014 (CET)
Why have the id numer as part of the tag name anyways, when it appears as an attribute value yet once more? --Purodha Blissenbach (talk) 23:29, 2 March 2014 (CET)
With PERL and PHP, we would not have any problem. With an asyncronous language like javascript we have a great deal of problem. The tag name is how I can link the expression to the dm and the dm to the syntrans. If the internet speed or the sequence of GET i8s not right, the wrong translation can be given to a dm. I got this problem sometimes when I try to get the entry for big here. So it is all about compatibility with js. -- 向榮 /Hiong3-eng5/ (talk) 02:01, 3 March 2014 (CET)
@Hiong: this is also what they did at Wikidata: e.g. [5] , for the properties, they use "P248" instead of just "248". So you can use either dmid5455 or DM5455 or anything similar. --Kip (talk) 19:46, 6 March 2014 (CET)

API function ow_add_definition[edit]

I have created a ow_add_definition function that mainly resembles ow_add_syntrans but adds a defintition in a given language to a definedmeaning, provided, a definition is not already there for the language. I cannot test it, because the installation routine of WikiLexicalData is broken and I was not yet able to fix ist. So I do not have a test instance. I would like to commit my code to Gerrit so as to allow someone else with a working installation to give reviewing it a try. Anybody there who would like to? --Purodha Blissenbach (talk) 23:40, 1 March 2014 (CET)

Just commit the code, We will check it. Thanks. -向榮 /Hiong3-eng5/ (talk) 01:16, 2 March 2014 (CET)
Here it is:
Thanks! --Purodha Blissenbach (talk) 02:17, 2 March 2014 (CET)
Ah indeed you did say that you didn't test it. Sorry then for my comment at gerrit... :)
What does not work in the installation routine for you? It works for me and Hiong3.Eng5. Maybe we can help. --Kip (talk) 12:49, 9 March 2014 (CET)
Bug 56220 - I managed to come to a partial fix, but was unable to further persue it last year. I found a bunch of more prefix-related problems once the in initial one had been fixed. --Purodha Blissenbach (talk) 20:11, 9 March 2014 (CET)
I think that if we follow Kip's suggestion here and used 01 instead of 01_, the problem might be solved. The point of the prefix is that multiple databases can be added. so ow_expression might be the main db, 01_expression is for your personal db, etal_expression can be for others. It is probable that the double underscore might be the problem here. So try using
01 which will produce 01_defined_meaning
instead of
01_ which might produce 01__defined_meaning ('note the double underscore)
Since I have not tested it, the program might produce the correct 01_defined_meaning, but wreck havoc on language. --向榮 /Hiong3-eng5/ (talk) 01:45, 10 March 2014 (CET)
Moving this diskussion back to the Insect Room before it gets too off-topic here. --Purodha Blissenbach (talk) 10:07, 10 March 2014 (CET)

Bot flag for Purbo_T[edit]


may I kindly ask for a bot flag for my account Purbo_T?

Is this the right place to ask?

The first data that I want to upload is in the page User:Purbo_T/uploads/1 and User talk:Purbo_T/uploads/1 holds a short description. --Purodha Blissenbach (talk) 11:09, 4 March 2014 (CET)

Yes this is the right place. Just be a bit patient until I have some time to check the page and also your commits. I am currently overbusy (as you can see from my lack of activity...) Maybe this week-end. Sorry for the delay. --Kip (talk) 19:39, 6 March 2014 (CET)
Done! Regarding your test file, the bot seems to stop at the second blank line. (so, after the first block of four lines). Probably the blank lines should be deleted --Kip (talk) 12:30, 9 March 2014 (CET)
Strange, sometimes, it accepts an initial blank line, sometimes it does not. Since blank lines are helpful in making the input somewhat more readable and more easily, I'll look into this when I have the time to do so. --Purodha Blissenbach (talk) 20:11, 9 March 2014 (CET)

More languages wanted[edit]

The languages Cherokee (chr), Seneca (see), Cayuga (cay), Mohawk (moh) and Oneida (one) are currently not selectable. They should be enabled. I have some data to be uploaded for them. --Purodha Blissenbach (talk) 19:37, 4 March 2014 (CET)

