Proposal talk:More multi dialect wikis

From Strategic Planning

This is a problem of the individual community. Anyway, w:lmo: already does this, because "Lumbard" does not exist. --Nemo 07:06, 3 October 2009 (UTC)[reply]

The repercussions are major for individual communities, but it is strategic for wikimedia, do we prefer creating different views within multi dialect wikis or to create separate wikis for each dialect? To harness en:the wisdom of crowds we should prefer multi dialect wikis where possible. WereSpielChequers 11:23, 3 October 2009 (UTC)[reply]
Move this if it's in the wrong place. I have no idea how this page is organized. I was born in London and have lived most of my life in the U.S., so I'm very familiar with the accents, spellings and usage on both sides of the Atlantic. The differences are just not as great as people sometimes make out ("two great peoples, divided by a common language" as Churchill put it), and I'm not sure that a whole new version or sub-version [pun unintended] is really necessary. And within those dialects, folks have other preferences (e.g. U.S. & British non-Christians who prefer CE/BCE), so you'd have to redivide again. I think the literary forms of French and French-Canadian are even smaller, although the Brazilian/Portuguese split might be a bit wider. Shakescene 00:59, 11 October 2009 (UTC)[reply]
This isn't a proposal to divide Wikipedia into English and American English versions, this is about retaining EN as one wiki but allowing users to choose how they want certain data presented. So three users looking at the same article might see:
  • "came to the throne in 37 AD, and appointed his favorite horse a consul"
  • "came to the throne in 37 CE, and appointed his favorite horse a consul"
  • "came to the throne in 37 CE, and appointed his favourite horse a consul"
If this was implemented perennial proposals such as Proposal:American English Wikipedia would be rarer, and newbies shifting between BCE and BC and between American English and English spelling would be much rarer. WereSpielChequers 23:21, 11 October 2009 (UTC)[reply]
  • Currently we have a proposal at pt.wikipedia for configuring LanguageConverter for Portuguese, so that we could use it in portuguese wikis: Conversor de idiomas para as variantes do português. We have drafts for two of the conversion tables being developed: for pt-pt (Portugal) and pt-br (Brazil). Furthermore, there is a draft of LanguagePt.php to be reviewed by someone who knows PHP better. From what was said at the talk page, the language conversor seems to be a good option to minimize some of the current problems we have at portuguese wikiprojects. Helder 13:24, 17 June 2010 (UTC)

See also

Would it be possible to use this for

HenkvD 07:36, 3 October 2009 (UTC)[reply]

Definitely the American and Canadian English Wikis, I will leave it to French and Portuguese speakers to answer for their languages, but I would certainly hope so. WereSpielChequers 11:13, 3 October 2009 (UTC)[reply]
Yes. It would be useful for pt and pt-br too. Helder 14:40, 9 December 2009 (UTC)
Strongly agree with usefulness for pt and pt-br Vapmachado 18:17, 25 June 2010 (UTC)[reply]

