Re: 6 months free Claude Max 20x for working on Maxima
David Scherfgen via Maxima-discuss <[email protected]> Mon, 20 Jul 2026 15:36:29 +0200
| Newsgroups | gmane.comp.mathematics.maxima.general |
|---|---|
| Message-ID | <CAMTHLKgLYX3LHSxhGtm_FCXaQutoJtpzCM29pM48AEpymNzWxA@mail.gmail.com> |
--===============3984330098473409032== Content-Type: multipart/alternative; boundary="0000000000009d3d5306570b0022" --0000000000009d3d5306570b0022 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable I agree with you, Viktor, and thanks for the words of caution. I'm not letting myself get pushed out of the loop. I explicitly instructed Claude to not try to run any code on its own or make modifications without my consent. The new "Fable" model seems very capable. I've thrown entire Maxima modules at it with the simple task to find potential bugs, and it found lots of them, no hallucinations whatsoever. Am So., 19. Juli 2026 um 19:40 Uhr schrieb Viktor T. Toth <[email protected]= m >: > It is indeed awesome that Anthropic does this but allow me to offer, from > experience, words of caution when it comes to "agentic" use of LLMs in > development work. > > I've been using LLMs actively as a tool since early 2023. The productivit= y > boost is unbelievable. I was able to complete things in hours that > previously would have taken days or weeks, with a result much better in > quality than anything I'd have created on my own. > > At first, there was of course only one way of doing things: ask the AI to > write code, add it to your project, test it, fix it or ask the AI to fix > it. Or maybe ask the AI to analyze existing code, especially code with > known bugs, and offer suggestions to fix or improve it. > > But then, coding agents arrived automating much of that. The crucial > difference: in conventional prompting, it is I, the human, who decide -- > after reviewing the code, inserting it into the project, testing it -- > whether or not the code delivers the expected result or behavior. In the > agentic model, the human is mostly removed from this low-level loop: the > agent writes the code, tests the code in a sandbox, evaluates and revises > it without human intervention. The human only sees the log and the final > result presented for approval. The process, though not completely opaque, > becomes far less transparent. > > This, in my view, is a very dangerous slippery slope. Without the mental > history of active participation, it is impossible to "own" the resulting > code. The human is no longer acting as a senior developer with LLMs as > highly capable junior helpers, but more like a project manager who > evaluates project success not on the basis of actual, hands-on experience > but on the basis of KPIs on PowerPoint slides at the weekly meeting. The > difference, of course, is that the senior developers working for that > project manager are themselves experienced. LLMs are not "experienced". > They have incredible book knowledge, but ultimately that's all they have: > their so-called "hallucinations" arise because for an LLM, words exist on= ly > in relationship with other words, not lived experience. And all it takes = is > a slightly misguided, overly enthusiastic prompt, for instance, for the L= LM > to invent a bug where none exists, because it seems like a statistically > more satisfactory response to the prompt than simply confirming that a > piece of code is valid and requires no change or refactoring. > > Long story short, I tried coding agents and rejected that approach myself= , > at least for now. They automate the part of my work that needs automation > the