Jon,
yes, i remember you ;)
> Do you have a list of remaining features that you'd like to implement?
here's something i wrote a while ago as some kind of a plan to someone that
expressed interested in a sponsorship (which didnt turn out):
"I surely would like to see it getting more community-backed than it is now, i
am unsure though how to proceed. As of lately, some people have been looking
into the code to solve some of their own problems, and that is of course
something i appreciate very much. Up until now i have been able to satisfy most
ideas/wishes that i received, and i of course hope it continues that way.
We also seem to share the long-term idea of getting somewhere near flex. My
current thoughts mostly circle around the build process, i'd like to see flash
development go somewhere closer to the "traditional" compiler/linker,
static/shared lib models of C/C++ development. I'm also spending some time to
get the various component sets properly integrated. That direction is probably
the most exciting of all, it should be possible to get swfmill to support a
"declarative layout" model like MXML or XUL with existing UI things like SMX or
enFlash. There's a bunch of open issues there though.
Another track that I'd like to follow is general optimization/stabilization of
the core features. swfmill seems pretty successful as the small, focussed tool
it is now-- import functionalities based on XML/SWF conversion. I have my
doubts however about some of the current fundamental architecture and would
like to redesign the XML/SWF conversion in direction of an event-based
("streaming") model, similar to SAX. That would lower memory consumption
significantly for most processing tasks, and remove some "unfixable" problems
with swfmill 0.2.x, like importing hundreds of hi-res images or fonts like aral
unicode.
As for stability, there are two kinds of test suites that could/should be
developed-- one "classical" for the core features, and another one for the
various SWF tags, something like the SVG testsuite, separate small SWFs that
test all the features SWF supports. This would pretty much eliminate the
possibility of regressions. It's obviously a bunch of work though. However, it
could be a nice chance to integrate the community, in a way that i develop a
testing framework and pretty generic method for creating such SWFs, and
motivate people to contribute tests for the specific parts they're actually
interested in.
Finally, there is one thing i have been working on a lot lately, that is SVG
import. I didn't publically announce this yet, but the later prereleases are
actually able to convert most Inkscape-generated SVGs into SWF. I have some
problems still with the cubic->quadratic spline conversion, and CSS and
declarative Animation/Interaction features of SVG are left untouched. CSS would
indeed be wonderful i think, the animation/interaction part is much less
pressing. The most useful thing here is to be able to import vector graphics at
all, something that AFAIK is impossible now with open-source tools."
My wish to redesign the core structure has strengthened since then, so if i can
allocate some substantial time to it, that's how i would proceed.
But, the community and documentation thing is really of pristine importance.
swfmill can do a lot more than even most of its users know ;)
Also, I think a lot can be done with XSLT, producing swfml-lowlevel (the whole
swfml-simple is just a stylesheet :).
And, finally, the test cases would be great to have.
These things could all be done without touching any C++, even completing the
supported tags should mostly be a matter of writing some XML into
codegen/source.xml. But of course you're free to fiddle with C also.
-dan
_______________________________________________
swfmill mailing list
swfmill@osflash.org
http://osflash.org/mailman/listinfo/swfmill_osflash.org
