[reactive] Using Reactive in a commercial video game - thoughts?

[Conal Elliott]

Thanks, Bob, for the summary.  Reactive has appeared to be tauntingly close
to working for about a year now, and it's still unclear what it will take.
Debugging/diagnosing has been difficult, termination/laziness issues have
been subtle, and in some cases unamb turned up ghc rts bugs.

And, yes, I'm exploring licensing alternatives.  None of the standard
licenses capture what I want.  I switched to GPL-dual as an invitation to
dialog and collaboratively explore possible arrangements.  I'm not attached
to keeping it that way, as the coercive aspects of GPL leave a bad taste in
my mouth.  I'm looking for a sustainable model for myself and others to
create and fun & useful software and share it with each other.  Reactive is
a handy test case for the exploration.

  - Conal

On Tue, Jun 16, 2009 at 12:19 AM, Thomas Davie <tom.davie at gmail.com> wrote:

>
> On 16 Jun 2009, at 08:45, Ryan Trinkle wrote:
>
> I'm a co-founder of a game studio that will be writing in (primarily)
> Haskell.  We're currently assembling the technology stack for our first
> game, and I'm very interested in FRP (Reactive in particular) as a way of
> making video game programming easier and less error-prone.  However, from
> reading this list and discussing it with people in #haskell on
> irc.freenode.net, it seems that there may be issues blocking us from using
> it within our timeframe (we plan to release our first game around November).
>
> What known issues would prevent Reactive from being used today in a
> commercial video game?  Would you anticipate the discovery of many critical
> issues in the future?  What would it take to overcome these issues?
>
>
> At the moment there are a couple (maybe more) of issues
> with reactive that make it basically unusable for this kind of project:
> 1) The way that unambiguous choice works is still not settled, every time
> someone thinks they have it right, some other little sneaky git of a detail
> comes up and shows it to be incorrect.  There's been some progress recently
> on getting this right.
> 2) The snapshot, mappend and join functions on events currently block in
> inconvenient ways with certain inputs.  This doesn't sound like a major
> problem, until you discover it's the *useful* inputs that they block on.
>  This issue is just a matter of getting lazyness right, unfortunately,
> various people have tried to get the lazyness right over the past year or
> so, and no one has yet succeeded.  If you were to get it right, I would be
> extremely happy.
>
> A final note:
> From past conversations with conal, I believe he's in the process of
> migrating reactive to a different license because he wants to make sure that
> while it remains open he still gets paid for its use in industry.  You'll
> have to check this with conal though.
>
> Bob
>
> _______________________________________________
> Reactive mailing list
> Reactive at haskell.org
> http://www.haskell.org/mailman/listinfo/reactive
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.haskell.org/pipermail/reactive/attachments/20090616/921d5b4f/attachment.html