least, at the price of removing me from the very loop that would allo= w > me to retain mental ownership of the code that bears my name. Especially = in > the case of a mature project like Maxima, with decades-long legacy, writt= en > in a nontrivial programming language, implementing its own language with > notable idiosyncrasies, and relying on sophisticated mathematics, I worry > that the use of LLMs as coding agents, as opposed to coding assistants, > will lead to trouble. > > > Viktor > > > > > On 2026-07-19 10:57, Raymond Toy wrote: > > On 7/18/26 11:49 PM, David Scherfgen via Maxima-discuss wrote: > > Dear all, > > Anthropic (the company behind the Claude AI) launched a programme > <https://claude.com/contact-sales/claude-for-oss> that allows Open Source > developers/contributors to apply for 6 months free Claude Max 20x, which > normally costs $200 per month. > > A friend recommended me to apply for it (I didn't know about it), citing > my recent work on Maxima, and indeed I got selected! To those > > Awesome! > > who've made significant contributions (see terms > <https://www.anthropic.com/claude-for-oss-terms>): I encourage you to > apply, too. In my application, I mentioned that Maxima lives on SourceFor= ge > and selected the Maxima GitHub mirror <https://github.com/calyau/maxima> > as the project. If you like, let me know before you apply, then I can > update the GitHub mirror so that the most recent commits are visible ther= e, > if that helps you. > > It=E2=80=99s a pain that that can=E2=80=99t be automated AFAIK. I have so= me projects on > gitlab that get automatically mirrored to github whenever I check in > anything into gitlab. (This is mostly for a kind of backup. And also for > Claude since gitlab is my canonical repo.) > > > I'm still getting used to Claude, but I can already say that it's > incredibly powerful and far superior to my previous Gemini-based workflow= . > The best part is that it can access code (read & write) and execute > commands (e.g. run Maxima to test things) locally on your computer. I'm > still careful with that, though. > > I=E2=80=99ve been using Claude for a while. It works quite well for me. B= ut do be > careful. It was hallucinating problems and solutions for me, making up > things that didn=E2=80=99t actually exist. I eventually added a rule that= it > shouldn=E2=80=99t guess but actually look at the source code and nothing = else > unless I ask for it. > > I also only use the browser version. I didn=E2=80=99t want to give it acc= ess to > any of my actual machines or to modify any code. This works fine except I > have to download and apply suggested changes. But I since I have gitlab > mirrored, it can fetch the latest github code to see what I did and and > what=E2=80=99s next. > > > Do you have suggestions for new features that I might try to tackle durin= g > these 6 months? Of course I can also keep fixing bugs, or both. > > Both, of course! But it=E2=80=99s up to you. > > There are personal projects I wanted to do but haven=E2=80=99t really mad= e any > progress. One big one was support for theta functions. A long time ago, I > (or this list?) got a screenshot from Bill Gosper (yes, that Bill) about > the work he had done with theta functions. Really impressive. Even had th= e > code, but so much of was in lisp on real Macsyma so I never figured out > what those functions actually did. I think he also sent some nice > algorithms for computing 2F1 over the entire plane. Barton has solved thi= s > using a different approach, I think. > ​ > > > _______________________________________________ > Maxima-discuss mailing [email protected]://li= sts.sourceforge.net/lists/listinfo/maxima-discuss > > _______________________________________________ > Maxima-discuss mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/maxima-discuss > --0000000000009d3d5306570b0022 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr">I agree with you, Viktor, and thanks for the words of caut= ion.