Talk:Macsyma

Macsyma as a

Macsyma as a Common Lisp program: today you would likely use a Common Lisp version of Macsyma. Maclisp is long gone. In the page there are mentioned the implementations in Gold Hill Lisp (which is a Common Lisp) and CLOE (Common Lisp Operating Environment). Maxima is also written in Common Lisp. Common Lisp was designed to be backwards compatible, so that ancient software like Macsyma can be ported with relatively little effort. Anybody now using Macsyma and derivatives (and not looking at the historical ancient Macsyma) would see a Common Lisp program and would try it in a Common Lisp implementations. -- RJ (comment by 85.182.78.172 at 2007-12-22T13:42:47)

Yes, I agree that Common Lisp is largely backwards compatible with MacLisp (in which Macsyma was written), and indeed most old Lisp programs can be moved to Common Lisp without too much difficulty. So I suppose if you're interested in a catalogue of software which can be run under Common Lisp, Macsyma qualifies. But Wikipedia is not a software catalogue, it is an encyclopedia; if you're interested in the programming style and history of Macsyma, it's bizarre to say that it's a Common Lisp program. After all, almost all Unix/Linux programs can run under Windows (using cygwin etc.); should we then say that grep is a "Windows program"? That seems silly. --Macrakis (talk) 07:01, 23 December 2007 (UTC)

Paragraph removed

I have removed a unsourced paragraph on numerical capabilities of Macsyma for several reasons. Firstly, it asserts that the weakness of Macsyma was to be inefficient for tasks for which it was not designed (numerical capabilities well performed by Fortran). Secondly, it asserts that the weak impact of Macsyma in the engineering world was due to these weak numerical capabilities. This was, maybe, the opinion of some of the developers, but certainly not a consensus. Moreover, this weak impact was not a specificity of Macsyma, but was shared, until must later, by all computer algebra. Around 2002, I participated to a discussion panel on this subject, where we could not reached. Thus comparing Macsyma and Fortran, and discussing the impact of Macsyma in industry is WP:OR, and does belongs to Wikipedia. D.Lazard (talk) 07:25, 27 September 2015 (UTC)

Please don't truncate the contributions list

The development section has always listed contributions throughout the entire history of the development of Macsyma. Please to not attempt to truncate it at some arbitrary date. This would be difficult to do in a reasonable manner (e.g. because various projects had long development cycles and went through various stages of prereleases and limited releases. Be aware that the contributions list is far from complete. Hopefully others will help to complete it. --Bill D. — Preceding unsigned comment added by 71.233.247.156 (talk) 20:23, 28 September 2015 (UTC)

This contribution list appears in the section devoted to the development of Macsyma at MIT. It is thus misleading to include here later developments. If you want to include later contributions, this must appear in another section. In any case, everything must be sourced, not only because it is a rule of Wikipedia, but also for distinguishing the contributions specific to Macsyma from the implementation in Macsyma, without specific added value, of algorithms developed elsewhere. D.Lazard (talk) 09:29, 29 September 2015 (UTC)

The Development section is not constrained to the MIT period - nor is their any good reason that it should be. — Preceding unsigned comment added by 71.233.247.156 (talk) 21:06, 10 October 2015 (UTC)

Discussion with an IP editor

I have had a mail discussion with an IP user that has recently edited this article. As a mart of his mail may be of general interest, I reproduce it here:

" You wrote > There is, alas, no written history of the latter half of > Macsyma's life (and the brief accounts of the earler half are neither > balanced nor complete). So the best we can rely on are the > recollections of the members of the group. I hope that they will > contribute while they are still with us, since otherwise it will prove > diifcult if not impossible to reconstruct the precise history.

I agree with this. However, Wikipedia is not the right place to reconstruct this precise history. In fact, one of the main Wikipedia policies, often called "No original research" is that every assertion in Wikipedia must be verifiable, in the sense that it relies on sources that are published elsewhere. This means that the contributions of the members of the group must be published elsewhere before being inserted in Wikipedia page on Macsyma.