Playing the European field

User Nemo above said : This is a problem of the individual community.
I respectfully wish to disagree. Imho This propossal should indeed not be construed to subdivide wiki projects, however the way this is proposed migth eventually have that undesired side effect. Allow me to expatiate on this subject by using the Dutch example to clarify what are the current effects and what could / imo -should- be done about it.
The Dutch (speaking) wiki currently covers 22 mio people 16 mio in the Netherlands, 6 mio in the northern half of Belgium the region of Flanders and a few thousands elsewhere in the world. What do we see in the wiki project : the majority of regular contributors and mods reside in The Netherlands, and a minority in Flanders, however their ratio is not even near the 37.5% that the population ratio would suggest. It's much closer to 18% (or in the region of 3 Flemish vs 16 Dutch). Despite a good collaboration within the wiki community the Flemish participation in the Dutch wiki has declined (vs the dutch growth) and is actually nearly entirely responsible for the net decline of participants in the dutch wiki. Currently - and I do participate in that effort - there is an effort on the way to create a Belgian Luxemburg chapter which would have to operate in at least 3 and possibly 5 languages : English as the administrative platform, and dutch, french for contributions perhaps ading German and Letzebuergish as further official contribution languages, there is even a "movement" around Limburgisch.
The consequences are that next to the current Dutch, French and German wiki projects a tri-lingual Belgian project would emerge. That is a terrible waste of resources. Even if a Belgian wiki chapter is created I'm in favour of NOT allowing it to set up its own infrastructure in the Wikipedia/wiktionary project they should be allowed to manage a sub-part of the Dutch, French and German, wiki's and those part ought to be able to interlink with each other.
If we take the Dutch wiki as an example : there would be pages that are clearly Flanders-related and others clearly Netherlands-related, and there would be a common ground. A collaboration protocol should be estalished how moderation should be organised as there is clearly a difference between what people in the Netherlands and those in Flanders find important and also on how they bring that difference into wording. The fact that first-time editors abandon their efforts within the first year is the way symptomatic : certain mods of Netherlands origin are clearly insensitive (or choose to be so) to those differences and have bots running to particularly edit away the use of so called "typical" regionalistic use of dutch, and the fact that the reference dictionary (Van Dale's woordenboek) is made in the netherlands and does list typical "flemish" as regionally used language while listing typical "northern" language as standard language is revolting to users from the south, and making them feel unwelcome and unappreciated, unrespected even.
This ,leads us to my final argument of the 22 mio dutch the largest minority lives in the area roughly determined as the provinces of North-, South-Holland and Utrecht (also referred to as "de randstad", an upside down diamond between Utrecht, Gouda, Rotterdam, Scheveningen, Amsterdam and back to Utrecht). They speak what is often referred to as "standard dutch". Outside that area are les populated area's with their own accents and dialects. Especially in the south (south of the river Maas) are bordercrossing regions like Limburg, Brabant and the Flanders-provices that have maintained their dialects as primary language of communication using standard dutch like a foreign language only at school and for official purposes. Non-dialect speakers are often quickly categorised and isolated as "fremkörper"(sociological strangers). So as a Brabant speaking Belgian I would be more readily accepted by Netherlands Limburgers in say Maastricht then say a tourist from Amsterdam. Currently pages are kept in standard dutch only, many associations exist that wish to defend the heritage of their dialect, and broader of their region, without being overly-nationalistic about it. I would suggest that particular pages are made including wiktionary pages that would be written in dutch and dialect, about the dialects and that would allow to find and cross-refence standard and dialect vocabulary.
My suggestion would be to have regional portals with particular edit/moderation teams : to gain access to that editing/moderating crew and to allow you to edit the dialect version of a page you would have to be a recognised speaker of that dialect, and the local universities would be able to lend linguistic support to the theoretical description/grammar of the dialects, any non-accepted editor can make suggestions for a lemma through the discussion page of a lemma. Your knowledge of local customs, or produce, would make you eligibale for the editor's crew. The contents of the regional portals would be geographical, historic, cultural items, local item such as regional products, events, even regional cuisine, traditions, language, regional and local "lifestyle" etc etc. Those regional/ local / dialect portals would be "off-limits" (for unsupervised editing) to users from outside the region, that would be a promotion of the region/ dialect. A place where people "proudly" show off their region, their dialect, having a "home in wikipedia" where they are not swallowed by the largest minority, as people in Flanders say : "we like to have our particular flavour of sausage, we don't like to be mixed into the unified (unflavoured) saucage from up north, but in the end we all are part of the string of sausages". Being able to feel at home about your region is to an extend nationalistic, but giving them that retreat it makes them less defensive and therefore they will more readily accept the diversity around them.

1- Regional/ local / dialect portals within the unified wiki of their linguistic region, and technological support to implement such portals, with support for dialect pages, multi-lingual pages.

2a- Collaboration and moderation protocols that would leave regional minorities and dialects off limits to those unfamiliar with the region, perhaps in an initial phase supervised editing based on your affiliation, so that you request access to the moderation team with your credentials, with a temporary supervised editing feature until a modus vivendi between regional editors and those of the linguistic region.

2b- Establishing rules on how to makes regional minorities feel respected and welcome in Wikipedia

3- technological solutions to establish quicklink (similar to "de:referenzwort" or "fr:mot de référence") inside the text across different languages

This is a first brainwave please make your comments and suggestions further on this page --DerekvG 17:58, 16 November 2009 (UTC)[reply]