<div>I'm not letting myself get pushed out of the loop. I explicitl= y instructed Claude to not try to run any code on its own or make modificat= ions without my consent.</div><div>The new "Fable" model seems ve= ry capable. I've thrown entire Maxima modules at it with the simple tas= k to find potential bugs, and it found lots of them, no hallucinations what= soever.</div></div><br><div class=3D"gmail_quote gmail_quote_container"><di= v dir=3D"ltr" class=3D"gmail_attr">Am So., 19. Juli 2026 um 19:40=C2=A0Uhr = schrieb Viktor T. Toth <<a href=3D"mailto:[email protected]">vttoth@vtto= th.com</a>>:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:= 0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">= <u></u> =20 =20 =20 <div> <p><font face=3D"monospace">It is indeed awesome that Anthropic does this but allow me to offer, from experience, words of caution when it comes to "agentic" use of LLMs in development wor= k.</font></p> <p><font face=3D"monospace">I've been using LLMs actively as a tool since early 2023. The productivity boost is unbelievable. I was able to complete things in hours that previously would have taken days or weeks, with a result much better in quality than anything I'd have created on my own.</font></p> <p><font face=3D"monospace">At first, there was of course only one way of doing things: ask the AI to write code, add it to your project, test it, fix it or ask the AI to fix it. Or maybe ask the AI to analyze existing code, especially code with known bugs, and offer suggestions to fix or improve it.</font></p> <p><font face=3D"monospace">But then, coding agents arrived automating much of that. The crucial difference: in conventional prompting, it is I, the human, who decide -- after reviewing the code, inserting it into the project, testing it -- whether or not the code delivers the expected result or behavior. In the agentic model, the human is mostly removed from this low-level loop: the agent writes the code, tests the code in a sandbox, evaluates and revises it without human intervention. The human only sees the log and the final result presented for approval. The process, though not completely opaque, becomes far less transparent.</font></p> <p><font face=3D"monospace">This, in my view, is a very dangerous slippery slope. Without the mental history of active participation, it is impossible to "own" the resulting co= de. The human is no longer acting as a senior developer with LLMs as highly capable junior helpers, but more like a project manager who evaluates project success not on the basis of actual, hands-on experience but on the basis of KPIs on PowerPoint slides at the weekly meeting. The difference, of course, is that the senior developers working for that project manager are themselves experienced. LLMs are not "experienced". They = have incredible book knowledge, but ultimately that's all they have: their so-called "hallucinations" arise because for an LLM= , words exist only in relationship with other words, not lived experience. And all it takes is a slightly misguided, overly enthusiastic prompt, for instance, for the LLM to invent a bug where none exists, because it seems like a statistically more satisfactory response to the prompt than simply confirming that a piece of code is valid and requires no change or refactoring.