This policy against "original research" would imply that the list of contributors and contributions should be removed (except for the part that appears somewhere). As I have not taken the time to verify the part of this list that appears in references, and as it is nevertheless interesting to know about the main contributors, I have only tagged the list with "citation needed". Moreover, as this list appears in the section devoted MIT period, it should not contain later contributors/contributions.

The list of contributions sets another problem: The content of Wikipedia must be "notable". IMO, for being notable, a specific contribution to a software such as Macsyma, must either to have been the state of the art at some moment, or to have been widely used. This must be attested by publications of either the contributor or users. Otherwise, a contribution is probably the transcription in Macsyma of others' work and does not deserve to be mentioned in an encyclopedia. This is typically the case of contributions "Grobner": Although I know rather well the literature about Gröbner bases, I do not remember any paper mentioning a "Grobner" package in Macsyma, nor papers on the subject by the cited contributors.

Another Wikipedia policy is that all Wikipedia article must have a "neutral point of view". This implies that every opinion must be attributed explicitly to its author. The paragraph "The greatest product weakness of Macsyma was its numerical analysis" is a controversial analysis of the reasons for which computer algebra was very long to "get a piece of a much larger engineering and lower-end scientific markets". Presently, computer algebra is widely used by engineers and non-mathematicians, and it is clear that it is not because of its numerical capabilities. At Symbolics it was already clear for many people that Macsyma (and more generally computer algebra) and numerical computation are completely different tasks, which are not concurrent (as suggested by this paragraph) but complementary. I am quite sure that contrarily what is said in this paragraph, this opinion was shared by early Macsyma developers and some of lisp developers at Symbolics. Thus, this paragraph (that is not essential for the history of Macsyma) must be removed.

I will copy this answer in the talk page of Macsyma, and editing the article accordingly. Please do not revert my edits without before discussing on the talk page. "

Following the conclusion of my mail, I'll restore the {{cn}} tags on the list of contributions, and remove the paragraph beginning by "The greatest product weakness of Macsyma ...". D.Lazard (talk) 11:09, 6 October 2015 (UTC)

I'll second the above.

I have no problem with including contributors but tagging them with cn.

Macsyma was not intended for numerical math. I believe one could evaluate expressions (plug in a float), but that was not its primary focus and I doubt there were any guarantees about the accuracy of the result.

Glrx (talk) 19:11, 8 October 2015 (UTC)

Mr. Lazard: please refrain from making false and misleading edits. You were never a member of the Macysma group nor did you ever have any close affiliation with it, so you have little knowledge on these matters. — Preceding unsigned comment added by 71.233.247.156 (talk) 20:15, 10 October 2015 (UTC)

71.233.247.156 - I highly recommend that you read Wikipedia's No original research policy - the one that D.Lazard referenced in his response. Your claims of affiliation with the group do not constitute a higher merit or ownership of the article. Any edits you make that aren't attributable or cited by a reliable source can constitute original research, and can be removed. Please do not hesitate to respond with any questions. ~Oshwah~ ^{(talk) (contribs)} 20:23, 10 October 2015 (UTC)

There is already a global "citation needed" on the entire contributions list so there is no need to annotate individual entries.

I reverted the removal because it is under discussion here. The cn tags target individuals and contest their contributions. I offered a possible ref for Groebner that seems to dispute a cn objection, but I don't know if it is relevant. I suspect many authors are not indispute; for example, Zippel and EZGCD sounds familiar. Glrx (talk) 21:33, 10 October 2015 (UTC)

Most of the contributions list would be difficult if not impossible to back up with links to online sources. It is accurate (though not complete by any means).

---

Not that hard, really.

For some, contributors, the only reference available would be annotations in the source. For example, I didn't write any Macyma papers, but I signed a comment "--RWK" in this source file: https://sourceforge.net/p/maxima/site-xml/ci/9bdedbde09f7590a301f5a3252bc5fa362f00ae8/tree/maxima/src/solve.lisp (I am not listed -- I contributed less than the ones now on the list, but it was non-zero. I did more on maintaining Maclisp, NIL, ITS than writing Macsyma. I'll leave it to others to decide whether I ought to be included).

Rich Zippel (RZ) and Mike Genesereth (MRG) appear in some of the source filenames (you can point to the above link's containing directory as a reference).

