Re: 6 months free Claude Max 20x for working on Maxima
David Scherfgen via Maxima-discuss <[email protected]> Wed, 22 Jul 2026 21:50:34 +0200
| Newsgroups | gmane.comp.mathematics.maxima.general |
|---|---|
| Message-ID | <CAMTHLKgFXj6AUjmVQWmuD06AuNOBW16EBWA52aM-oJRce3GK2Q@mail.gmail.com> |
--===============8788453674142352802== Content-Type: multipart/alternative; boundary="000000000000948c740657387440" --000000000000948c740657387440 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable You can see here what kind of bugs Claude was able to find just by looking at the source code: https://sourceforge.net/p/maxima/bugs/search/?q=3Dreported_by:tomasriker Almost all of these were identified (and mostly fixed) by Claude. All I told it was "*You're a bug hunter! Today, you'll be focusing on the file xyz.lisp. Read the source code carefully, also consult other files when you encounter unknown functions, macros or variables, to make sure you understand their purpose. Report all potential bugs, each with a brief description, a Maxima user-level reproducer and an idea for resolving the issue.*" It then kept thinking for several minutes and came back with a nice collection of mostly real bugs and solutions. I think Claude is pretty good at differentiating between what code is supposed to do and what it actually does. For example, it found out that the bigfloat code makes some rounding errors in addition and division (fix in the pipeline). It can use tools like Git to get an idea of how the code evolved over time, when and why it was changed, etc. It can write and execute helper scripts in Python to check results, do simulations, etc. And in theory, it can directly modify Maxima's source code, run Maxima commands, debug, test, ... - but I haven't let it do that. If you haven't experimented with LLMs, I strongly recommend to give it a try. Make sure to select an advanced model with reasoning (thinking) support. Honestly: The amount of objectively real bugs in Maxima that it has found out within minutes of processing is truly mind-blowing. It would probably take a human hundreds, maybe thousands of hours. Best regards David Scherfgen Richard Fateman <[email protected]> schrieb am Mi., 22. Juli 2026, 20:19: > Some AI types seem to be viewing the Wolfram results as ground truth. If > the Maxima and Mathematica and Sympy ... results are different, that at > least suggests a closer look is warranted. > > I am somewhat skeptical about Claude figuring out the difference between > case 1: the code, as written, is the only concrete definitive > specification of an algorithm and therefore agrees with the spec and is > inherently correct > and > case 2: same as case 1 except there is an undiscovered bug (not revealed > by any test cases). > > and then refactoring the code, adding comments, etc. > > Of course humans might not be so great at this either. Depends on the > human... > > Cheers > Richard > > On Wed, Jul 22, 2026 at 10:34=E2=80=AFAM Raymond Toy <[email protected]= om> > wrote: > >> On 7/21/26 12:00 PM, Gunter K=EF=BF=BDnigsmann via Maxima-discuss wrote: >> >> >> One other non-mathematical thing would be asking it to clean up the buil= d >> dependencies so incremental builds actually yield the result one