</fo= nt></p> <p><font face=3D"monospace">Long story short, I tried coding agents and rejected that approach myself, at least for now. They automate the part of my work that needs automation the least, at the price of removing me from the very loop that would allow me to retain mental ownership of the code that bears my name. Especially in the case of a mature project like Maxima, with decades-long legacy, written in a nontrivial programming language, implementing its own language with notable idiosyncrasies, and relying on sophisticated mathematics, I worry that the use of LLMs as coding agents, as opposed to coding assistants, will lead to trouble.</font></p> <p><font face=3D"monospace"><br> </font></p> <p><font face=3D"monospace">Viktor</font></p> <p><font face=3D"monospace"><br> </font></p> <p><font face=3D"monospace"><br> </font></p> <p><font face=3D"monospace"><br> </font></p> <div>On 2026-07-19 10:57, Raymond Toy wrote:<br> </div> <blockquote type=3D"cite"> =20 =20 =20 <div> <p style=3D"margin-bottom:19.2px;margin-top:0px">On 7/18/26 11:49 PM, David Scherfgen via Maxima-discuss wrote:<br> </p> <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rg= b(119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;bord= er-left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,11= 9);margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;paddi= ng-right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-empha= sis-color:rgb(119,119,119)"> <div dir=3D"ltr">Dear all, <div><br> </div> <div>Anthropic (the company behind the Claude AI) launched a <a href=3D"https://claude.com/contact-sales/claude-for-oss" t= arget=3D"_blank">programme</a>=C2=A0that allows Open Source developers/contributors to apply for 6 months free Claude Max 20x, which normally costs $200 per month.</div> <div><br> </div> <div>A friend recommended me to apply for it (I didn't know about it), citing my recent work on Maxima, and indeed I got selected! To those</div> </div> </blockquote> Awesome! <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rg= b(119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;bord= er-left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,11= 9);margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;paddi= ng-right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-empha= sis-color:rgb(119,119,119)"> <div dir=3D"ltr"> <div>=C2=A0who've made significant contributions (see <a hr= ef=3D"https://www.anthropic.com/claude-for-oss-terms" target=3D"_blank">ter= ms</a>): I encourage you to apply, too. In my application, I mentioned that Maxima lives on SourceForge and selected the <a href=3D"https://gith= ub.com/calyau/maxima" target=3D"_blank">Maxima GitHub mirror</a> as the project. If you like, let=C2=A0me know before you apply, then= I can update the GitHub mirror so that the most recent commits are visible there, if that helps you.<br> </div> </div> </blockquote> It=E2=80=99s a pain that that can=E2=80=99t be automated AFAIK. I h= ave some projects on gitlab that get automatically mirrored to github whenever I check in anything into gitlab. (This is mostly for a kind of backup. And also for Claude since gitlab is my canonical repo.) <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rg= b(119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;bord= er-left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,11= 9);margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;paddi= ng-right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-empha= sis-color:rgb(119,119,119)"> <div dir=3D"ltr"> <div><br> </div> <div>I'm still getting used to Claude, but I can already sa= y that it's incredibly powerful and far superior to my previous Gemini-based workflow. The best part is that it can access code (read & write) and execute commands (e.g. run Maxima to test things) locally on your computer. I'm still careful with that, though.