Talk:Requests for comment/Adopt OmegaWiki

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search
This page is for discussions related to the Requests for comment/Adopt OmegaWiki page.

  Please remember to:


Wikimedia Community Logo.svg

The OmegaWiki community is a priori in favor of it[edit]

The most active people at OmegaWiki also belong to the Wikimedia Foundation community (mostly Wiktionary and Wikipedia). The source code is an extension of MediaWiki, on the same svn server. So, basically, becoming an official WMF project is for us no problem, and actually can only be a good news (better server, and better visibility).

Often, when people start at OmegaWiki, they ask: "Why aren't you a WMF project?" I don't know ;-). --Kip 09:00, 19 July 2010 (UTC)


Advantages and disadvantages of OmegaWiki vs. Wiktionary[edit]

The WikiLexicalData extension should be improved (from user experience and linguistic points of view) before its implementation could provide the functionality currently expected on Wiktionary.

Advantages[edit]

  • It is possible to create one billion entries per Wiktionary: All synthetic languages could import at least ~10M of words, but probably more if all common phrases are counted. Thus, it means that we need just 100 synthetic or polysynthetic languages to create one billion entries per Wiktionary. This is very large number and while it is possible to keep technically one such project, presently it is hardly possible to keep a number of projects with more than billion of entries.
    Please explain this, I didn't understand anything here. One billion entries per each Wiktionary? Etc. Nemo 22:24, 19 July 2010 (UTC)
     :) Yes. I'll give the example through Serbian (Croatian, Bosnian) language, as a synthetic language with moderate number of inflected forms. The biggest dictionary of Serbian language has ~500.000 words. It can be said that Serbian language has around 1M of words. Every noun in Serbian has two numbers and seven cases, thus 14 semantically different forms, but usually ~10 different phonological forms. Every adjective has ~200 semantically different forms (probably ~100 phonological forms) and every verb has ~500 semantically different forms (probably ~200 phonological forms). So, without going into proverbs and their forms, we can easily have 100M of articles. And all Wiktionaries are multilingual dictionaries, too. As Wiktionary is a multilingual dictionary, too, Serbian Wiktionary need to add just 10 more similar synthetic languages (let's say, just other Slavic languages) to reach 1 billion of articles. (Thus, not 100, I've made mistake in counting first time.) So, all of the Wiktionaries are in that sense fully redundant. Instead of solving that issue just once in OmegaWiki, we need to solve it ~200 times. --Millosh 06:23, 20 July 2010 (UTC)
  • It is structured formally, and therefore easier to export to other dictionary projects/formats.
  • It is possible to have the interface and add definitions and translations in any language without the burden of requesting the opening of a new project.

Disadvantages[edit]

  • Philologists like more open form for dictionaries.
  • OmegaWiki is distant from the wiki principle. Software fixes should make it closer.
  • It could partially vampirize Wiktionary community.

Not instead of the Wiktionaries, but in parallel please[edit]

The arguments of the people contributing to the Wiktionaries are the following:

  1. The lingua franca at OmegaWiki is English.
    • We have beer parlours in every languages to discuss issues specific to languages, and the interface is also multilingual, but talks about definitions of concepts, which are multilingual (i.e. have a definition and translation in several languages) are best held in English. It is not viable to translate the discussion in several languages and have them run in parallel.
    • So, many people, for example at the French Wiktionary, will not be willing to move to OmegaWiki, because they are not fluent in English.
  2. OmegaWiki does not have all functionalities (but it already has good functionalities).
    • The most cited missing feature is inflexions.
    • It is understandable that the people from the Wiktionaries are not willing to move to OmegaWiki if they lose features.
  3. The data from the Wiktionaries cannot be moved to OmegaWiki automatically (contrary to what was believed at the beginning), because the data structure is different.
    • It is not possible to match together definitions from different Wiktionaries
    • Also, some entries at Wiktionary are defined by synonyms, which is not possible at OmegaWiki, because synonyms share the same definitions and translations.
    • We do not want to lose the data from Wiktionary. We do consider it for OmegaWiki (mostly when adding translations), but has to be copied manually, and that takes time.