expects >> and don't forget dependencies or hit circular ones. >> >> A few months ago I asked Claude to check the dependencies between files >> to make them explicit and the dependencies between modules too. Many of >> these changes have been incorporated. Along the way, some modules got sp= lit >> into smaller ones to simplify the dependencies. >> >> My original goal was to get the build order correct so the def-simplifie= r >> compiler macros were always defined ahead of use. What I learned is that >> there are a lot of circular dependencies. Some of these were fixed but m= any >> were not. I think this would require a huge refactoring of the existing >> code moving all kinds of things around. I stopped at that point because = it >> was a lot of work. Claude could probably do it fairly easily, but it wou= ld >> make the file names and modules a lot less meaningful and I didn=E2=80= =99t know how >> to keep them simple. >> >> >> We could ask it to clean up our mess with global variables as a first >> step towards multi-threading. >> >> I=E2=80=99ve worked on this over the years doing it by hand. I haven=E2= =80=99t tried with >> Claude but it=E2=80=99s on my list. If nothing else, it will at least ma= ke it >> easier to find all the uses. However, it=E2=80=99s also hard because lot= s of places >> use the same (short) name for different things. >> >> >> We could ask it to perform a stability review. >> >> I don't know how good it is in maths. But when it can change stuff so it >> is testable, add tests and then test the results of its work it really >> excels. Or we could ask it what needs to be done so it can port Rubi. >> >> For some other projects, Claude has been quite good at creating tests. >> For those projects, it=E2=80=99s easily verifiable that it=E2=80=99s cor= rect. Could be >> harder in maxima, if, say, Claude says here=E2=80=99s a definite integra= l with >> corresponding result. It could be difficult the verify that that=E2=80= =99s correct. >> We could compute it numerically, but that=E2=80=99s not a proof of corre= ctness. >> ​ >> _______________________________________________ >> Maxima-discuss mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/maxima-discuss >> > _______________________________________________ > Maxima-discuss mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/maxima-discuss > --000000000000948c740657387440 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"auto">You can see here what kind of bugs Claude was able to fin= d just by looking at the source code: <a href=3D"https://sourceforge.net/p/= maxima/bugs/search/?q=3Dreported_by:tomasriker">https://sourceforge.net/p/m= axima/bugs/search/?q=3Dreported_by:tomasriker</a><div dir=3D"auto"><br><div= dir=3D"auto">Almost all of these were identified (and mostly fixed) by Cla= ude. All I told it was "<i>You're a bug hunter! Today, you'll = be focusing on the file xyz.lisp. Read the source code carefully, also cons= ult other files when you encounter unknown functions, macros or variables, = to make sure you understand their purpose. Report all potential bugs, each = with a brief description, a Maxima user-level reproducer and an idea for re= solving the issue.