</div> </div> </blockquote> I=E2=80=99ve been using Claude for a while. It works quite well for= me. But do be careful. It was hallucinating problems and solutions for me, making up things that didn=E2=80=99t actually exist. I eventually added a rule that it shouldn=E2=80=99t guess but actuall= y look at the source code and nothing else unless I ask for it. <p style=3D"margin-bottom:19.2px;margin-top:0px">I also only use the browser version. I didn=E2=80=99t want to give it access = to any of my actual machines or to modify any code. This works fine except I have to download and apply suggested changes. But I since I have gitlab mirrored, it can fetch the latest github code to see what I did and and what=E2=80=99s next.</p> <blockquote type=3D"cite" style=3D"border-color:rgb(119,119,119) rg= b(119,119,119) rgb(119,119,119) rgb(114,159,207);border-left-width:2px;bord= er-left-style:solid;color:rgb(119,119,119);column-rule-color:rgb(119,119,11= 9);margin:19.2px 0px;outline-color:rgb(119,119,119);padding-left:16px;paddi= ng-right:16px;quotes:none;text-decoration-color:rgb(119,119,119);text-empha= sis-color:rgb(119,119,119)"> <div dir=3D"ltr"> <div><br> </div> <div>Do you have suggestions for new features that I might try to tackle during these 6 months? Of course I can also keep fixing bugs,=C2=A0or both.</div> </div> </blockquote> Both, of course! But it=E2=80=99s up to you. <p style=3D"margin-bottom:19.2px;margin-top:0px">There are personal projects I wanted to do but haven=E2=80=99t really made = any progress. One big one was support for theta functions. A long time ago, I (or this list?) got a screenshot from Bill Gosper (yes, that Bill) about the work he had done with theta functions. Really impressive. Even had the code, but so much of was in lisp on real Macsyma so I never figured out what those functions actually did. I think he also sent some nice algorithms for computing 2F1 over the entire plane. Barton has solved this using a different approach, I think.</p> </div> <div style=3D"height:0px;width:0px;max-height:0px;max-width:0px;overf= low:hidden;font-size:0px;padding:0px;margin:0px" aria-hidden=3D"true" title= =3D"MDH:PGJyPjxicj48ZGl2IGNsYXNzPSJtb3otY2l0ZS1wcmVmaXgiPk9uIDcvMTgvMjYgMTE= 6NDkgUE0sIERhdmlkIFNjaGVyZmdlbiB2aWEgTWF4aW1hLWRpc2N1c3Mgd3JvdGU6PGJyPjwvZG= l2PjxibG9ja3F1b3RlIHR5cGU9ImNpdGUiIGNpdGU9Im1pZDpDQU1USExLaGZ5b0V1NUZPWEd6M= mVzTzRhbitiaDVIdENLRVJNcWFDa3dPXzBBZD11a3dAbWFpbC5nbWFpbC5jb20iPgo8bWV0YSBo= dHRwLWVxdWl2PSJjb250ZW50LXR5cGUiIGNvbnRlbnQ9InRleHQvaHRtbDsgIj48ZGl2IGRpcj0= ibHRyIj5EZWFyIGFsbCw8ZGl2Pjxicj48L2Rpdj48ZGl2PkFudGhyb3BpYyAodGhlIGNvbXBhbn= kgYmVoaW5kIHRoZSBDbGF1ZGUgQUkpIGxhdW5jaGVkIGEgPGEgaHJlZj0iaHR0cHM6Ly9jbGF1Z= GUuY29tL2NvbnRhY3Qtc2FsZXMvY2xhdWRlLWZvci1vc3MiIG1vei1kby1ub3Qtc2VuZD0idHJ1= ZSI+cHJvZ3JhbW1lPC9hPiZuYnNwO3RoYXQgYWxsb3dzIE9wZW4gU291cmNlIGRldmVsb3BlcnM= vY29udHJpYnV0b3JzIHRvIGFwcGx5IGZvciA2IG1vbnRocyBmcmVlIENsYXVkZSBNYXggMjB4LC= B3aGljaCBub3JtYWxseSBjb3N0cyAkMjAwIHBlciBtb250aC48L2Rpdj48ZGl2Pjxicj48L2Rpd= j48ZGl2PkEgZnJpZW5kIHJlY29tbWVuZGVkIG1lIHRvIGFwcGx5IGZvciBpdCAoSSBkaWRuJ3Qg= a25vdyBhYm91dCBpdCksIGNpdGluZyBteSByZWNlbnQgd29yayBvbiBNYXhpbWEsIGFuZCBpbmR= lZWQgSSBnb3Qgc2VsZWN0ZWQhIFRvIHRob3NlPC9kaXY+PC9kaXY+PC9ibG9ja3F1b3RlPkF3ZX= NvbWUhPGJsb2NrcXVvdGUgdHlwZT0iY2l0ZSIgY2l0ZT0ibWlkOkNBTVRITEtoZnlvRXU1Rk9YR= 