- From: David Morris <dwm@xpasc.com>
- Date: Thu, 9 Aug 2007 13:08:29 -0700 (PDT)
- cc: <ietf-http-wg@w3.org>
On Thu, 9 Aug 2007, Jamie Lokier wrote: > > Paul Marquess wrote: > > Using the zlib wrapper means an implementation can quickly determine whether > > the content is what it claims to be by checking the zlib signature. With > > 1951, you have to try to uncompress it before you can tell. > > We have HTTP headers for that. If they aren't trustworthy - what > makes the zlib signature trustworthy? As part of a team that produces a value added proxy facility, I can tell you that HTTP headers are frequently wrong when it comes to labeling content, encoding, etc. Data streams are also corrupt, but much less often and if the stream starts with a known label, we don't need to crash test alternatives. Dave Morris
Received on Thursday, 9 August 2007 20:08:47 UTC