</i>"</div><div dir=3D"auto">It then kept thinking f= or several minutes and came back with a nice collection of mostly real bugs= and solutions.<br><div dir=3D"auto"><br></div><div dir=3D"auto">I think Cl= aude is pretty good at differentiating between what code is supposed to do = and what it actually does. For example, it found out that the bigfloat code= makes some rounding errors in addition and division (fix in the pipeline).= </div><div dir=3D"auto"><br></div><div dir=3D"auto">It can use tools like G= it to get an idea of how the code evolved over time, when and why it was ch= anged, etc. It can write and execute helper scripts in Python to check resu= lts, do simulations, etc. And in theory, it can directly modify Maxima'= s source code, run Maxima commands, debug, test, ... - but I haven't le= t it do that.</div><div dir=3D"auto"><br></div><div dir=3D"auto">If you hav= en't experimented with LLMs, I strongly recommend to give it a try. Mak= e sure to select an advanced model with reasoning (thinking) support.</div>= <div dir=3D"auto"><br></div><div dir=3D"auto">Honestly: The amount of objec= tively real bugs in Maxima that it has found out within minutes of processi= ng is truly mind-blowing. It would probably take a human hundreds, maybe th= ousands of hours.</div><div dir=3D"auto"><br></div><div dir=3D"auto">Best r= egards</div><div dir=3D"auto">David Scherfgen</div></div></div></div><br><d= iv class=3D"gmail_quote gmail_quote_container"><div dir=3D"ltr" class=3D"gm= ail_attr">Richard Fateman <<a href=3D"mailto:[email protected]">fateman@= gmail.com</a>> schrieb am Mi., 22. Juli 2026, 20:19:<br></div><blockquot= e class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px s= olid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr"><div dir=3D"ltr"><= div class=3D"gmail_default" style=3D"font-family:arial,sans-serif;font-size= :small">Some AI types seem to be viewing the Wolfram results as ground trut= h.=C2=A0 If the Maxima and Mathematica and Sympy ... results are different,= that at least suggests a closer look is warranted.</div><div class=3D"gmai= l_default" style=3D"font-family:arial,sans-serif;font-size:small"><br></div= ><div class=3D"gmail_default" style=3D"font-family:arial,sans-serif;font-si= ze:small">I am somewhat skeptical about Claude figuring out the difference = between</div><div class=3D"gmail_default" style=3D"font-family:arial,sans-s= erif;font-size:small">case 1:=C2=A0 the code, as written, is the only concr= ete definitive specification of an algorithm and therefore agrees with the = spec and is inherently correct</div><div class=3D"gmail_default" style=3D"f= ont-family:arial,sans-serif;font-size:small">=C2=A0 and</div><div class=3D"= gmail_default" style=3D"font-family:arial,sans-serif;font-size:small">case = 2: same as case 1 except there is an undiscovered bug (not revealed by any = test cases).</div><div class=3D"gmail_default" style=3D"font-family:arial,s= ans-serif;font-size:small"><br></div><div class=3D"gmail_default" style=3D"= font-family:arial,sans-serif;font-size:small">and then refactoring the code= , adding comments, etc.</div><div class=3D"gmail_default" style=3D"font-fam= ily:arial,sans-serif;font-size:small"><br></div><div class=3D"gmail_default= " style=3D"font-family:arial,sans-serif;font-size:small">Of course humans m= ight not be so great at this either.