However, for Wiktionaries with small community that are not growing much, I think it is more relevant to contribute to OmegaWiki and maybe shut down the Wiktionary. --Kip 09:17, 19 July 2010 (UTC)

Licensing[edit]

The idea at OmegaWiki is to have the less restrictive license possible. We do not want to license the data, we want it to be used by others, as much as possible. So, license is really not an issue. --Kip 08:55, 19 July 2010 (UTC)

www.wiktionary.org sounds like a good home[edit]

Not a heck of a lot is being done with www.wiktionary.org currently, so that address sounds like a great home for this worthy multilingual project.--Pharos 13:49, 19 July 2010 (UTC)

Ok for me as well. --Kip 14:28, 19 July 2010 (UTC)
Things get complicated if we add them to the main domain – how about mul.wiktionary.org ? Cbrown1023 talk 15:42, 19 July 2010 (UTC)
In that case, I think we should perhaps call it something like 'Multi-Wiktionary'.--Pharos 17:52, 19 July 2010 (UTC)
Maybe the main domain would be hard to implement for now, but it still might be a thing to consider for the future. Also instead of mul.wiktionary.org, we could go more with the existing community tradition and just call it omega.wiktionary.org.--Pharos 15:54, 19 July 2010 (UTC)
Omega.wiktionary.org will probably not be accepted I think. --OosWesThoesBes 15:59, 19 July 2010 (UTC)

Activity statistics[edit]

The statistics currently don't display the number of active users, but only the number of registered users (2600). How active is the project? --Nemo 22:21, 19 July 2010 (UTC)

The number of users who can edit the data is those of the "wikidata-omega" group, i.e. 862 users.
Among this group, I'd say we have about 10 regular users. --Kip 07:51, 20 July 2010 (UTC)

Are those statistics correct? 0,12 edits per page looks a bit impossible.. --OosWesThoesBes 07:51, 20 July 2010 (UTC)

I don't know. It is possible in that we have imported two corpus (UMLS and SwissProt, which are in a different database than the community corpus), and the corresponding pages in the "Expression" namespace have been created directly in the database with sql queries (I think). So it'll not show as modifications. --Kip 07:59, 20 July 2010 (UTC)

Extension:Wikidata[edit]

http://www.mediawiki.org/wiki/Extension:Wikidata does not exist.
http://www.mediawiki.org/wiki/Wikidata does exist.
Is all the code for OmegaWiki in the SVN repository? John Vandenberg 01:12, 20 July 2010 (UTC)

Yes http://svn.wikimedia.org/svnroot/mediawiki/trunk/extensions/Wikidata/OmegaWiki/ --Kip 07:52, 20 July 2010 (UTC)

These links are outadated by now. --46.115.82.219 18:29, 29 December 2012 (UTC)

New link: http://www.mediawiki.org/wiki/Extension:WikiLexicalData
The code is now on Git https://gerrit.wikimedia.org/r/gitweb?p=mediawiki/extensions/WikiLexicalData.git;a=summary

comment[edit]

I support this generally, if we can find the right way forward for wiktionary, et al. SJ talk | translate   11:39, 8 July 2011 (UTC)

I am an active developer/contributor at OmegaWiki. Is there anything I can do? --Kip 12:43, 10 October 2011 (UTC)
You could write some specifics about the implementation could look like, how it would work (in parellel) with Wiktionary, start asking around what extensions you use might be a problem for the WMF and would need to be reviewed etc., request some comments on wikitech-l and Wiktionaries village pumps... Only if the situation is well assessed there's some chance that this could happen, IMHO. Anyway, try writing to SJ in private, he'll probably miss messages like this. Nemo 20:18, 28 October 2011 (UTC)

Although redundant in some ways, both Wiktionary and OmegaWiki are great projects. When it comes to translations OmegaWiki is both technologically and conceptually superior. On the other hand Wiktionary is much broader in its linguistic and dictionary goals. Besides that it simply has a larger community and a greater dataset. What I would like is that the functionality regardings translations (adding a translation will be shown on all wiktionaries) will be slowly implemented in Wiktionary. Of course this should go step for step and maybe will not include all wiktionaries at once. What also would help is links between both wikis, diff possibilities, extensions for wiktionaries, conversion scripts and that Omegawiki would have a xml (dump) format similar to Wiktionary. OmegaWiki is too important to get neglected! User Raymondm (Dutch wiktionary)

A roadmap[edit]