I am sure that we have at least Cherokee. The other ones probably not, I will add them in a few days. --Kip (talk) 19:40, 6 March 2014 (CET)
Same question with the Torres Strait Creole (tcs) language. Thank you! --Purodha Blissenbach (talk) 01:28, 7 March 2014 (CET)
I added all of the above (except for Cherokee that we already had). --Kip (talk) 12:45, 9 March 2014 (CET)

Tajik writing systems.[edit]

According to, the Tajik language is written in the Cyrillic, the Latin, and the Arabic alphabets. That means that we have to create three variants of Tajik, does it not? Currently, we have only one. --Purodha Blissenbach (talk) 19:03, 17 March 2014 (CET)

This does not apply only to Tajik; Azerbaijani, Uzbek, Kazakh, and Kyrgyz, and Tatar have the same three writing systems. Many other languages also have more than one writing system that are not supported - Malagasy (Arabic/Latin), Kashmiri (Perso-Arabic/Devangri/Sharada), Swahili (Latin/Arabic). Historically, there are more - pre-1918 Belarusian had Arabic/Cyrillic/Latin, 1930-1938 Bashkir had been written in Latin, and pre-1930 Bashkir was written in Arabic. Воображение (talk) 05:30, 19 March 2014 (CET)
Hmmm, more work... will do when I have some time :). Is any of these needed right away by a contributor? --Kip (talk) 13:36, 8 April 2014 (CEST)


To get a botflag you need to create an extra account with the word 'bot' in it? If not I prefer my current name and have the botbit set, please. TIA,  Klaas|Z4␟V:  15:11, 18 March 2014 (CET)

In principle it can be done. The problem if the bot flag is given to your user account, all your contributions will be hidden from the recentchanges by default, which we don´t want. So personally, I would prefer if you create a separate account (whatever name it has), unless you can convince me otherwise, or there is a way to have your contributions appear in recentchanges still. :) --Kip (talk) 13:32, 8 April 2014 (CEST)

Done. My BotAccount is User:KlaasBot Thanks in advance for setting the flag,  Klaas|Z4␟V:  12:14, 21 April 2014 (CEST)

bugfix patch for 56220 MediaWiki WikiLexicalData prefix compatibility[edit]

This is a big patch, needed volunteers to check it for new bugs. Any volunteers? --向榮 /Hiong3-eng5/ (talk) 06:47, 19 March 2014 (CET)

I have not had time to review the newest amedments. If I got that right, you commented in favour of adding underscores to the tale prefixes of WikiLexicalData so as to not break existing installations. I strongly suggest, not to use any underscores anywhere, so as to have table prefixes compatible with MediaWiki proper. Anything else would be too bewildering. If underscores are already there, they need to go into the prefix fields of the table of datasets. The mainenance script update.php should probaly be doing this. --Purodha Blissenbach (talk) 15:07, 8 April 2014 (CEST)

Yes, temporarily if you wish. I want to finish the bug fix before worrying about the underscores. I find it hard to add other commits when files changes overlaps. Your suggestion seems doable... but bear with me a little more, let me finish squash this bug first, let Kip merge the bug fix and then we solve the underscore issue. Ok? :) --向榮 /Hiong3-eng5/ (talk) 17:22, 8 April 2014 (CEST)

DefinedMeaning identifier[edit]

In Wikidata there is a discussion if there is a way to input the DefinedMeaning identifier in a URL without inputting the name of the page:$1 is there a URL to input "5781" insteed of "apple (5781)"?

regards Bigboss (talk) 19:23, 7 April 2014 (CEST)
Then click the search button. Cheers,  Klaas|Z4␟V:  01:41, 8 April 2014 (CEST)
Bigboss, we will check if it is possible, thanks for supporting OmegaWiki :) --向榮 /Hiong3-eng5/ (talk) 06:49, 8 April 2014 (CEST)
Currently you can use which should redirect you to the correct page. Getting rid of the parentheses is a work in progress. --Kip (talk) 13:24, 8 April 2014 (CEST)
Kip, I have committed a patch for review at
This, will make using possible, sans bugs, that is :)
Kindly check if ok. But since this patch depends on the latest changes in MediaWiki (json etc), when we use this, we must upgrade to the latest MediaWiki version. Testing on older omegawiki version, this works, but with the latest version, the internationalization did not display well, since I do not have the latest MediaWiki version yet. --向榮 /Hiong3-eng5/ (talk) 13:57, 8 April 2014 (CEST)
I tested with the latest MediaWiki, it works :) --向榮 /Hiong3-eng5/ (talk) 14:41, 8 April 2014 (CEST)
The URL is:$1)
you don't have to have a string there, but you need the underscore and the brackets. --Purodha Blissenbach (talk) 15:14, 8 April 2014 (CEST)
Thanks Hiong3.eng5 :) I will upgrade MediaWiki and patch the extension later. I will probably do that over the week-end, to have more time in case something goes bad with the MW upgrade. My weeks are busy. --Kip (talk) 22:57, 8 April 2014 (CEST)
Thank you for all! --Bigboss (talk) 00:50, 9 April 2014 (CEST)

