Re: Validator Multiligual Support

Hello Kenny,

On 15-Oct-08, at 9:35 PM, kennyluck@csail.mit.edu wrote:
> I came from Taiwan, and I wonder whether the validator team is  
> planning to do some multiligual support. Ideally, it's nice if  
> validator.w3.org returns Chinese when the client is Chinese and  
> sends a 'Accept-Language: zh-tw'. I am able to hack the validator  
> code and make a (Traditional) Chinese version, but do you mind  
> adding the language content negotiation to the server side  
> (validator.w3.org)? That'all be very helpful because language is  
> always a barrier to standards.

The groundwork for localization has been done on the markup validator  
a very long time ago and having multilingual UI and results has been a  
plan for very long indeed. So why hasn't this happened?

1) lack of coding power. There has been a lot of great offers to  
translate the validator but what we would need first is someone with  
perl proficiency to finish the coding groundwork and make the  
localization actually work. At this point the coding power for the  
markup validator is really low, and all resources have been put into  
fixing bugs, updating support for document types, etc.

2) multiple text sources. The localization of a single piece of  
software is HARD. My experience from the maintenance of the  
multilingual CSS validator is that in spite of the great help provided  
by translators, getting translations updated or added in 10+ languages  
every time a piece of code changes is a source of headaches, and  
translations always constitute the worst bottleneck for releases. For  
the markup validator, the conditions are even more complicated, since  
the validation messages come from at least 4 sources:
* the markup validator code itself, for UI, a number of warnings/ 
errors, and most error explanations
* opensp, the DTD validator, for most of the error messages
* libxml2, the XML parser, for XML parsing errors
* validator.nu for HTML5
Some of these (opensp and IIRC libxml2) have some translations  
available, but getting them to work in the validator has been tricky  
so far. Making sure that all the components are translated and that  
the translations work in their interaction with the validator would be  
feasible, but difficult.

All this considered, I still think, as you do, that making the markup  
validator available in more languages than english would be a major,  
major improvement for its usability by web developers and designers  
worldwide. As outlined above, there are no showstopper, but a few  
hurdles that have kept us from making this happened. Nothing, however,  
that can't be overcome if this open source project can gather:
- a translation coordinator
- some (perl) dev power
- ... the translation help will follow

Would you, or anyone on this list, be interested in filling either role?

Thanks!
-- 
olivier

Received on Friday, 17 October 2008 16:07:35 UTC