Jeff Golden (JPG) and Richard Bogen are listed as authors of the Macsyma Reference Manual, volume 2

http://americanhistory.si.edu/teachingmath/html/enlarge/2001_11052.htm

The Acknowledgement section of rev 16 of the manual provides support for more. I'll call out ones that are relevant as contributors.

https://people.eecs.berkeley.edu/~fateman/macsyma/docs/refman16.pdf

Richard Bogen -- original author of manual

Ellen Golden -- editorial, and see below as Ellen Lewis

Michael Genesereth -- also listed above

Richard Pavelle, Michael Wester, Charles Karney, Kent Pitman, Richard Fateman, Alexander Doohovskoy -- all members of the group

Paul Wang is also listed there as affiliated with Kent State, but he spent a period of time at MIT as part of the group.

Paul Wang and Ellen Lewis are also supported by this reference: http://people.csail.mit.edu/psz/LCS-75/knowledge.html

Stavros Macrakis is easily verified as a current contributor to Maxima, but I'm having trouble locating a MIT-era Macsyma-specific citation for him (I found one for his work on Maclisp). I can personally attest he was there and involved; I'll continue to try.

Joel Moses, Bill Martin, Richard Petti, Richard Fateman are supported by The Macsyma Book, which is also a significant historical reference that some of the article should be checked against.

http://maxima.sourceforge.net/docs/maximabook/maximabook-19-Sept-2004.pdf

Richard Fateman has made available the 1977 Macsyma Users' Group proceedings, which cites many, including John Kulp, CFF Karney, Dave Barton.

https://www.academia.edu/29190838/An_improved_algorithm_for_the_isolation_of_polynomial_real_zeros?auto=download

David Barton's Taylor series representation is discussed in this book: https://books.google.com/books?id=xP7iBQAAQBAJ&pg=PA147&lpg=PA147&dq=david+barton+macsyma&source=bl&ots=K9qWTY2pX8&sig=olLU970TSvwUH3Xe-wEsa1taWOg&hl=en&sa=X&ved=0ahUKEwil7uaqgd3TAhVqllQKHaqVBnMQ6AEILDAB#v=onepage&q=david%20barton%20macsyma&f=false

Finally, I'll note, in jest, that the entire list has been published in a book. However, it's a circular reference -- it quotes Wikipedia!

https://books.google.com/books?id=cuJMzvdIWMcC&pg=PA75&lpg=PA75&dq=stavros+macsyma&source=bl&ots=BCteeeGXLc&sig=YDy06tV3df80-jV2S_gTGUFpKnk&hl=en&sa=X&ved=0ahUKEwig7MKS-9zTAhXHzFQKHZwjBQAQ6AEIPjAF#v=onepage&q=stavros%20macsyma&f=false

Bob Kerns (talk) 05:03, 7 May 2017 (UTC)

False and misleading edits reverted

I have a reverted a number of edits that were blatantly false and/or misleading. First, someone misinterpreted a reference to the Bronstein(-Risch) algorithm and changed it to an entry claiming that Bronstein had contributed to the development of various Macsyma modules (false). Second, someone changed some entries referring to development of standard bases algorithms in the MIT Mathlab (Macsyma) group to instead refer to independent work by others. These have been reverted and corrected, and some citations have been added. Please be much more careful in the future, esp. if you have no firsthand knowledge of these matters. — Preceding unsigned comment added by 71.233.247.156 (talk) 00:30, 12 October 2015 (UTC)

There's no polite way to say this, and it's been suggested to you before above. Wikipedia does not value firsthand knowledge, and in fact requires those with direct involvement with a topic to stand back from it. Instead, Wikipedia values published references from reputable sources. Macsyma has a great history of publication, going back to its origins in Project MAC, some of which are already in use in this article. That's how you support claims in a good Wikipedia article. And thanks - I've been a Macsyma fan since 1977. RossPatterson (talk) 02:07, 14 October 2015 (UTC)

Ross, that isn't even what he said. He was purely going off references available. Do you have a vested interest in giving undue credit to people who didn't actually contribute the way they said they did?