[A6] andromizer pc ports

[Chris Pickett]

Paul Evenblij wrote:

>--------- Original Message ---------
>DATE: Sun, 12 Oct 2003 00:32:48
>From: Chris Pickett <chris.pickett at mail.mcgill.ca>
>
>  
>
>>Hey Paul,
>>
>>Some folks on the A6 list are interested in Win32 and possibly GNU/Linux 
>>ports of your tool.
>>    
>>
>
>I know ;)
>
>This has come up before. Since I have added my voice to the chorus clamoring for an open-source release of the Andromeda OS, I suppose I should put my money where my mouth is, and release Andromizer's source code... :)
>  
>
well, that's good news! :)

>There are 2 issues here:
>
>1. Time. I can't release the code 'as is'. There's too much stuff in there that either needs to be removed, or is waiting for development into a new feature. And for me, this was also an exercise in multithreaded programming in C++ and Objective-C. This shows.
>Also, even though the code is documented, I owe it to my professional pride to polish it, so people will have an easier job of understanding it. 
>Currently I am much too busy to do that.
>  
>
well, imho you could release it "as is", and just say, "this is not a 
final, cleaned-up, polished source tree".  and then if it's too 
difficult for people to work on, you can just get to it when you feel 
like it.  i mean, it's not like 50000 lines of code or anything.  if you 
decide it's not worth the bother to clean everything up and port it to 
other platforms, and would rather rewrite it all, then at least the 
source is out there for others.

>2. Effectiveness. Keep in mind that a large part of this tool's usefulness stems from its user interface. C++ user interface code is NOT easily ported to other platforms. In fact, any Linux or Windows programmer would probably have to rewrite this entire part. The MIDI part, of course, depends on system libraries as well. In my opinion, the only stuff that could actually be ported with any degree of effectiveness is the randomizing routines. Luckily, I have kept these parts functionally separate as much as I thought feasible.
>
that was what i was thinking.  if you (or somebody else) rewrote the 
interface to use GTK-OSX (based on GTK 1.2), then it wouldn't be so hard 
to port to windows and gnu/linux, because GTK runs on these platforms as 
well.  as for MIDI stuff, yeah, i have no idea.  i agree the actual 
internal randomizing stuff is what's most interesting to look at -- it's 
more the structure and algorithms we care about than the interface -- 
and what's also easiest to port.

i was thinking . . . one could also write everything in java (including 
the midi bit), this would perhaps be the easiest way to make it portable 
(java is plenty fast enough for small midi apps like this).

>That said, it would be nice if other people could build on my efforts. I will have to look up the sources (they're archived, and I don't have a MacOS 9 development system on my current machine...), and see if there is some way in which I can share the work. (Might take a while.)
>
>Bottom line: I'll do my best, but I can't promise anything.
>  
>
Ok, that's totally fine.  If you want a little help setting up a 
sourceforge project (including the CVS repository), I can pitch in 
there.  Not sure about being able to do any programming with my current 
schedule though.

Cheers,
Chris