ODP organisational question

I have argued with ODP people in the past about one thing that I feel is tremendously short-sighted on the part of the ODP organisation, and I understand from others here that this forum is unlikely to make any difference to these sorts of issues. However I'd like to get an understanding of the reaso9ns it is organised as it is in one crucial area: that of language *then* topic. To me, this simply makes no sense whatsoever. I know the vast majority of US citizens probably speak only English, and that the vast majority of ODP contributors and editors are probably US citizens (based on internet demographics alone), but to the rest of the 5.86 billion people on the planet, speaking more than one language is an ability likely to be found in a large proportion of the people.

Why not have an interface language choice, and then translations for each category into appropriate number of languages, links and descriptions shown in as many languages as they have been entered in, and each resource given a number of language options to allow the user to see what languages are available at specific urls? With tools such as Babelfish, sites that are in Japanese are decipherable to an English speaker, and certainly worth the time to decipher when their content is excellent, albeit in a language not natively spoken by the user.

By splitting the resources on language first, it seems that the ODP has forked itself n ways (n being the number of languages as categories) right from the get-go.

Cheers

I think this is an issue of legacy. When the ODP was started 5 1/2 years ago there were very few non-English sites around. Now over 1/4 of the total sites we have listed are not in English.

We do have links at the bottom of each category page which lets you view the same category in another language, should there be sufficient sites to warrant an equivalent category in that language.

That's not a great solution though as the category has to exist in the other languages before the link will show. The best resource in the world on a certain topic may also be in a language other than English - why deny it a listing in the main (read: English) portion of the directory when it is (hypothetically) better than all the English sites even when viewed through the 'fish?

I dunno, it just seems kinda broken to do it this way around. In a perfect world (ok mostly perfect - not everyone is fully multilingual ;-) you'd possibly have language preferences and the urls arranged according to that or some other criteria. In other words - surely the topic/category of the site and content should be of greater import than the language chosen to convey it in?

Thanks - I'll stop now

Cheers

I think that as totalXSive mentioned it is a legacy issue. It was built this way to suit the situation at the time, but that doesn't mean that it won't change in the future. There are a lot of things that we would like to see changed, but something like this would probably be part of a major overhaul if the directory ever gets rebuilt. We have ongoing discussions internally about these things.

zak, I agree, and I like your suggestion a LOT - I speak two languages well enough to be able to read web sites, and I would love the idea of being able to say that, and see the amalgamated listings in those two languages.

It might be something that is considered in future.

Thanks for your suggestion.

Actually, what you are suggesting sounds like a better task for a downstream data user - combining content of equivalent categories and providing translation of all pages into any language.

The ODP's main focus is on listing unique content. A site which offers content on a topic in multiple languages has unique content for each language. A site in one language which uses an automatic translator to other languages still only offers unique content in one language. As you've probably noticed, automatic translation leaves a lot to be desired in many areas, and often doesn't even convey the correct basic meaning of what is translated, much less important nuances.

We are focussed on classifying websites, not presenting information to users. Downstream data users can identify target audiences and modify the presentation far better and more effectively than we can. That's why the arrangement is set up the way it is.

By the way, English is the primary language of several major countries other than the United_States. Another common misperception is that ODP "Main" = United States, which is not true.

While I can see some benefits to the general idea of your suggestion, the idea of building a category structure to make use of really bad translation tools like babelfish is not a good idea. There are some really fabulous sites that I know of that are the definitive sites for their topics. They're not in English, though, and using a machine translator renders the sites virtually unreadable. That makes them useless for listing outside of the language they're written in.

There are also some small problems that I guarantee would develop into large problems if we tried to combine the ontologies. There is more than just language that separates the World categories. How things are categorized are often different from one language to another.

What would we do when predominantly English speakers insist a certain topic fits in one spot, and predominantly French speakers insist it should be somewhere else?* Even the differences between the US and UK, both English speaking though there are a few differences between the two Englishes, are quite large.

(* Note: I used English and French only as examples, there are quite a few very active language categories at ODP)

More than once I've fixed links between languages, using babelfish to be certain that I was linking correctly, only to be informed that the subtle differences in the categories weren't apparent from a babelfish translation--which is still sorely lacking in most languages.

I'm not saying that there isn't be a better way to do this--though there actually wasn't a better way to do it before UTF-8; and while conversion to UTF-8 is underway at the moment, there's still a way to go before the character sets between languages will be compatible--I'm only saying that there are a lot of problems that would not immediately be evident in doing it as proposed here.

Creating an ontology by which to classify such widely varying sites, languages and cultures is far more complicated than any classification project I've seen proposed in the past. Personally, I applaud the fore-site of the ODP creators in realizing that the effort should be International, despite the fact the the Internet at the time of ODP's conception was English dominated, and the staff totally US oriented.

I think that it is convenient from an editing and user perspective that non-English sites have their own "mini directories". You could, for example, take the World/Deutsch category and have a self-contained directory of all German language sites covering every topic with 320,000 sites.

Let's look at this another way though, consider Punjabi, a language with perhaps 80 million speakers in south Asia, but with only 46 listed sites. It only makes sense to list those sites together, because they would be completely lost if listed under topics - 46 sites in millions of listings - and from a visitor perspective they would be worthless.

There are certainly advantages in what you suggest, but there are disadvantages too and even if it was determined that there would be some sort of benefit to the visitor (and I don't personally believe it would) then I think the amount of work to be done would be prohibitive and that editors and staff have much more important things to do.