MediaWiki 1.22[edit]

I upgraded MW to 1.22, with the new way to access DM by number (Thanks Hiong-eng)

There might be bugs, please tell.

In particular, some other developer modified our code to make it fit for MW 1.23, so that some parts of it are not compatible with 1.22 anymore (but 1.22 is the stable version, so we keep it). Such errors would be e.g. "missing getPageTitle function". --Kip (talk) 21:09, 10 April 2014 (CEST)


It seems there are problems in displaying these. They are displayed well in my local copy. Maybe the problem is in the saved cache? I do not know how to refresh them. --向榮 /Hiong3-eng5/ (talk) 03:09, 12 April 2014 (CEST)

Where are the problems? For me they are working fine. --Kip (talk) 19:22, 12 April 2014 (CEST)
English interface, OmegaWiki NS. ex. ow_language. --向榮 /Hiong3-eng5/ (talk) 23:19, 12 April 2014 (CEST)
Ok fixed, it was a problem with cache (as you guessed :) ). All other languages were apparently working fine already. --Kip (talk) 14:12, 13 April 2014 (CEST)

Missing languages[edit]

Can we have these languages enabled: Modern Low Saxon (nds), Westphalian (wep), беларуская in тарашкевіца orthography (be-tarask), Modern North Frisian (frr), Swabian (swg) ? -- Purodha Blissenbach (talk) 22:48, 21 April 2014 (CEST)

We still need to have the languages:
--Purodha Blissenbach (talk) 00:45, 17 December 2015 (CET)
Added the three as editable. --Kip (talk) 20:33, 17 December 2015 (CET)
done OK Thank you! --Purodha Blissenbach (talk) 20:59, 17 December 2015 (CET)

Bureaucrat and my absence[edit]

I will have no Internet access in July (thank you Deutsche Telekom), because I am currently changing house (which explains also my low activity the last months).

During that time I temporarily give bureaucrat access to Hiong3.eng5 so that he can give user rights to new users. (Hiong3.eng5 already has developer access to the server, so that he is not actually gaining anything new with bureaucrat rights.).

--Kip (talk) 14:55, 16 June 2014 (CEST)


Wikidata does it. We too?  Klaas|Z4␟V:  16:38, 26 June 2014 (CEST)

It does seem ideal that we improve our UI. The problem would be the lack of programmers. I do contribute, but my knowledge is limited. There are parts of the WikiLexicalData code that I still do not understand. Kip understands them, but is not able to contribute due to being limited to mobile internet. You could give suggestion on what we can improve upon and when either... 1. Kip regains internet connection and has time to code, 2. I miraculously understand the essential code to work upon it, or 3. Someone willing to contribute comes along... then maybe we might have a cooler UI :) --向榮 /Hiong3-eng5/ (talk) 11:24, 12 July 2014 (CEST)
I would be willing to contribute but have limited time and other ressources and still not even a working installation for tests. :-( --Purodha Blissenbach (talk) 16:31, 14 July 2014 (CEST)

2 usages[edit]


I have a problem with the French noun "germain", meaning "sibling", which has 2 usages: one is archaic (It seems it was common French up to the 19th century.), one is scientific (This word is still used or used again, but only by scientists, notably geneticists.). If I just check the boxes "archaic" and "technical", how would I indicate a word being "archaic technical", such as the French "humeur" ("humour"): "Any of the fluids in an animal body, especially the four "cardinal humours" of blood, yellow bile, black bile and phlegm that were believed to control the health and mood of the human body."? In other words, are the usages supposed to be understood with and "AND" coordinating conjunction, or with an "OR" coordinating conjunction? and how to indicate the other one? (talk) 17:13, 9 July 2014 (CEST)