I agree that OmegaWiki's definition-based structure is a better design than Wiktionary's word-based structure. The question is how to convert from one structure to the other. I would like to suggest that this can be achieved by cross-indexing in both systems. Every defined meaning (DM) under an entry in Wiktionary should have an external link to a DM in OmegaWiki and every synonym/translation under a DM in OmegaWiki should have an external link to an entry in Wiktionary. The people who would be working on this task ("integrators") would just be setting up cross-links. Content addition should be done only on Wiktionary, keeping in mind the requirements of OmegaWiki, that a definition should define, not just give a synonym. Then, bots can carry cross-linked content from Wiktionary to OmegaWiki. Once all the cross-references are done, you would have one integrated system that is organized differently at two different web systems. At either Web site, you would see two different perspectives of the same data depending on where you click. When this hyperlink-based integration is achieved, the two systems can be also merged physically and administratively. I think this should be the process to adopt OmegaWiki into MediaWiki. --InfoCan (talk) 20:20, 10 March 2012 (UTC)

Waiting for what?[edit]

With Wikidata.org gaining notoriety in 2012, I thought that was the project that soon would finally give the Wiktionary the means to become what it is meant to be. I recently discovered OmegaWiki and it comes to me as a bit of a shock that it has existed for years and with apparently so little resources.

If I understand correctly some people would like the respective roles of / the relationship between the existing Wiktionaries and OmegaWiki to be clearly defined before starting to invest in the latter. I disagree.

It is true that OmegaWiki would be very useful to improve the Wiktionary. More accurately OmegaWiki is necessary to achive some of the Wiktionary's goals. Namely the project to provide in every language a translation dictionary from any language to any language is cubic in the number of languages (i.e. if we consider 100 languages the result should be roughly equivalent to 1 million printed books) with of course a ridiculous amount of data duplication and it would be equally ridiculous to hope to ever finish it without some structure like OmegaWiki's.

However OmegaWiki is not just a tool to make good Wiktionaries; OmegaWiki is very interesting and usable by itself. It features an interface available in many languages which allows to read and edit the database. It presents the information in a structured way that I already find superior to the Wiktionary's. OmegaWiki will achieve what the Wiktionaries can't not to please them but for the sake of it. However OmegaWiki is a database that the Wiktionaries and the other projects should be able to use if it helps them achieve their own goals.

Therefore I think OmegaWiki should just be set free in parallel of the existing Wiktionaries. If these projects would only collaborate as much as it serves their own goals they would improve greatly. However by "set free" I don't mean "left as it is". OmegaWiki needs 1/ some developer time to improve it technically and 2/ some advertising so that it can start to "vampirize" existing communities; in other words, people might decide that it is a good project and start contributing to it if only they heard about it!

OmegaWiki needs and deserves the foundation's support. Please! --Rinaku (t · c) 11:33, 27 January 2013 (UTC)

I totally agree. I was very shocked when I discovered OmegaWiki. I link this page where it says: "it would be nice to become a WMF project, even without the need to implement whatever missing features. We would also be happy to be hosted on the WMF servers, for a faster website."--Luca Ghio (talk) 10:47, 28 January 2013 (UTC)
See also OmegaWiki#Merging_with_the_Wikimedia_Foundation with the more recent thoughts about integrating OmegaWiki in the WMF. Maybe we should merge the two pages? or move the relevant section to the page Requests_for_comment/Adopt_OmegaWiki. --Kip (talk) 13:24, 28 January 2013 (UTC)
Can this request be proposed here as a new idea for improving the Wiktionary project?--Luca Ghio (talk) 20:41, 30 January 2013 (UTC)
The idealab may be able to help. Please also see below. SJ talk  07:38, 11 February 2013 (UTC)
Someone already posted a similar idea on the Idealab: Grants:IEG/Wiktionary - the way it should be --Kip (talk) 14:57, 18 February 2013 (UTC)

Next steps[edit]

Kip, some suggested next steps:

  • Add a section where people can indiciate support and interest in OmegaWiki. I don't know why that wasn't there already, but it is essential :) Revise the OmegaWiki page so that it is clearly for a proposed Wikimedia project. For comparison, two good examples are the Wikikids page and the former Wiki Travel Guide page. (JamesA can help you out here - he indicated interest on my talk page and did a fine job with the Wikivoyage process.)
    • Let people know you are actively pursuing this idea - on OW, on the wiktionaries you frequent. Get them to sign up directly on the OmegaWiki page; once there is enough support there - at least 30-40 people - you can pursue a specific RfC. (by then you should have a clearer idea of what sort of adoption makes sense.)
  • Rewrite this RFC page so that it is purely about a specific proposal to adopt OmegaWiki. I suggested a few sentences at the top of the page. Leave all of the details about how OW works and how it compares to Wiktionary, in its own page.
  • The section about merging with the WMF can be even more clear. (And those details should be on the OmegaWiki page, not on both pages. Just link to that detail from this RfC.) Ways you can improve that part:
    • Be more specific. Who would the technical contact be on the OW side, for such a switch? What current maintenance tasks do you carry out, and what are the details of your server setup? What's a timeline for a potential migration (e.g., physical move, Unified Login, merging any images, &c)? You can look at what was done for Wikivoyage for ideas. Talk to Stefan2 or LtPowers or DerFussi or JamesA, or others in the Wikivoyage/Lounge - they can help you understand what the process was like for them.
    • Outside of this sort of migration and updating, leave out requests for extra programmer manpower to optimize OmegaWiki. Developer support probably has to come through recruiting new contributors. Move talk about "extra programmers" to the talk page (along with brainstorming about where to find said programmers :).
  • Link to existing docs that describe how OW and Wiktionary work together. If there is a long-term plan for OW features, such as the Wikivoyage long-term goals, you might link to that as well.
    • Ask for interested participants to start brainstorming how to better align OW with Wiktionary and TranslateWiki so that they work together to provide a coherent language and definition framework. I know there are already some historical drafts like this available.

SJ talk  06:22, 11 February 2013 (UTC)

Thanks! It is a lot of work, but at least I know where to start :) --Kip (talk) 10:47, 12 February 2013 (UTC)
I reorganised the page, as suggested by Sj.
The page OmegaWiki is now about a proposal to adopt OmegaWiki. Details about OmegaWiki itself have been moved to http://www.omegawiki.org
The page Requests for comment/Adopt OmegaWiki is now small and only about commenting the proposal. Details about the proposal have been moved to OmegaWiki.
--Kip (talk) 12:49, 14 February 2013 (UTC)

Finding new contributors[edit]

If there is enough community support for doing this - and I believe there is - it would certainly attract attention. This would add new tech and language experts to the community of users. (compare the traffic and visibility Kiwix received after there was a banner campaign for it on a single Wikipedia-language for a short amount of time.) SJ talk  08:10, 11 February 2013 (UTC)

Technical input from supporters[edit]

This has been a good discussion so far. The only major issues raised have been security concerns, and questions of how this would eventually integrate w wikidata.

  • Can you find some community developers to comment on the current codebase / give it a code review / help bring it up to security standards?
  • Can you add a section about Omegawiki discussing how it would work with Wikidata? Is there already some sort of automatic synchronization?

Regards, SJ talk  23:00, 11 March 2013 (UTC)

Hoi, as far as I am concerned the best option will be to have OmegaWiki functionality be implemented in Wikidata. There is a huge overlap anyway in the functionality. One challenge will be to have a user interface that is as good as OmegaWiki... The next challenge will then be to have a NICE user interface.
When OmegaWiki would be hosted by the WMF, it would be for as long as it takes to no longer need the old code.
Once a concept in OW is linked to Wikidata, the Wikipedia articles in any language are available.
Thanks, GerardM (talk) 09:16, 12 March 2013 (UTC)
Hello GerardM,
Could you explain a bit more the advantage of reimplementing OmegaWiki in Wikidata compared to putting links on each one to the other, or interfacing the two in any other way? (If this has already been discussed somewhere else, a link to the discussion would be fine.) --Rinaku (t · c) 19:45, 14 March 2013 (UTC)
OmegaWiki and its technical merits are re-implemented in Wikidata as it is. Maintaining two platforms is prohibitivly expensive. When OW and WD are functionally the same, it means that functionality that IS expected of Wikidata like searching Commons in multiple languages can be achieved. OW has shown that this can be done. When keeping OW is not what anybody wants, why discuss it further ? GerardM (talk) 14:12, 18 March 2013 (UTC)

Wikitionary future[edit]

Please consider to involve yourself in the Wiktionary future topic which is parallel to the present one. --Psychoslave (talk) 11:31, 23 March 2013 (UTC)

Closing[edit]

@Kipmaster: This proposal has been around since 2013. So much time has passed, and so much has happened around Wikidata, that the comments have mostly lost their meaning. I propose to close this RfC as 'no conclusion' and suggest that if you're still feeling strong about this, you start a new RfC, with fresh data and information. Effeietsanders (talk) 02:50, 23 January 2018 (UTC)

Support Support --Liuxinyu970226 (talk) 08:19, 20 February 2018 (UTC)