Hi Derek, This proposal might solve that problem, provided Flemish and other Dutch dialects could be programmaticaly linked. WereSpielChequers 18:46, 17 November 2009 (UTC)[reply]
Hi Werespielchequers, This proposal is far from finished it deals first and foremost with the user-behaviour matters, and perhaps a establishing a protocol to help decide in conflicts. Your reflection is a technical one and it brings those aspects into view of my proposal, could you expand what your statement implicates, what is involved to make this happen?
I happen to belong to a society in the making that is reflecting on a written representation, on consistent and historically correct pronunciation and a vocabulary (with in excess of 3000 entries) of the Antwerp dialect for instance. We have made certain rules regarding representation that stem from both from the origins of our dialect as from its current place within the dutch-speaking language area. We could contribute some of these aspects to spellcheckers and thesauri. what could we contribute to advance the matter?
However I think there are other matters to resolve in order to implement multi-dialect wikis :
# It could happen that we make dialect contributions on standard language pages : is there a feature to make sure that spell checking bot use the correct checker and thesaurus to check a dialect section and how would that operate, if anyone could provide an explanatory link that would be helpful
# Certain pages could be written both in dialect and in standard language how can we implement the switch feature
# another face of the same coin is that a word like "law" has a different explanation between 2 countries operating in the same linguistic area. The current explanantion Wet(law) is specific to the Netherlands, there is no explanation about Belgium, and despite it being closely similar it is different. We need protcols to differentiate that. If I extrapolate this example "LAW" in (primarily) English speaking countries there should be different explanations of law according to it's British (perhaps even English, Welsh, Scottish, Ulster), US, Australian, Canadian definition(s)
my proposal is to think about these matters, and to address the problem as a whole rather than as different singular problems (currently there are fragmented proposal in different discussions in the strategy-wiki's , cause the solution of these interlinked/interdependent problems needs/merits a coherent and consistent solution.
--DerekvG 14:57, 18 November 2009 (UTC)[reply]
Hi Derek, Currently there are three different approaches to multi - dialect support. EN wiki does not dictate one version of English, but defaults to the creator of each article. This is becoming less satisfactory over time as many of the article creators have gone, and it leads to situations such as a Featured article writer expanding a stub in a dialect other than theirs to FA status whilst trying to write in an unfamiliar dialect. The alternative route which I think you describe in Dutch is that one version of spelling is official for that project. I suspect the former solution loses us writers from the dominant dialect, and the latter solution loses writers from the rarer dialects. Moving to the third approach as presented in this proposal would require community agreement in each project where this was implemented, and then the appropriate technology, and probably some processing to identify dialect words. So to use some EN wiki examples, If you set your preferences to American English, then unless you were in edit mode colour, transport and pavement would simply display as color, transportation and sidewalk. Other words are more difficult because they have multiple meanings, but this could be resolved by tagging them, so {{ENvar|bonnet}} would display as hood to someone viewing Wikipedia in American English but otherwise bonnet would appear as bonnet. This would require a certain amount of semiautomated editing to achieve, but in future anyone changing stuff between different versions of English could simply be shown how to do so usefully rather than as at present, simply be told they are wrong to do so. If the differences between two dialects can be resolved by a combination of one simple table "all the words in this table display one way if you choose to display in one dialect and other ways in other settings" and one complex table "Words in this table only display in different dialects if they have been tagged in the article as a dialect word". Then the software should be able to do this - obviously not altering anything in quotations. Things might get complicated with words like Fag, which has two meanings in English and a third in American English. But it is still doable. WereSpielChequers 09:36, 16 July 2010 (UTC)[reply]
I don't understand the real problem, here. Italy has been unified only recently, language unification is some 40-50 years old, we have tens of millions of citizens who use their dialect, Veneto is the region with the most dialect-proud population, we have some parties which required some official recognition and teaching of dialects etc., and still they switch to standard Italian immediately when they have to talk with a non local or in official environment, and most people find that it's silly to use dialect for things like an encyclopedia (apart from some specific local topics), so we have lots of editors from Veneto and so on, while regional wiki are almost desert and populated by nostalgic non Italian users with some Italian ancestor (Italians who migrated long time ago to foreign countries didn't know Italian, but only their dialect). I fail to see how your situation could be worse (and how this proposal could help). --Nemo 13:35, 22 July 2010 (UTC)[reply]
If it isn't a problem for Italians then we don't need to do this for the Italian project, though if you have popular regionalist parties that want their regional dialect supported then perhaps IT Wiki should do this. But it is a problem/opportunity in several languages, and the only explanation I find credible for our relative lack of success in the US as opposed to Britain and Australia. WereSpielChequers 13:14, 28 July 2010 (UTC)[reply]