Ray Saintonge wrote:

Yann Forget wrote:

...

Hi, I think I found some errors in the Dutch Wiktionary (wrong language):

• http://nl.wiktionary.org/wiki/Catalee (not Marathi)
• http://nl.wiktionary.org/wiki/%D0%9C%D1%83%D1%85%D1%88%D0%B8 (not Tamil)
• http://nl.wiktionary.org/wiki/Catal_Ke (not Bihari)
• http://nl.wiktionary.org/wiki/Denmark (not Malayalam)
• http://nl.wiktionary.org/wiki/air (not Malayalam)
• http://nl.wiktionary.org/wiki/Ingraji (not Marathi)

This seems like an awfully difficult problem to police. We are basically at the mercy of the person who puts up this kind of material. Although I'm very active on the English Wiktionary I seldom touch the translations except for matters affecting formatting. This is because I have absolutely no idea whether most of these are correct. Often these translations are imported from some other source without any attempt at fact checking whatsoever; some people seem to take pride in having the longest possible lists of translations. Having persistently dead wrong information can affect credibility even more than rampant POVs. At least with a POV issue there is a recognition that there is more than one way to look at a subject.

A possible solution might be to forbid the inclusion of translations unless

1. There is an active Wiktionary in that language, or
2. A source has been cited for a particular translation.

Are there any other fairly easy ways for us to fact-check this kind of information?

Ec

The problem here is not that we need to POLICE it, this is a situation where we meet our strengths and our weaknesses. The model that we use is one of collaboration; the more people work on the same data the better the quality. Our problem is that to a large extend we do not work together as much as we could; we all have our own little wiktionaries and we hardly notice what happens outside our view. This increases the likelyhood of errors. The proposal to come to one integrated wiktionary is a way of solving many of these problems.

I am happy with the increased cooperation between many wiktionaries, I see it as one of the prerequisites of the maturation of the Wiktionary.

The idea of forbidding translations is contrary to what wiktionary aims to be. With Wiktionary content it is simple; it is correct or it needs improvement. The idea that errors are persistent is wrong as the idea is to correct any and all errors. The conclusion should be to increase any means that will improve cooperation and not restrict the inclusion of new data that is believed to be correct.

Thanks, GerardM

--- Ray Saintonge saintonge@telus.net wrote:

...

Yann Forget wrote:

...

Hi, I think I found some errors in the Dutch Wiktionary

.......

...

This seems like an awfully difficult problem to police. We are basically at the mercy of the person who puts up this kind of material.

A possible solution might be to forbid the inclusion of translations unless

1. There is an active Wiktionary in that

language, or 2. A source has been cited for a particular translation. Are there any other fairly easy ways for us to fact-check this kind of information?

Well there is always the talk page for each word. One of the benefits of a distributed project is that everybody can have their say and wandering experts and native speakers are always welcome.

The real problems are very different instead: 1) co-operation 2) time 3) computer / internet literacy

1) co-operation Many people don't "think global and act local" - they just would like to see their name at top and that's all - whenever they have to contribute it is hard to motivate them. Out of approx. 1000 translators I tried to interest only three would contribute and not even one of those three online. There are many, many small wiktionaries and only a small group is now interacting (and this is great :-)

2) time Not everyone of us has enough time to do everything he/she would like to do. Many would like to do more and many do a lot on single wiktionaries - the same work done could be quite easily used for all wiktionaries --> database solution

3) computer / internet literacy Like I said above: the few people that write me and who would like to contribute have one huge problem: most of them are not used to wikis - they send me glossaries to be included (and these glossaries often represent years of work for a translator) - on one hand they just don't have the time to restart from the beginning uploading one word after the other (we have one of approx. 7000 checked translations) and on the other: they just don't know how to do this. Many of my colleagues use the computer just like a typewriter. As soon as they have to work with something else but Word they indeed have difficulties. So the language experts are there, but they are not willing/not able/don't have time to care about online work. I know from others I talked to on the phone that they just don't offer their help as they belive not to be able to do anything online or that their work (very valuable glossaries made by specialists) are not of interest.

I have been thinking about these themes for some weeks now, discussed with colleagues, discussed with my former business partner (who is always full of ideas and checks my Italian writings for wiktionary) - the only real solution would be: a human interface between many language specialists and wiktionary, but not every single wiktionary - a global one - so every contribution would have immediate effect and only the UI would be different. Yes, I know, some don't like this idea, but witktionary is very different to wikipedia, wikisource, wikibooks etc - it uses data that is interchangeable if prepared wisely.

There are many potential contributors that could and would help with their expertise, but they need an interface (a human one), someone to talk to and we should find the way to include their expertise, the linguist one.

I'll write a second mail that centers on a small project (that hopefully will become very, very huge) we are trying to achieve and includes some other basic questions. Even this one depends on proof reading, new entries etc. I am just coping with the above difficulties within this project and step by step it becomes very clear what we need.

Ciao,

Sabine