Re: meta content-language

Hi Martin,

On Aug 22, 2008, at 5:16 AM, Martin Duerst wrote:

> The original idea of the http-equiv meta data was to provide data
> to the server to put into the HTTP headers. So even if everything
> went according to plan, it wouldn't be surprising, because
> Richard is testing the browser side.
>
> It turned out that parsing the file itself on the server side
> was too slow when you just want to throw out pages, so this
> never caught on. [except for the cases that Roy mentions]

I would take only a slight issue with your comment, that is an  
important one. Parsing the file itself on the server side did not  
catch on for some time, but with recent developments that is starting  
to change. Technologies such as Apple's Spotlight metadata  
architecture solve the problem of 'too slow' by indexing files in  
advance and reindexing upon any modification. The same approach is  
being applied to HTTP in CalDAV solutions. It is very likely that this  
approach will spread to traditional/basic HTTP services too. So I  
think it is even more important now that we don't decouple the meta  
http-equiv pragmas from their http equivalents by redefining what  
those http-equiv pragmas mean in HTML5.

Take care,
Rob

Received on Friday, 22 August 2008 11:07:28 UTC