W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 2005

Re: XML InfoSet and property value preservation

From: Lisa Dusseault <lisa@osafoundation.org>
Date: Tue, 15 Nov 2005 10:02:38 -0800
Message-Id: <08f24d13d07d293695cdd6f0e9d84a42@osafoundation.org>
Cc: WebDav WG <w3c-dist-auth@w3.org>
To: Julian Reschke <julian.reschke@gmx.de>

On Nov 15, 2005, at 12:01 AM, Julian Reschke wrote:

> Lisa Dusseault wrote:
>> In WebDAV implementations today, the only InfoSet piece that we know 
>> isn't always preserved on properties is the namespace prefix (and 
>> thus the namespace declaration attributes might also be rewritten).   
>> I'd
> I'm not sure how you can say that. I just tried with comments and PIs 
> in Apache/moddav, and they aren't preserved. Do you have any data 
> supporting your claim?

No -- I only said we don't *know* of other InfoSet items being lost, 
and now you've provided fresh data for this list.  If it's been said 
before, I'd either missed it or forgotten.

> I don't want to be lax. We should either document what we have today 
> (if we want to advance RFC2518), or document what we tjink would be 
> good (staying at proposed). Throwing in things like PIs, but keeping 
> out *essential* things like prefixes doesn't make any sense to me, 
> though.

Great; so what's a principled basis on which to make this decision?  I 
can see one extreme option, plus an "exclusionary" choice and an 
"inclusionary" choice based on what we know implementations do today.  
This isn't full text for each point but just illustrative:

Extreme option:  All InfoSet items MUST be preserved.  [This clearly 
has the disadvantage of making existing implementations change their 
code to comply, but has the advantage of simplicity and enforcing the 
greatest consistency between servers.]
Exclusionary option: All InfoSet items MUST be preserved, except for 
the ones we know aren't preserved in some server implementations 
(comments and prefixes... )  [This encourages nearly as great 
consistency as the extreme option.]
Inclusionary option: Of the InfoSet items, only the ones we know client 
implementations really need (text, element, attribute name and 
value...) MUST be preserved.  [The difference between this one and 
exclusionary option covers InfoSet items like Processing Instructions.  

Has anybody considered this issue besides myself and Julian?  Does 
anybody with an implementation wish to speak up and argue why it might 
be important to preserve or useful to discard various pieces of the 
InfoSet in property values?

Received on Tuesday, 15 November 2005 18:03:03 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:33 UTC