=C2=A0 Depends on the human...</div><di= v class=3D"gmail_default" style=3D"font-family:arial,sans-serif;font-size:s= mall"><br></div><div class=3D"gmail_default" style=3D"font-family:arial,san= s-serif;font-size:small">Cheers</div><div class=3D"gmail_default" style=3D"= font-family:arial,sans-serif;font-size:small">Richard</div></div><br><div c= lass=3D"gmail_quote"><div dir=3D"ltr" class=3D"gmail_attr">On Wed, Jul 22, = 2026 at 10:34=E2=80=AFAM Raymond Toy <<a href=3D"mailto:toy.raymond@gmai= l.com" target=3D"_blank" rel=3D"noreferrer">[email protected]</a>> w= rote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0p= x 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><u></u> =20 =20 =20 =20 =20 <div> <div> <p style=3D"margin-bottom:19.2px;margin-top:0px">On 7/21/26 12:00 PM, Gunter K=EF=BF=BDnigsmann via Maxima-discuss wrote:<br> </p> <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rgb(= 119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;border= -left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,119)= ;margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;padding= -right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-emphasi= s-color:rgb(119,119,119)"> <div dir=3D"auto"><br> One other non-mathematical thing would be asking it to clean up the build dependencies so incremental builds actually yield the result one expects and don't forget dependencies or hit circular ones.=C2=A0<br> </div> </blockquote> A few months ago I asked Claude to check the dependencies between files to make them explicit and the dependencies between modules too. Many of these changes have been incorporated. Along the way, some modules got split into smaller ones to simplify the dependencies. <p style=3D"margin-bottom:19.2px;margin-top:0px">My original goal was to get the build order correct so the <code style=3D"backg= round-color:rgba(175,184,193,0.2);border-radius:3px;font-family:Consolas,In= consolata,Courier,monospace;font-size:16px;font-weight:550;margin-left:2.4p= x;margin-right:2.4px;padding-left:4.8px;padding-right:4.8px">def-simplifier= </code> compiler macros were always defined ahead of use. What I learned is that there are a lot of circular dependencies. Some of these were fixed but many were not. I think this would require a huge refactoring of the existing code moving all kinds of things around. I stopped at that point because it was a lot of work. Claude could probably do it fairly easily, but it would make the file names and modules a lot less meaningful and I didn=E2=80=99t k= now how to keep them simple.</p> <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rgb(= 119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;border= -left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,119)= ;margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;padding= -right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-emphasi= s-color:rgb(119,119,119)"> <div dir=3D"auto"><br> We could ask it to clean up our mess with global variables as a first step towards multi-threading.