[Haskell-cafe] CEF3 bindings and game dev in Haskell
jo at durchholz.org
Fri Mar 10 12:54:05 UTC 2017
Am 10.03.2017 um 13:05 schrieb MarLinn:
> Simplifying massively, a browser engine seems to offer four things a GUI
> can use: A set of parsers, a complex model of the interface layout, a
> set of tools to render said model, and a system to get a stream of
> meaningful user events related to the model. There's more but let's
> pretend that's all that's needed. Let's further say you had implemented
> all the bindings and tools you're imagining. Would it be possible to
> replace one of the four parts with a pure Haskell solution without
> changing the others?
Not for the usual browser engines, not easily, and not very usefully
either because of interdependencies.
I.e. any Haskell solution that replaced just one part would have its
design massively constrained by having to server all the other parts.
to establish whatever GUI model you want to see implemented (vanilla
establish a smooth communication mechanism between backend and frontend,
and write a library on the backend side that drives the whole thing.
It's tremendously complicated, but it's well-isolated because the
program sees only a socket interface through which it drives the GUI and
receives events - which is pretty similar to any existing GUI toolkit
where you have an outgoing interface (function calls) and a stream of
I wouldn't have recommended doing such a thing a decade ago, but this
approach has become pretty well-understood so you can expect to see
solutions to any architectural decisions, though maybe a tad "too
imperative" for the Haskellish mind so you may want to design the socket
interface to make it easier to deal with on the Haskell side.
People may not be doing it because they do not have enough FPL
experience, so there's enough interesting stuff to do on that side.
More information about the Haskell-Cafe