Re: [Gnash-dev] Gnash, Flash, Adobe, and cash

[John Gilmore]

> We are simply willing to pay to get Flash 9 playback in our product,
> if this ends up being within our budget.

This is the classic "free software support / development" business
contract, as practiced for many years by Cygnus Support (which I
founded, and where Rob was a key employee).  The key to making it work
is for both parties to understand what exact specs you require the
free product to meet.  Then the developers can work up a price for
which they think they can do that work.  If that price is within your
budget, you conclude a contract; otherwise, you and they start
shifting the specs, to see if you can come up with a subset that you
can live with, which is also cheap enough to implement.

If your spec is "perfectly implements every possible Flash 9 movie"
then it'll be expensive -- multiple years of work by at least three or
four talented programmers.  If your spec is "plays the ten Flash 9
movies that we ourselves are authoring" then there's a lot more leeway
for adjustment -- small changes in those movies can probably make
large improvements in the delivery time and cheapness of the player.
If your spec is "plays Youtube and nine other Flash 9 sites of our
choosing", then not only is this a development contract, but it should
also include a post-delivery support contract (because those sites
sometimes change their flash movies, and you'll want Gnash to improve
to play those websites as they evolve, after your product's first
release).

> 1.   What is stopping the Gnash team from fully implementing the Flash 9 
> file format?  Where could we help the most?  ...
> someone taking the time to explain where the largest issues lie.  We 
> have some programming resources available, although we have no 
> experience with the Gnash codebase at all, as well as a potentially 
> large number of sample Flash movies that play correctly in the Adobe 
> player but not in Gnash.

One area that would speed up development is if you could make good,
small test cases.  If you find a flash movie that fails in gnash,
simplify it down to the essential part that fails, and report that.
Ideally that test case would be your original work under copyright
law, so you could assign it to the FSF and it could be added to
Gnash's regression test suite.  Then not only can a developer fix the
issue, and know when they've fixed it, but every future developer will
find out rapidly if they have broken it in making some later change.

> 2.   What kind of monetary investment would be necessary to 
> significantly speed up Gnash development?
> 
> I realize that this may be a difficult question to answer, but we are 
> quite serious.  We were prepared to pay Adobe to license their player, 

I am happy that Adobe keeps declining to serve significant parts of
the market; it makes life easier for the free software competition.  I'm
sorry it makes your life harder, though.

Most businesses don't want a "significant speedup"; they want a
"product in which these features work by this deadline".  If that's what
you really want, please give Rob an idea of which particular features,
and a vague idea of the deadline.

        John