=C2=A0<br> </div> </blockquote> I=E2=80=99ve worked on this over the years doing it by hand. I haven= =E2=80=99t tried with Claude but it=E2=80=99s on my list. If nothing else, it wi= ll at least make it easier to find all the uses. However, it=E2=80=99s also= hard because lots of places use the same (short) name for different things. <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rgb(= 119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;border= -left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,119)= ;margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;padding= -right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-emphasi= s-color:rgb(119,119,119)"> <div dir=3D"auto"><br> We could ask it to perform a stability review.<br> <br> I don't know how good it is in maths. But when it can change stuff so it is testable, add tests and then test the results of its work it really excels. Or we could ask it what needs to be done so it can port Rubi.</div> </blockquote> For some other projects, Claude has been quite good at creating tests. For those projects, it=E2=80=99s easily verifiable that it=E2= =80=99s correct. Could be harder in maxima, if, say, Claude says here=E2=80= =99s a definite integral with corresponding result. It could be difficult the verify that that=E2=80=99s correct. We could compute it numerical= ly, but that=E2=80=99s not a proof of correctness. </div> <div style=3D"height:0px;width:0px;max-height:0px;max-width:0px;overflo= w:hidden;font-size:0px;padding:0px;margin:0px" aria-hidden=3D"true" title= =3D"MDH:PGJyPjxicj48ZGl2IGNsYXNzPSJtb3otY2l0ZS1wcmVmaXgiPk9uIDcvMjEvMjYgMTI= 6MDAgUE0sIEd1bnRlciBL77+9bmlnc21hbm4gdmlhIE1heGltYS1kaXNjdXNzIHdyb3RlOjxicj= 48L2Rpdj48YmxvY2txdW90ZSB0eXBlPSJjaXRlIiBjaXRlPSJtaWQ6MDk4ODQxODAtRDY3NS00M= TYwLUI5RUQtREU5NUEyQkFGNjVCQHBldGVycGFsbC5kZSI+CgogICAgPG1ldGEgaHR0cC1lcXVp= dj0iQ29udGVudC1UeXBlIiBjb250ZW50PSJ0ZXh0L2h0bWw7ICI+CiAgCiAgPGRpdiBkaXI9ImF= 1dG8iPjxicj5PbmUgb3RoZXIgbm9uLW1hdGhlbWF0aWNhbCB0aGluZyB3b3VsZCBiZSBhc2tpbm= cgaXQgdG8gY2xlYW4gdXAgdGhlIGJ1aWxkIGRlcGVuZGVuY2llcyBzbyBpbmNyZW1lbnRhbCBid= WlsZHMgYWN0dWFsbHkgeWllbGQgdGhlIHJlc3VsdCBvbmUgZXhwZWN0cyBhbmQgZG9uJ3QgZm9y= Z2V0IGRlcGVuZGVuY2llcyBvciBoaXQgY2lyY3VsYXIgb25lcy4mbmJzcDs8YnI+PC9kaXY+PC9= ibG9ja3F1b3RlPkEgZmV3IG1vbnRocyBhZ28gSSBhc2tlZCBDbGF1ZGUgdG8gY2hlY2sgdGhlIA= pkZXBlbmRlbmNpZXMgYmV0d2VlbiBmaWxlcyB0byBtYWtlIHRoZW0gZXhwbGljaXQgYW5kIHRoZ= SBkZXBlbmRlbmNpZXMgCmJldHdlZW4gbW9kdWxlcyB0b28uJm5ic3A7IE1hbnkgb2YgdGhlc2Ug= Y2hhbmdlcyBoYXZlIGJlZW4gaW5jb3Jwb3JhdGVkLiZuYnNwOyAKQWxvbmcgdGhlIHdheSwgc29= tZSBtb2R1bGVzIGdvdCBzcGxpdCBpbnRvIHNtYWxsZXIgb25lcyB0byBzaW1wbGlmeSB0aGUgCm= RlcGVuZGVuY2llcy48YnI+Cjxicj4KTXkgb3JpZ2luYWwgZ29hbCB3YXMgdG8gZ2V0IHRoZSBid= WlsZCBvcmRlciBjb3JyZWN0IHNvIHRoZSBgZGVmLXNpbXBsaWZpZXJgIGNvbXBpbGVyIG1hY3Jv= cyB3ZXJlIGFsd2F5cyBkZWZpbmVkIGFoZWFkIG9mIHVzZS4mbmJzcDsgJm5ic3A7V2hhdCBJCiB= sZWFybmVkIGlzIHRoYXQgdGhlcmUgYXJlIGEgbG90IG9mIGNpcmN1bGFyIGRlcGVuZGVuY2llcy= 4mbmJzcDsgU29tZSBvZiAKdGhlc2Ugd2VyZSBmaXhlZCBidXQgbWFueSB3ZXJlIG5vdC4mbmJzc= DsgSSB0aGluayB0aGlzIHdvdWxkIHJlcXVpcmUgYSBodWdlIApyZWZhY3RvcmluZyBvZiB0aGUg= ZXhpc3RpbmcgY29kZSBtb3ZpbmcgYWxsIGtpbmRzIG9mIHRoaW5ncyBhcm91bmQuJm5ic3A7IEk= gCnN0b3BwZWQgYXQgdGhhdCBwb2ludCBiZWNhdXNlIGl0IHdhcyBhIGxvdCBvZiB3b3JrLiZuYn= NwOyBDbGF1ZGUgY291bGQgCnByb2JhYmx5IGRvIGl0IGZhaXJseSBlYXNpbHksIGJ1dCBpdCB3b= 