Re: HTML Validator HTTP Accept

Am 10.12.2008 um 18:35 schrieb Andreas Prilop:
>
> I do not agree as I do not have any problem validating these three  
> URLs
> with current version 0.8.4 of the validator:
> http://www.unics.uni-hannover.de/nhtcapri/ruby-annotation.x.html
> http://www.unics.uni-hannover.de/nhtcapri/ruby-annotation.xhtml
> http://www.unics.uni-hannover.de/nhtcapri/ruby-annotation.var
>
> Note that these three URLs point to one and the same document.
>
> It is not necessary for the validator to send an Accept or Accept- 
> Charset
> or Accept-Language HTTP header.

Andreas, please not again...!
In some cases, and more generally, it is very useful, if it sends  
reasonable Accept Headers. And you know that. Your examples concerning  
content negotiation, you point to again and again, are NOT reflecting  
the majority of what's out there in real web life. Real web life and  
real web needs aren't such easy as you tend to paint it. And you know  
that, because we discussed this topic several times and month ago.  
There IS, in fact, a need for reasonable accept headers (like every  
smart web client out there also provides -- the validator service is  
nothing else than a web client), even if you don't want to accept  
that. And it is very welcome, that the validator in the meantime  
provides such headers and informations that make it easier to  
cooperate and to communicate with thousands of webservers, that use  
several mechanisms to provide a most possible wide range of clients  
and to enable distinguishing between the several needs of the clients.  
That presumes, that the HTTP communication between client and  
webserver tells the webserver some useful information about the asking  
client, what it accepts. "Accept all" isn't a useful information, as  
one can see with Internet Explorer, this particular browser tells "I  
accept all", but isn't able to keep its promise, as we all know.


Sierk

Received on Wednesday, 10 December 2008 18:29:23 UTC