Re: svg textLength issue (and a little thing about use...)

Hi Jim,

On Fri, 16 Aug 2002, Jim Ley wrote:

> The textLength attribute on SVG's textLength causes accesibility problems
> for users who use their own stylesheets to overide author selections.
> Consider:
>    <text font-size="20px" textLength="38px"
> lengthAdjust="spacingAndGlyphs">Hello</text>
> Which will work fine if the font-size is close to the 20px, the
> adjustments whilst making the font slightly less clear shouldn't have too
> much effect on legibility.
> However if the user takes advantage of CSS2 to increase the font-size to
> something they find legible, the resulting narrow letters will likely be
> unreadable for them.  I do not see any mechanism for a user to override
> the textLength property in their user stylesheet, yet a user would need
> this ability to adjust font-size.
> If the lengthAdjust="spacing" then as a single character of the users
> font could easily be 38px, I'm not sure what's supposed to happen then,
> the spec. appears to be silent on the issue of what happens when the two
> calculatated values are radically different, but Batik stacks the letters
> on top of each other which seems reasonable.
> I do not have any ideas to propose as a solution unfortunately.

It seems like we need some way to override the textLength.
We'll discuss this and get back to you (www-svg).

> The other is a little thing about use:
> | When a 'use' references another element which is another
> | 'use' or whose content contains a 'use' element, then the
> | deep cloning approach described above is recursive.
> I believe it should be made explicit what should be done in the situation
> where a 'use' element, references a 'use' element which refers back to
> the original one. ( ) as reading
> it currently a conforming UA should keep on recursing forever... (and
> Batik does!)

We added wording to SVG 1.1 which describes that this in error.
Basically as soon as the UA notices a self-reference (or reference
loop) it goes into error - stops rendering.

I think we should also publish an errata for this issue in
SVG 1.0.


Received on Friday, 16 August 2002 07:01:47 UTC