Fw: RE: SVG Tiny 1.2 comparisons of IRIs for resource documents

For ACTION-2076 I was to ask the BitFlash guys about post-redirect
comparison of IRIs for resource documents.  Turns out the BitFlash
player doesn’t implement this and just compares the pre-redirect IRI, so
we should be safe in changing this.

Their replies forwarded with permission, below.


----- Forwarded message from Stephane Heintz <sheintz@opentext.com> -----

From: Stephane Heintz <sheintz@opentext.com>
Date: Fri, 27 Jun 2008 09:36:43 -0400
To: 'Cameron McCormack' <cam@mcc.id.au>
Cc: Andrew Emmons <andrew.emmons@quickoffice.com>,
 Lee Martineau <lee.martineau@quickoffice.com>
Subject: RE: SVG Tiny 1.2 comparisons of IRIs for resource documents

Hi Cameron, 

No, I am not aware of any post-redirect comparison in our player. The
internals of our SVG engine make no assumption of 2 documents with
different IRIS being potentially identical due to redirection, so they
would be treated as truly different documents. I'd like to know what
leads that "someone" to believe otherwise...

We would also lean towards removing the requirement and make comparisons
simply look at the absolue IRI. It is simnpler and much less confusing.

Cheers !
Steph

PS. As we are transitioning from OpenText to Quickoffice, it is best to
not use our Opentext emaid addresses anymore.




-----Original Message-----
From: Cameron McCormack [mailto:cam@mcc.id.au] 
Sent: Friday, June 27, 2008 12:06 AM
To: sheintz@opentext.com
Cc: aemmons@opentext.com
Subject: SVG Tiny 1.2 comparisons of IRIs for resource documents

Hi Stephane.

Since Andrew is away at the moment (congratualations, Andrew!) and the
WG has an issue that we need to discuss BitFlash's behaviour for, I
thought I would mail you. Please pass this on to whoever's the most
appropriate person to ask, if you could.

We have ISSUE-2003 (http://www.w3.org/Graphics/SVG/WG/track/issues/2003)
which is about the requirement of SVG Tiny 1.2 to consider two IRIs to
be equivalent by comparing their absolute, post-redirect values. For
example if you had:

  <!-- test.svg -->
  <svg xmlns='http://www.w3.org/2000/svg'>
    <use xlink:href='a.svg#x'/>
    <use xlink:href='b.svg#y'/>
  </svg>

  <!-- a.svg -->
  <svg xmlns='http://www.w3.org/2000/svg'>
    <g id='x'/>
    <g id='y'/>
  </svg>

and if b.svg redirected to a.svg, then the 'a.svg#x' and 'b.svg#y'
references would be considered to be to the same document, and thus it
is loaded as a resource document only once.

Rob O'Callahan from Mozilla raised the issue to say that there's not
much of a use case for looking at post-redirect IRIs to determine if
it's the same or a separate resource document and that in fact, if you
had the same IRI referenced multiple times in your main document, that
you would have to fetch it from the network (or at least HEAD it) to
determine if it gives you a redirect back each time.

The WG members who have discussed this are leaning towards removing
this requirement, and making comparisons just look at the absolute IRI.
Someone pointed out though that the BitFlash player already implements
this post-redirect IRI comparison behaviour, so we would like your
opinion on the matter.

Thanks,

Cameron

--
Cameron McCormack ?  http://mcc.id.au/

----- End forwarded message -----

----- Forwarded message from Lee Martineau <lee.martineau@quickoffice.com> -----

From: Lee Martineau <lee.martineau@quickoffice.com>
Date: Fri, 27 Jun 2008 08:45:25 -0500
To: Stephane Heintz <sheintz@opentext.com>,
 'Cameron McCormack' <cam@mcc.id.au>
CC: Andrew Emmons <andrew.emmons@quickoffice.com>
Subject: RE: SVG Tiny 1.2 comparisons of IRIs for resource documents

Hi Cameron,

Like Stephane says, we do not attempt to do any optimization in this area.

>From the SVG meeting minutes, you said "... bitflash does the double
network access already". I don't have the context of your previous
discussions, what were you referring to here?

--
Lee

----- End forwarded message -----

-- 
Cameron McCormack ≝ http://mcc.id.au/

Received on Monday, 30 June 2008 11:37:26 UTC