Because gradients were not implemented in IE, I did have that one
feature in mind.

But the main thing that made me guess IE/VML might need the help of a
smaller feature set to support vector graphics reliably was just the
kinds of IE bugs (every single bug was an IE bug) I encountered when
using GWTCanvas. The most demoralizing, can't-I-even-rely-on-VML-to-do-
that, bugs:

http://code.google.com/p/google-web-toolkit-incubator/issues/detail?id=293
http://code.google.com/p/google-web-toolkit-incubator/issues/detail?id=281

I'd greatly prefer an API that was NOT bifurcated, though, and, after
reading Ray and your replies to **, I imagine that the scene-graph
approach (don't really know what that is, but presume an extra level
of abstraction is involved) could make it easier to cleanly avoid
(what I assumed were VML related) bugs like the ones above.

** =
http://groups.google.com/group/Google-Web-Toolkit-Contributors/browse_thread/thread/ce4dfae5c70a5651/30df5bca65a553c2?#30df5bca65a553c2


On Jul 17, 7:25 am, Joel Webber <[email protected]> wrote:
> RichTextArea was indeed designed the way it was for that reason. The
> difference between the basic and extended formatters in particular were
> needed for Safari2 support.
> As for canvas/svg/vml, do you have any particular features in mind? The
> difference in capability between svg and vml seems to be fairly small,
> though IIRC, vml is fairly week on gradients.
>
> On Thu, Jul 16, 2009 at 10:47 PM, John Gunther
> <[email protected]>wrote:
>
>
>
> > RichTextArea supports a basic and an extended feature set:
>
> >http://google-web-toolkit.googlecode.com/svn/javadoc/1.5/com/google/g...
>
> > I imagine this approach was taken because the RichText capability gap
> > was too large cross browser to feasibly emulate everything in the
> > worst browsers.
>
> > Would it make sense to take the same approach with GWTCanvas?  (The
> > vector graphics capability gap between IE6 and FF3 may also be too
> > large to be feasibly spanned via emulation alone).
>
> > For example, Client-side GChart has gained a lot of value from this
> > relatively small subset of GWTCanvas's features:
>
> >http://gchart.googlecode.com/svn/trunk/doc/com/googlecode/gchart/clie...
>
> > Many flowsheet simulation packages could also be comfortably
> > implemented via some such subset (plus the standard GWT Widgets, of
> > course). If such a bifurcation allowed GWTCanvas to graduate from the
> > incubator much faster, I think it would be worthwhile.
--~--~---------~--~----~------------~-------~--~----~
http://groups.google.com/group/Google-Web-Toolkit-Contributors
-~----------~----~----~----~------~----~------~--~---

Reply via email to