Hi Arthur, > Let me preface my remarks by saying that my main focus with panaxia has > been and still is developing a web interface. I restrict my efforts > simply because the time I can afford to devote to the project is quite > limited.
Yes, I've been spending time on other things recently, I haven't done any further development on the scenegraph graphics framework as it met my needs and there didn't seem to much initial interest in it. > I've never much liked HyperDoc and it's always been in the back of my > mind to work on that, like what I did with the Axiom book in > xhtml+mathml, although now with your work we need something like > xhtml+mathml+svg+x3d+webgl.... I'll have to look at your html work too > but my immediate interest is in using your work, in conjunction with > webgl capability appearing in recent browser builds, to enable graphics > in my browser interface. > I would say that yours will become default if people like it better. I > never much liked the old framework either. Both it and hyperdoc seem > clunky and dated. In terms of using the old framework for my own > project I didn't like the one way communication, spad -> C, and of > course the near total lack of documentation in the C code is annoying. > I wondered whether or not it was worth reusing the C code or throwing it > out. With your work, and for my purposes, it looks like throwing out is > looking good. > I'm not sure what you mean by render itself directly from openaxiom. > The webgl capability being introduced in the latest browser builds is > based on opengl es 2.0 which is a stripped down opengl for mobile > devices. The mozilla people are talking about adding more opengl > capability as extensions to webgl. > Certainly I think doing as much as possible in openaxiom makes sense. > So right now to do a 3d plot in the browser interface I could presumably > send your x3d representation to the browser and then parse it with > javascript and convert it into the webgl commands, but it makes sense to > me to do more of that in openaxiom before sending it out. However > ultimately webgl is a javascript binding to opengl es 2.0 so the > javascript has to be executed in the browser. > > If you're interested in desktop rendering then we need an opengl > package/domain that can produce either full opengl or the stripped down > webgl. I guess the ideal would be for graphics that works across all platforms: (Linux command line, emacs, windows and web interface) and all flavors of lisp: (GCL,sbcl, clisp, Clozure,ECL, cmucl) and all flavors of panAxiom. but perhaps that is too ambitious? I understand some flavors of lisp have an openGL extension so that openGL calls can be made directly from lisp code so presumably could be made directly available to SPAD (perhaps that's what I mean by render itself directly from openaxiom). It seems to me that this might be more portable across the different platforms but less portable across the different flavors of lisp? The other issue is: could such a solution open a new graphics window? and would it block the normal execution of the command line functions? I suspect that each graphics window would have to run in its own thread and so some C/C++ code might still be required: 1) as a thin wrapper for the openGL commands and to put them in their own thread. 2) to support all flavors of lisp. As you can probably tell, I'm out of my depth on these system integration issues. > I agree something like this is the way to go. I can't see openaxiom, or > the other panaxia for that matter, becoming very popular without a more > agreeable and modern interface. I first tried Axiom in the late 90's. > At the time I was fully immersed in Mathematica, provided for free at > the applied math computing centre, and when I looked at the Axiom > command line interface I wasn't very impressed. Subsequently I saw the > light and now won't touch Mathematica, but the issue with the axiom > interface remains. > > I plan to continue working on this as time permits and having somebody > else thinking similarly will be motivating. I agree, I would like to continue working scenegraph graphics framework. To extend it in the ways we are discussing here I think it needs someone who understands the system integration issues and is prepared to do the development, debugging and testing it takes to get something like this working across all platforms and lisps. Martin ------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl _______________________________________________ open-axiom-devel mailing list open-axiom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open-axiom-devel