That's not a technical, but a medical meaning. If it's still in use among doctors it's in my opinion not archaic. Some Dutch words are archaic here and/or in the northern part of Belgium, but still in use in South Africa. So you have to be careful calling any word archaic, it's POV to speak in Wikipedia lingo ;-D  Klaas|Z4␟V:  09:30, 10 July 2014 (CEST)
Not all geneticists are medical doctors. Many are biologists. If fact, rather than "technical" or "medical", "scientific" should be better, but is not in the list. Anyway, this is not my question, which is: How to distinguish a word of archaic medical usage from a word with 2 usages: an archaic one in common language, and a medical current usage? This is just an example. You point out another complication: a work can sound archaic in a part of the word (and OmegaWiki readers may need this piece of information), and not in another one. Another point is that, quite often, words lose their strength, little by little. For instance the French "étonné" (literally "struck by the thunder") used to be very strong ("stunned", "stupefied"), and now just means "surprised". Sometimes, words change connotation. For instance the notion of human "race" was of common usage up to World war II. So it seems that the "usage" fields should be both localised and dated... (talk) 11:00, 10 July 2014 (CEST)

Why you don't answer questions on your Talk page?  Klaas|Z4␟V:  11:20, 10 July 2014 (CEST)

simply make it two words having two definitions having spelling and likely few more things in common. --Purodha Blissenbach (talk) 16:21, 14 July 2014 (CEST)

Social networks[edit]

I discovered that except our blog, Facebook and Twitter we're on Google+ as well, though contributions are few. Anyone cares to help? You can find it here. I think we could become more known if we show off more in public.  Klaas|Z4␟V:  09:30, 10 July 2014 (CEST)

I created that page, but have no time to update. Sabine wanted to do it at some point, but then became busy as well. --Kip (talk) 18:43, 17 August 2014 (CEST)

She helped me with the TwitterAccount as well. Now I try to it on my own again so posts are irregular. Not a big deal, is it?  Klaas|Z4␟V:  15:47, 3 October 2014 (CEST)

Question marks in lemmata[edit]

E.g. DefinedMeaning:Selbstmörder_(1181892). Is this right or does in English suicider exist?  Klaas|Z4␟V:  17:57, 12 August 2014 (CEST)

Both are correct according to Wiktionary. --Kip (talk) 18:48, 17 August 2014 (CEST)

Usage: Hyperbolic[edit]

I added the defined meaning "A great quantity (of a liquid)" for the English word "sea" and I believe that the Lexical Annotation usage="hyperbolic" should be indicated. Could this choice be added to the selection list for "usage" please? --InfoCan (talk) 20:50, 11 September 2014 (CEST)

I also suggest the usages "literary", "scientific" and "affective". (talk) 04:36, 17 September 2014 (CEST)

added "literary", "scientific", "hyperbolic".
Not sure about "affective". Do you mean "affectionate"? (fr: affectueux). --Kip (talk) 16:17, 15 October 2014 (CEST)
Thank you! Yes, I mean "affectionate", like the French words "mignon", "menotte" (hand), "frimousse", "sœurette"... (talk) 05:37, 19 October 2014 (CEST)
Added affectionate. I hope it is ok with the translations. Otherwise we would have to add a new definition for Expression:affectionate that is specifically for the linguistic meaning. --Kip (talk) 16:33, 21 October 2014 (CEST)
Thank you Kip! (talk) 03:38, 22 October 2014 (CEST)

Kazakh word for red was misspelled[edit]

Can someone change кызыл into қызыл.--Bauer (talk) 03:42, 13 October 2014 (CEST)

Done! --Kip (talk) 16:09, 15 October 2014 (CEST)

No english definition on pomegranate[edit]

Can someone write an english definition for this word?--Bauer (talk) 15:36, 13 October 2014 (CEST)

done! --Kip (talk) 16:08, 15 October 2014 (CEST)

Duplicated words for sister in kyrgyz[edit]

I accidentally pressed save button 2 times, can someone remove these duplicates here --Bauer (talk) 06:41, 15 October 2014 (CEST)

Done! --Kip (talk) 16:02, 15 October 2014 (CEST)