Re: On the ordering of header entries

But if:

0) Blocks are fundamentally unordered from a processing point of view.

then the de-signing code will need to verify the correct order
for processing anyway, right?
-Dug

Rich Salz <rsalz@zolera.com>@zolera.com on 05/21/2001 08:25:46 AM

Sent by:  rsalz@zolera.com


To:   Doug Davis/Raleigh/IBM@IBMUS
cc:   xml-dist-app@w3c.org
Subject:  Re: On the ordering of header entries



> Just wondering, what's the rational behind not allowing someone
> to reorder the headers?  Not that I think people will be itching
> to do it, I'm just curious.  8-)

(Some subset of) the headers might be signed.

I could imagine some intermediate actor "slurping up" the entire
package, and then outputting the headers in arbitrary order (say,
alphabetically by namespace URI).  And they shouldn't do that.
     /r$

Received on Monday, 21 May 2001 08:30:36 UTC