Re: Review of Content-Encoding: value token

Mark Nottingham wrote:
> My concern is that many implementations treat content-coding like 
> transfer-codings, in some ways; that is, they layer it in automatically 
> (e.g., mod_gzip). The confluence of this with things like byteranges, 
> etag comparison, etc. may be quite prone to bugs.
> 
> You'd also start to get into discussions like "can't JPEG just be a 
> content-coding of GIF?" and so forth. The media type system isn't 
> perfect by any means, but I think it's a better fit for this type of thing.
>...

I agree with that, but right now the spec doesn't really state that.

Should we change it?

BR, Julian

Received on Friday, 30 January 2009 08:30:09 UTC