Position statement for W3C i18n Workshop - François Yergeau

My experience in the domain of the workshop is quite extensive and goes back to 1995:

I agree with the idea of orienting future i18n work towards the creation of guidelines, although not exclusively. Of the three basic areas of Web technology (formats, protocols, URLs), the first two are now very advanced with respect to i18n, but the last one is still in a very sorry state. Although strictly speaking within the purview of the IETF, URLs are central to the Web architecture and W3C should play a more active role in their development, most notably in correct i18n.


One area that would, IMHO, benefit from guidelines is design of multilingual sites. Such guidelines could cover, among other things:

Another area that would benefit from guidelines is Localizability of Web formats and applications. The Golden Rule in i18n is to separate code from (localizable) data and make the data easily localizable. Most Web formats are very poor in this respect. Localizing ASP/JSP or HTML cum Javascript is a major pain; it only gets worse when databases and other data sources are involved. Apart from separating code and content, other aspects of localization-enabling Web formats should be adressed. Such guidelines would be targeted to specifications (W3C or other), although not exclusively; Charmod is not the end of the story as far as guidelines for specs are concerned.


Although it has had an i18n activity for a number of years, W3C has not been examplary in its own behaviour, on its own site. The basics are of course covered (charset identification, etc.) but it has taken very long for e.g. the validators to go beyond Latin-1 and, more importantly perhaps, the W3C is not multilingual at all. The W3C does pay lip service to multilingualism by explicitly encouraging translation of its specs (and other pages, with permission), but at the same time discourages it by refusing to host translations. In effect, the W3C is telling the world, and in particular its volunteer translators, to please make the effort of translating, but we don't really care about your work nor about disseminating our specs in any language but English. The result is that the (numerous) translations remain on flaky servers at volatile URLs and that access to them is not reasonable easy for someone who doesn't know English (see for instance http://www.w3.org/XML/#trans where the list of translations is buried in the middle of a long English-only page and which lists the language names and spec titles in English only).

Of course the W3C cannot afford to translate all of its specs let alone the rest of the site! but it could go much further than it currently does in managing volunteer translations, as part of i18n outreach. Specifically, W3C should:


The WAI structure of Guidelines+Techniques seems appropriate: the guideline says what is desirable to achieve, the techniques (analogous to the IETF's BCP, Best Current Practice), say how to achieve it with current standards and technology.

I don't have any new ideas as to the organization of the i18n Activity in Wgs and IGs at this time. Too early to tell.