I have run into what I hope are minor glitches in running (all of) Rubi on mockMMA; I suspect that the more recent Rubi will have similar or more glitches. (You are, of course, welcome to include mockMMA in Sage). Why didn't I go further? (a) It seemed that some parts of Mathematica were being used explicitly or implicitly in the patterns or replacements, parts that I had not implemented or implemented differently. I forget the details, though I seem to recall these led to loops, which should be impossible structurally in the rule set unless we are (for example) missing simplifications. Anyway, after conversations with Albert Rich and a lack of interest by others (previous to this...) for someone to pick it up and run with it, I put it on a back burner.
(b) Albert wrote this stuff as Mathematica rules. I can run Mathematica. He can run Mathematica. So far as I can tell, the number of users of MockMMA is quite small, and I hadn't heard any clamoring for Rubi. (c) I think the right approach to integration of hairy expressions requiring optimal answers that do not fall out of algorithms is what I did a while back, with TILU. which ran for a few years and then I retired it when the last of the servers was tossed out (It ran on Sun, DEC, and HP at various times). TILU responded to maybe 100,000 integration questions and used table lookup. Not nearly as sophisticated as Rubi on Rubi's terrain. TILU may have included a more varied set of patterns in some aspects. It included a pile of stuff we just through in about Bessel functions, for example. Anyway, we didn't have the terms available when we started (or ended!) TILU, but I think it now would be called Software as a Service that was available "in the Cloud". Jargon like this makes me cringe. The goal then for Rubi (and it could be used for other parts of Sage, especially those that are bulky and rarely used, but whose use typically represents a minuscule amount of computing time) would be to install it in one or a few places as a web server responding to requests from remote but www-connected computer systems for Rubi-lookups. These could be done in parallel with local algorithmic methods. Additionally (as was proposed for TILU), new patterns could be incrementally added to the servers from remote sites for augmenting the data base. (Trusted sites, data verified, I suppose). (d) If there are people who are encountering difficulty with MockMMA, they can contact me. (e) It seems to me there was another imitation of Mathematica produced recently. Mathics? If that supports Rubi, and is license-friendly, why not try that? I gather it is written in Python. RJF -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/groups/opt_out.
