- From: Thomas DeWeese <Thomas.DeWeese@Kodak.com>
- Date: Tue, 25 Jan 2005 16:35:46 -0500
- To: Dean Jackson <dean@w3.org>
- CC: Cameron McCormack <cam-www-svg@aka.mcc.id.au>, www-svg@w3.org
Hi All, Cameron just raised this to my attention again. I don't think I actually followed what Dean had suggested the first time I read this. I might agree that the document is in error, I would disagree on the reason. I think it is a really bad idea to allow users to change the version attribute after the document has been loaded (or it should have no effect). If you want to enable this you would have to specify what implementations are expected to do with existing references to DOM objects (do they lose the SVG 1.2 methods? Try that with a strongly typed language ;). At the very least it would imply that the implementation would have to walk the DOM checking for elements or references that are now invalid (re-evaluate switches). This just strikes me as an almost impossible to implement feature with almost no reason for existing. Also given the issues with the Object model I have a hard time imagining any kind of consistent behavior across browsers/binding languages. Dean Jackson wrote: > No such thing as a stupid question. Only stupid people. > (I'm in that group) > > On 14 Jan 2005, at 15:13, Cameron McCormack wrote: > >> ...but what happens if you change the value of the version attribute on >> the document element? >> >> <svg xmlns="http://www.w3.org/2000/svg" version="1.2"> >> <solidColor solid-color="red"/> >> <rect width="100%" height="100%" >> onclick="document.documentElement.setAttributeNS(null, >> 'version', '1.1')"/> >> </svg> >> >> After clicking the rect, is the document in error? > > My thinking is that it is error. This is the same > as a script adding an element that doesn't exist, or > as a child of an element that can't contain it, or > doing something else that is equally bad. > > Do you agree? > >> >> Also, while I'm talking about solidColor, what is the purpose of the >> stop element as a child of solidColor? The schema says that it is >> allowed but I'm not sure why. > > > That's a mistake. It shouldn't be there. > > Dean > >
Received on Tuesday, 25 January 2005 21:35:49 UTC