Re: Merging zh and en versions of clreq (urgent)

@r12a,

Thanks for the explanation. ;)

Sincerely,
Chen Yijun (@ethantw)


> On 2015年7月7日, at 18:51, Richard Ishida <ishida@w3.org> wrote:
> 
> On 07/07/2015 11:41, Yijun Chen wrote:
>> Agree to the proposal. But what’s wrong with the original `lang` attribute that we had to use `data-lang` instead?
> 
> hi Yijun,
> 
> yes, i thought about that, too.  I might be convinced otherwise, but my reasoning was that we need a key for a script to do the separation, and if the key is the lang attribute then it may be more difficult to tell which instances of the lang attribute should be used for splitting the text, and which are there to really indicate the language of the text.
> 
> in the end, i thought it safer to use a dedicated attribute. I also thought that the final documents would be better if they didn't have lang attributes on every p, hx, etc, but again removing just those that aren't needed and leaving those that are may not be straightforward.
> 
> 
> ri
> 
> 
> 
> PS: fwiw, jlreq used lang attributes, but each language string appeared within an extra element with a class and id that grouped the pair together. That sounded like far too much effort to create and maintain.  Also the method of dealing with inline stuff with ids, such as dfn, required a lot of indirection and scripting.  I'm aiming to keep the extra markup and effort minimal for clreq.
>