Apologies if that doesn't make much sense, I should've proofread it :) S++
----- Original Message ---- > From: Stuart Axon <[email protected]> > To: [email protected] > Sent: Mon, September 20, 2010 10:33:34 PM > Subject: Re: [shoebot-devel] shoebot android someday? > > > From: "architetto francesco fantoni <hva - hermanitos verdes architetti>" > ><[email protected]> > > > To: [email protected] > > Sent: Mon, September 20, 2010 9:55:31 PM > > Subject: Re: [shoebot-devel] shoebot android someday? > > > > though pyglet need ctypes that at present is not part of SL4A python > > > Also it's for Opengl, not Opengl ES as android needs. > > > > > Il giorno lun, 20/09/2010 alle 22.48 +0200, architetto francesco fantoni > > ha scritto: > > > cool! > > > wasn't there already an attempt to port shoebot on pyglet? > > > > > > f. > > > > > A while back I got it rendering to a window in pyglet, but not > really usable (and still using cairo for everything). > > Still, in future it would be useful to be able to render onto > OpenGL textures this way. > > > > I've had a bit of a look at the nodebox code (just a little mind), > and it looks the way fill, stroke etc are just imported and share > state through global variables. > > In future I'd like to be able to do things like call one bot from > another (e.g. render a frame of another bot into rectangle). > > I think with the nodebox approach state will be leaked and this > won't be possible (without having seperate python processes). > > Now... I might be wrong and not actually understand their > architecture, but if I'm right I prefer our approach using classes. > > > Anyway... it might not matter, as at the moment we don't have a > functioning release...but: > It would be cool if in future we > could prove a nice architecture for supporting stuff like > drivers (e.g. for jython/pyglet) and solve the problem of > fragmentation of the nodeboxes :) > > [And not to forget; the ability to add different grammars > too]. > > > > Oh - while I'm braindumping this: > Something else to look at are the different ways of using > nodebox-gl: > > draw(canvas): > fill(...) > > And nodebox-1/shoebot: > > draw(): > fill() > > > Actually canvas is not a bad abstraction and we might want > to consider it for a 'pure python' mode. > We'd really want to reconcile this if we want to make all > the libraries shared. > > > > > > > > > > > > > > > Il giorno lun, 20/09/2010 alle 13.54 -0600, Sebastian Oliva ha scritto: > > > > On Sat, Sep 18, 2010 at 5:12 PM, Architetto Francesco Fantoni > > > > <[email protected]> wrote: > > > > > I just found out that cairo has a skia backend, I had not noticed > >before. Maybe porting a minimal cairo version to android could be done > >after > > >all, though probably not so easily... > > > > > > > > > > > > > > > "architetto francesco fantoni <hva - hermanitos verdes > > > > architetti>" > ><[email protected]> ha scritto: > > > > > > > > > >>Some very early, lazy and useless notes and suggestions about a > possible > > > > >>port of shoebot to android. > > > > >> > > > > >>With the development of the scripting layer for android going on > > > > >>( http://code.google.com/p/android-scripting/ ) that brings a > functional > > > > >>python environment, I started to figure out whether it is going to >be > > > > >>possible someday to have shoebot working on google's mobile os. > > > > >>Actually I can't imagine no real use or need, let's say it's just >for > > > > >>fun ;) > > > > >> > > > > >>obviously cairo has not been ported to android, and probably never >will > > > > >>(way too heavy and complex), > > > > >>and as shoebot is cairo based, a port would probably require the >cairo > > > > >>canvas to be replaced by something else. > > > > >>It seems android uses skia ( http://code.google.com/p/skia/ ) as low > > > > >>level library for 2D graphics, it is open source and I think it >could > > > > >>replicate the basic functions of cairo, but skia api is not even > >exposed > > > > >>through NDK and, more important, at present it has no python >bindings, > > > > >>though it seems it is in their roadmap. > > > > >>I suppose a lot of things would be rather hard to re-implement > > > with > skia > > > > >>(e.g. fonts, images, etc.) > > > > >> > > > > >>Another option I thought of could be openvg > > > > >>( http://www.khronos.org/openvg/ ). > > > > >>The most complete open source implementation I know of is ShivaVG > > > > >>(http://sourceforge.net/projects/shivavg/). It's built on top of > OpenGL, > > > > >>and a port to android, though not yet existing, could be done using > > > > >>OpenGL ES, and therefore being hardware accelerated, that's not bad. > > > > >>I tried an attempt of porting it using android NDK, but the lack > > > > in > GLES > > > > >>of some GL attributes implies patching the code and rewriting parts > > > > >of > > > > >>it in order to use OpenGL ES, and obviously it's a task far too >complex > > > > >>for me. > > > > >>There is a commercial openvg implementation called Amanith > > > > >>(http://www.amanith.org/project.html). which has already an > >experimental > > > > >>port to android > > > > >>(http://comments.gmane.org/gmane.comp.handhelds.android.ndk/3657) >that > > >I > > > > >>was able to run as simple test on my phone (but as I said it is > > > > >>commercial). > > > > >>The nice aspect is that it already exists a project that provides > python > > > > >>bindings for both ShivaVG and Amanith vg > > > > >>(http://code.google.com/p/pyopenvg/) though I must admit I have not > > > > >>tested it yet. > > > > >> > > > > >>I'm sure there are other smarter or simpler possibilities for > replacing > > > > >>cairo on android that I didn't think of. > > > > >> > > > > >>All of the shoebot gui of course would need a rewrite as well, as of > > > > >>course there won't be a gtk environment, but simple gui with python >is > > > > >>already possible on android, so I suppose that as long as the >scripting > > > > >>layer is further developed, more complex python programs can really > > > > >be > > >a > > > > >>possibility on android. > > > > >> > > > > >>Other non-pure-python modules can be ported to android already (PIL > > > > >>maybe?) but I must admit it's not a trivial task at all. > > > > >> > > > > >>best, francesco > > > > >> > > > > >> > > > > >> > > > > >> > > > > >> > > > > >>_______________________________________________ > > > > >>Shoebot-devel mailing list > > > > >>[email protected] > > > > > >>http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net > > > > > > > > > > ---- > > > > > Francesco Fantoni Architetto > > > > > via s.giacomo, 38 > > > > > 41121 modena / italia > > > > > -- -- > > > > > ph. & fax.: +39059217554 > > > > > web: www.hv-a.com > > > > > _______________________________________________ > > > > > Shoebot-devel mailing list > > > > > [email protected] > > > > > >http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net > > > > > > > > > > > > > I think an easier way would it be to use > > > > http://organisms.be/NodeBox_for_OpenGL as a base > > > > _______________________________________________ > > > > Shoebot-devel mailing list > > > > [email protected] > > > > http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net > > > > > > > > > _______________________________________________ > > > Shoebot-devel mailing list > > > [email protected] > > > http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net > > > > > > _______________________________________________ > > Shoebot-devel mailing list > > [email protected] > > http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net > > > > > > _______________________________________________ > Shoebot-devel mailing list > [email protected] > http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net > _______________________________________________ Shoebot-devel mailing list [email protected] http://lists.tinkerhouse.net/listinfo.cgi/shoebot-devel-tinkerhouse.net
