W3C home > Mailing lists > Public > www-svg@w3.org > June 2006

RE: SVGT 1.2: SVG fragment identifiers

From: Doug Schepers <doug.schepers@vectoreal.com>
Date: Tue, 27 Jun 2006 20:17:24 -0400
To: "'Bjoern Hoehrmann'" <derhoermi@gmx.net>
Cc: <www-svg@w3.org>
Message-Id: <20060628001725.7CE325621@postalmail-a3.dreamhost.com>

Hi, Bjoern-

Bjoern Hoehrmann wrote:
| * Doug Schepers wrote:
| >The SVG WG agrees that this behavior is not what was intended, and is
| >inconsistent with regards to the behavior of the fragment 
| identifier in
| >HTML. This is due to an omission in SVG 1.0, and we have 
| restored the intent
| >of the behavior.
| >
| >The specification now indicates that when a bare name 
| fragment identifier
| >indicates an existing element, that element will be centered in the
| >viewport. The exact wording will be available in the next 
| public draft. If
| >this behavior or the specific wording does not satisfy your 
| concern, please
| >respond promptly.
| I certainly object to the Working Group's decision to 
| disguise this new feature as correction of an "omission" in SVG 1.0.

As has been explained to me by members who were in the WG at the time, this
was the original intention of the behavior of fragment identifiers.  I'm
sorry that you don't approve of the change, but we felt that it was
important to address this in SVG Tiny 1.2, so that the behavior going
forward would be as authors would expect.

| As I pointed out in
| http://lists.w3.org/Archives/Public/www-svg/2006Jan/0526.html 
| the issue raised here is the result of thoughtless search and replace;

This fix is not intended to address that particular comment, although it
does obviate it.  Instead, it is meant to address several comments over the
course of years, including some that I made before joining the Working
Group, and specifically to address Maciej's concern that the behavior was
poorly specified.  

| what you propose is incompatible with SVG 1.1,
| which specified different but perfectly sensible
| behavior that is at least as consistent with "HTML".

This was discussed as a concern.  Since the behavior defined in the SVG 1.1
specification was ineffective and unimplemented, and thus has no content
relying on that behavior, it was judged that the trade-off with
incompatibility was neglible when considering the gains for authors.
However, we do intend to issue an errata for SVG 1.1 in this regard.

Speaking personally as an author, I think that this is a big win, since now
we can have sensible behavior for fragment identifiers, which has been a
frequent request.


www.vectoreal.com ...for scalable solutions.
Received on Wednesday, 28 June 2006 00:17:31 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 8 March 2017 09:47:08 UTC