On Tue, 28 May 2013 03:15:31 +1000
john skaller
<skal...@users.sourceforge.net> wrote:
> The desire is not surprising but the ability limited.
Heh...well put together. ;)
> I think this is something you can pick up with some patience.
It's true, but the GUI issue remains, iow. there are gtk2hs and wxhaskell.
Former simply sucks on non-Linux and although I like wxhaskell, seeing how long
they can't release 3.0 (it looks the project is almost one man band these days),
I prefer to try with Qt, although I'm not KDE user (using mostly GTK apps and i3
wm).
> The issue with Haskell I think is that there are no shortcuts to
> performance. You have to understand the compiler, and even then some
> things like mutable arrays, the cornerstone of performance, simply
> aren't available in a purely functional paradigm.
That's very true. Elegancy of the language is lost somewhere on the real-world
performance road.
> Really? D has been around a while.
D's minor releases break the code which is not pretty.
> My problem with it is that you just don't get enough back for giving up the
> primary advantages of C++: popularity and Standardisation: D is good but not
> good enough. IMHO of course.
And it seems it's getting bigger for not much gain, as you said.
> But Ada is standardised, and there are very high performance
> compilers. In fact my tests show Ada often beats C. The problem with
> Ada is that it is too old. It lacks modern polymorphism etc.
Ada 2012 is a bit more modern, isn't it?
> Nimrod looks good in principle, but the implementation doesn't appear
> to live up to the design yet. Eg the GC doesn't actually work.
I like(d) its syntax and it seems to be not too big/complicated while still
having decent performance...
> Lol. Felix is WAY ahead of Nimrod. Its been around for 15 years at
> least. And the Felix GC actually works. Although it isn't fast. And I
> have no idea how Nimrod devs got any ideas about Felix since to learn
> about it you have to join the mailing list -- and I'd notice. Good
> docs are one thing Felix does NOT have.
15 years? That's a lot.
How is it that the language, afaict, haven't caught more users?
> It's been working and useful for years. However based on the number
> of users its hard to claim robust testing which only a significant
> user base can bring.
Any rough estimation about the userbase and are you the only core dev?
> Felix can bind to C and C++ libraries. Executable glue logic is
> usually not required, however you DO have to provide "type glue".
I see.
> However if I recal Qt uses an archaic extension to an archaic version
> of C++ and so requires a special preprocessor. Felix could probably
> be modified to handle this "as if" it were an alternate compiler.
Hmmm...
> I am not going to write a Qt binding, or a binding for ANY large
> library. However I will support any such effort, modifying the
> compiler if required to ensure it can be done relatively easily.
Are there *any* GUI bindings available for Felix?
When it comes to C++, there are several choices like {qt, wx, fox, fltk...}
while the C world practically offers only gtk+ & EFL.
What would you do for GUI & Felix?
I hope that web & HTML5 is not the future of desktop apps...Just today saw
announcement about new Haxe release which generates JS, Flash, PHP...
> Not any more.
Uhh...it is a recent change?
Does it mean one has to do it manually now?
> I am phasing out fbuild. At this time, you need fbuild to build Felix
> first, then you use the result to rebuild it using Felix.
OK.
>
> *However* I suspect that's not what you're asking. If you're asking
> about building Felix programs the answer is simple.
>
> You don't. Its all completely automatic. Just execute the sources.
> Felix compiles stuff (with dependency checking and caching behind
> the scenes. Just pretend its like Python or Perl only it compiles down
> to machine code instead of byte code.
That's cool.
> As to end user docs: you can either write a document processor
> to do the job you want, or you can use the webserver and make
> a plugin that does what you need. The webserver generates HTML
> for various formats "on the fly". No separate document generation run
> is required.
Frankly speaking, generating docs is the least of the problem.
So, binding some C++ GUI lib is doable in Felix, but the problem is that I fight
to find enough time to work on the (open-source) project itself, what to speak
about improving language's ecosystem.
It also means that if we eliminate languages like Cobra/Haxe...it looks that the
Ada is giving us the most having decent compiler, fair size of community and
maintained GUI bindings...
Of course, I'll explore further...
Thank you for your input.
Sincerely,
Gour
--
As the ignorant perform their duties with attachment to results,
the learned may similarly act, but without attachment, for the
sake of leading people on the right path.
http://www.atmarama.net | Hlapicina (Croatia) | GPG: 52B5C810
------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
_______________________________________________
Felix-language mailing list
Felix-language@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/felix-language
