Re: [open-axiom-devel] [Ecls-list] Solaris 9 build report and suggested fixes with current cvs
A system for computer algebra and symbolic mathematics
Brought to you by:
dos-reis
From: Michael.Abshoff <mic...@go...> - 2007-08-30 00:32:43
|
Bill Page wrote: > On 26 Aug 2007 11:31:57 -0500, Gabriel Dos Reis wrote: > >> Michael Abshoff writes: >> Hello folks, >> ... >> | I also just checked the openaxiom devel mailing list archives and saw >> | that Bill Page has accepted axiom maintainership of the MinGW port >> | as well as the Sage integration. I was under the impression that FriCAS >> | had been favored (at least initially before the second Axiom fork) for Sage >> | integration, but things might have changed since the last time I checked. >> >> Bill has better answers than I about this. >> >> Most definitely, it is OpenAxiom's intent to work with SAGE people to >> facilitate `plugability' in SAGE. We had some brief discussion about >> that a week ago, I think. >> >> > > The previous version of the Axiom package installable by Sage (called > 'axiom4sage-0.1.1'), was based on an earlier version of the > build-improvements branch of the Axiom project. Build-improvements is > now continued by OpenAxiom so creating an 'axiom4sage' based on > OpenAxiom would be straightforward. But right now FriCAS is being used > for the next version of the Sage Axiom package (called > 'axiom4sage-0.3.1') because Waldek has implemented a very significant > speed-up for building Axiom based on pre-compiled cached Lisp code. > The generated Lisp is architecture independent and can be compiled to > a working Axiom system by Clisp in about 12 minutes on a reasonably > fast (e.g. > 1.5 GHz) system - compared to several hours building from > the usual source distribution. A fast build time is essential if we > hope to ever convince the Sage developers to include Axiom as a > "standard" package in Sage. Since Clisp is current a standard package > in Sage and is used to build Maxima, it is also the preferred Lisp for > building Axiom in Sage. > > Although this is not currently so important to Sage users, it is also > important to know that FriCAS includes some critical fixes to hyperdoc > that are not (yet) included in OpenAxiom. Without these changes it is > relatively easy to cause hyperdoc to crash while doing certain > searches. With the older 'axiom4sage-0.1.1' is was possible to start a > hyperdoc and graphics aware Axiom session using the command: > > $ sage -axiom > > but unfortunately this is not possible with 'axiom4sage-0.3.1' because > the current clisp-based version of Axiom in FriCAS does not support > the sman session interface. > > Except for the pre-compiled Lisp option, in general the FriCAS build > system attempts to building every possible from the basic sources. > This includes the all of the generated graphics and Axiom results > included in hyperdoc as well as the Axiom database that is used while > compiling the Spad code in the Algebra library "from the ground up". I > believe that this is a reasonable goal for OpenAxiom as well however > porting the FriCAS changes to the hyperdoc and Algebra build into > OpenAxiom might be rather challenging (at least it would be so for > me!). All of which raises the longer term issue of the continued > existence of two alternate forks of the original Axiom project. I have > to say that I have nothing strongly against the current status quo and > I can forsee continuing in this mode for at least the short term - at > least as long as the energy and dedication of the creators of these > forks remains at a high level. Currently I consider myself a > "developer" in all three of the Axiom-related projects and subscribe > to all of the mailing lists for these projects. Although this is a > little awkward, it is not impossible to work in this manner. > > My hope is that it will be possible to maintain a cordial and active > (but presumably rather selective) exchange of patches between the > various Axiom-related projects. For example I have seen recent patches > to FriCAS by Waldek labelled "ported from OpenAxiom". Shortly this may > also include the new open source Aldor project that will allow us to > officially add Aldor as a supported library compiler for Axiom > (non-commercial use only licensing issues aside). For example at the > recent Aldor Workshop we introduced (but without resolution) the issue > of which Axiom/Aldor project will adopt and maintain the Axiom Aldor > interface originally contributed by Peter Broadberry? It seems that > Tim Daly has (more or less) rejected the idea of official supporting > Aldor in the original Axiom project until or unless aldor.org can > offer a more compatible license. > > >> ... >> > > I suppose all of this is more than you really wanted to know, :-) but > I would be quite interested in your opinions and observations about > this current state of affairs. > > This is all quite interesting, especially since I followed the Axiom forks only via reading the archives after it had happened. Lacking political insight this summarizes the situation nicely. > Regards, > Bill Page. > > Cheers, Michael > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > open-axiom-devel mailing list > ope...@li... > https://lists.sourceforge.net/lists/listinfo/open-axiom-devel > > |