3VsZCBtYWtlIHRoZSBmaWxlIG5hbWVzIGFuZCAKbW9kdWxlcyBhIGxvdCBsZXNzIG1lYW5pbmdm= dWwgYW5kIEkgZGlkbid0IGtub3cgaG93IHRvIGtlZXAgdGhlbSBzaW1wbGUuPGJsb2NrcXVvdGU= gdHlwZT0iY2l0ZSIgY2l0ZT0ibWlkOjA5ODg0MTgwLUQ2NzUtNDE2MC1COUVELURFOTVBMkJBRj= Y1QkBwZXRlcnBhbGwuZGUiPjxkaXYgZGlyPSJhdXRvIj48YnI+V2UgY291bGQgYXNrIGl0IHRvI= GNsZWFuIHVwIG91ciBtZXNzIHdpdGggZ2xvYmFsIHZhcmlhYmxlcyBhcyBhIGZpcnN0IHN0ZXAg= dG93YXJkcyBtdWx0aS10aHJlYWRpbmcuJm5ic3A7PGJyPjwvZGl2PjwvYmxvY2txdW90ZT5JJ3Z= lIHdvcmtlZCBvbiB0aGlzIG92ZXIgdGhlIHllYXJzIGRvaW5nIGl0IGJ5IGhhbmQuJm5ic3A7IE= kgaGF2ZW4ndCB0cmllZCB3aXRoIENsYXVkZSBidXQgaXQncyBvbiBteSBsaXN0LiZuYnNwOyAmb= mJzcDtJZiBub3RoaW5nIGVsc2UsIGl0IHdpbGwgYXQgbGVhc3QgbWFrZSBpdCBlYXNpZXIgdG8g= ZmluZCBhbGwgdGhlIHVzZXMuJm5ic3A7IEhvd2V2ZXIsIGl0J3MgYWxzbyBoYXJkIGJlY2F1c2U= gbG90cyBvZiBwbGFjZXMgdXNlIHRoZSBzYW1lIChzaG9ydCkgbmFtZSBmb3IgZGlmZmVyZW50IH= RoaW5ncy48YmxvY2txdW90ZSB0eXBlPSJjaXRlIiBjaXRlPSJtaWQ6MDk4ODQxODAtRDY3NS00M= TYwLUI5RUQtREU5NUEyQkFGNjVCQHBldGVycGFsbC5kZSI+PGRpdiBkaXI9ImF1dG8iPjxicj5X= ZSBjb3VsZCBhc2sgaXQgdG8gcGVyZm9ybSBhIHN0YWJpbGl0eSByZXZpZXcuPGJyPjxicj5JIGR= vbid0IGtub3cgaG93IGdvb2QgaXQgaXMgaW4gbWF0aHMuIEJ1dCB3aGVuIGl0IGNhbiBjaGFuZ2= Ugc3R1ZmYgc28gaXQgaXMgdGVzdGFibGUsIGFkZCB0ZXN0cyBhbmQgdGhlbiB0ZXN0IHRoZSByZ= XN1bHRzIG9mIGl0cyB3b3JrIGl0IHJlYWxseSBleGNlbHMuIE9yIHdlIGNvdWxkIGFzayBpdCB3= aGF0IG5lZWRzIHRvIGJlIGRvbmUgc28gaXQgY2FuIHBvcnQgUnViaS48L2Rpdj48L2Jsb2NrcXV= vdGU+Rm9yIHNvbWUgb3RoZXIgcHJvamVjdHMsIENsYXVkZSBoYXMgYmVlbiBxdWl0ZSBnb29kIG= F0IGNyZWF0aW5nIHRlc3RzLiZuYnNwOyBGb3IgdGhvc2UgcHJvamVjdHMsIGl0J3MgZWFzaWx5I= HZlcmlmaWFibGUgdGhhdCBpdCdzIGNvcnJlY3QuJm5ic3A7IENvdWxkIGJlIGhhcmRlciBpbiBt= YXhpbWEsIGlmLCBzYXksIENsYXVkZSBzYXlzIGhlcmUncyBhIGRlZmluaXRlIGludGVncmFsIHd= pdGggY29ycmVzcG9uZGluZyByZXN1bHQuJm5ic3A7IEl0IGNvdWxkIGJlIGRpZmZpY3VsdCB0aG= UgdmVyaWZ5IHRoYXQgdGhhdCdzIGNvcnJlY3QuJm5ic3A7IFdlIGNvdWxkIGNvbXB1dGUgaXQgb= nVtZXJpY2FsbHksIGJ1dCB0aGF0J3Mgbm90IGEgcHJvb2Ygb2YgY29ycmVjdG5lc3MuPGJyPjxi= cj48dWwgY2xhc3M9ImRyb3Bkb3duLW1lbnUgdGV4dGNvbXBsZXRlLWRyb3Bkb3duIiBzdHlsZT0= iZGlzcGxheTogbm9uZTsgcG9zaXRpb246IGFic29sdXRlOyB6LWluZGV4OiAxMDAwOyIgY29udG= VudGVkaXRhYmxlPSJmYWxzZSIgcG9wb3Zlcj0iYXV0byI+PC91bD4=3D">&#8203;</div> </div> _______________________________________________<br> Maxima-discuss mailing list<br> <a href=3D"mailto:[email protected]" target=3D"_blank" r= el=3D"noreferrer">[email protected]</a><br> <a href=3D"https://lists.sourceforge.net/lists/listinfo/maxima-discuss" rel= =3D"noreferrer noreferrer" target=3D"_blank">https://lists.sourceforge.net/= lists/listinfo/maxima-discuss</a><br> </blockquote></div></div> _______________________________________________<br> Maxima-discuss mailing list<br> <a href=3D"mailto:[email protected]" target=3D"_blank" r= el=3D"noreferrer">[email protected]</a><br> <a href=3D"https://lists.sourceforge.net/lists/listinfo/maxima-discuss" rel= =3D"noreferrer noreferrer" target=3D"_blank">https://lists.sourceforge.net/= lists/listinfo/maxima-discuss</a><br> </blockquote></div> --000000000000948c740657387440-- --===============8788453674142352802== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline --===============8788453674142352802== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Maxima-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/maxima-discuss --===============8788453674142352802==--