3oyZXNPNGFuK2JoNUh0Q0tFUk1xYUNrd09fMEFkPXVrd0BtYWlsLmdtYWlsLmNvbSI+PGRpdiBk= aXI9Imx0ciI+PGRpdj4mbmJzcDt3aG8ndmUgbWFkZSBzaWduaWZpY2FudCBjb250cmlidXRpb25= zIChzZWUgPGEgaHJlZj0iaHR0cHM6Ly93d3cuYW50aHJvcGljLmNvbS9jbGF1ZGUtZm9yLW9zcy= 10ZXJtcyIgbW96LWRvLW5vdC1zZW5kPSJ0cnVlIj50ZXJtczwvYT4pOiBJIGVuY291cmFnZSB5b= 3UgdG8gYXBwbHksIHRvby4gSW4gbXkgYXBwbGljYXRpb24sIEkgbWVudGlvbmVkIHRoYXQgTWF4= aW1hIGxpdmVzIG9uIFNvdXJjZUZvcmdlIGFuZCBzZWxlY3RlZCB0aGUgPGEgaHJlZj0iaHR0cHM= 6Ly9naXRodWIuY29tL2NhbHlhdS9tYXhpbWEiIG1vei1kby1ub3Qtc2VuZD0idHJ1ZSI+TWF4aW= 1hIEdpdEh1YiBtaXJyb3I8L2E+IGFzIHRoZSBwcm9qZWN0LiBJZiB5b3UgbGlrZSwgbGV0Jm5ic= 3A7bWUga25vdyBiZWZvcmUgeW91IGFwcGx5LCB0aGVuIEkgY2FuIHVwZGF0ZSB0aGUgR2l0SHVi= IG1pcnJvciBzbyB0aGF0IHRoZSBtb3N0IHJlY2VudCBjb21taXRzIGFyZSB2aXNpYmxlIHRoZXJ= lLCBpZiB0aGF0IGhlbHBzIHlvdS48YnI+PC9kaXY+PC9kaXY+PC9ibG9ja3F1b3RlPkl0J3MgYS= BwYWluIHRoYXQgdGhhdCBjYW4ndCBiZSBhdXRvbWF0ZWQgQUZBSUsuJm5ic3A7IEkgaGF2ZSBzb= 21lIHByb2plY3RzIG9uIGdpdGxhYiB0aGF0IGdldCBhdXRvbWF0aWNhbGx5IG1pcnJvcmVkJm5i= c3A7IHRvIGdpdGh1YiB3aGVuZXZlciBJIGNoZWNrIGluIGFueXRoaW5nIGludG8gZ2l0bGFiLiZ= uYnNwOyAoVGhpcyBpcyBtb3N0bHkgZm9yIGEga2luZCBvZiBiYWNrdXAuJm5ic3A7IEFuZCBhbH= NvIGZvciBDbGF1ZGUgc2luY2UgZ2l0bGFiIGlzIG15IGNhbm9uaWNhbCByZXBvLik8YmxvY2txd= W90ZSB0eXBlPSJjaXRlIiBjaXRlPSJtaWQ6Q0FNVEhMS2hmeW9FdTVGT1hHejJlc080YW4rYmg1= SHRDS0VSTXFhQ2t3T18wQWQ9dWt3QG1haWwuZ21haWwuY29tIj48ZGl2IGRpcj0ibHRyIj48ZGl= 2Pjxicj48L2Rpdj48ZGl2PkknbSBzdGlsbCBnZXR0aW5nIHVzZWQgdG8gQ2xhdWRlLCBidXQgSS= BjYW4gYWxyZWFkeSBzYXkgdGhhdCBpdCdzIGluY3JlZGlibHkgcG93ZXJmdWwgYW5kIGZhciBzd= XBlcmlvciB0byBteSBwcmV2aW91cyBHZW1pbmktYmFzZWQgd29ya2Zsb3cuIFRoZSBiZXN0IHBh= cnQgaXMgdGhhdCBpdCBjYW4gYWNjZXNzIGNvZGUgKHJlYWQgJmFtcDsgd3JpdGUpIGFuZCBleGV= jdXRlIGNvbW1hbmRzIChlLmcuIHJ1biBNYXhpbWEgdG8gdGVzdCB0aGluZ3MpIGxvY2FsbHkgb2= 4geW91ciBjb21wdXRlci4gSSdtIHN0aWxsIGNhcmVmdWwgd2l0aCB0aGF0LCB0aG91Z2guPC9ka= XY+PC9kaXY+PC9ibG9ja3F1b3RlPkkndmUgYmVlbiB1c2luZyBDbGF1ZGUgZm9yIGEgd2hpbGUu= Jm5ic3A7IEl0IHdvcmtzIHF1aXRlIHdlbGwgZm9yIG1lLiZuYnNwOyBCdXQgZG8gYmUgY2FyZWZ= 1bC4mbmJzcDsgSXQgd2FzIGhhbGx1Y2luYXRpbmcgcHJvYmxlbXMgYW5kIHNvbHV0aW9ucyBmb3= IgbWUsIG1ha2luZyB1cCB0aGluZ3MgdGhhdCBkaWRuJ3QgYWN0dWFsbHkgZXhpc3QuJm5ic3A7I= EkgZXZlbnR1YWxseSBhZGRlZCBhIHJ1bGUgdGhhdCBpdCBzaG91bGRuJ3QgZ3Vlc3MgYnV0IGFj= dHVhbGx5IGxvb2sgYXQgdGhlIHNvdXJjZSBjb2RlIGFuZCBub3RoaW5nIGVsc2UgdW5sZXNzIEk= gYXNrIGZvciBpdC48YnI+PGJyPkkgYWxzbyBvbmx5IHVzZSB0aGUgYnJvd3NlciB2ZXJzaW9uLi= ZuYnNwOyBJIGRpZG4ndCB3YW50IHRvIGdpdmUgaXQgYWNjZXNzIHRvIGFueSBvZiBteSBhY3R1Y= WwgbWFjaGluZXMgb3IgdG8gbW9kaWZ5IGFueSBjb2RlLiZuYnNwOyBUaGlzIHdvcmtzIGZpbmUg= ZXhjZXB0IEkgaGF2ZSB0byBkb3dubG9hZCBhbmQgYXBwbHkgc3VnZ2VzdGVkIGNoYW5nZXMuJm5= ic3A7IEJ1dCBJIHNpbmNlIEkgaGF2ZSBnaXRsYWIgbWlycm9yZWQsIGl0IGNhbiBmZXRjaCB0aG= UgbGF0ZXN0IGdpdGh1YiBjb2RlIHRvIHNlZSB3aGF0IEkgZGlkIGFuZCBhbmQgd2hhdCdzIG5le= HQuPGJsb2NrcXVvdGUgdHlwZT0iY2l0ZSIgY2l0ZT0ibWlkOkNBTVRITEtoZnlvRXU1Rk9YR3oy= ZXNPNGFuK2JoNUh0Q0tFUk1xYUNrd09fMEFkPXVrd0BtYWlsLmdtYWlsLmNvbSI+PGRpdiBkaXI= 9Imx0ciI+PGRpdj48YnI+PC9kaXY+PGRpdj5EbyB5b3UgaGF2ZSBzdWdnZXN0aW9ucyBmb3Igbm= V3IGZlYXR1cmVzIHRoYXQgSSBtaWdodCB0cnkgdG8gdGFja2xlIGR1cmluZyB0aGVzZSA2IG1vb= nRocz8gT2YgY291cnNlIEkgY2FuIGFsc28ga2VlcCBmaXhpbmcgYnVncywmbmJzcDtvciBib3Ro= LjwvZGl2PjwvZGl2PjwvYmxvY2txdW90ZT5Cb3RoLCBvZiBjb3Vyc2UhJm5ic3A7IEJ1dCBpdCd= zIHVwIHRvIHlvdS4mbmJzcDsmbmJzcDs8YnI+PGJyPlRoZXJlIGFyZSBwZXJzb25hbCBwcm9qZW= N0cyBJIHdhbnRlZCB0byBkbyBidXQgaGF2ZW4ndCByZWFsbHkgbWFkZSBhbnkgcHJvZ3Jlc3MuJ= m5ic3A7IE9uZSBiaWcgb25lIHdhcyBzdXBwb3J0IGZvciB0aGV0YSBmdW5jdGlvbnMuJm5ic3A7= IEEgbG9uZyB0aW1lIGFnbywgSSAob3IgdGhpcyBsaXN0PykgZ290IGEgc2NyZWVuc2hvdCBmcm9= tIEJpbGwgR29zcGVyICh5ZXMsIHRoYXQgQmlsbCkgYWJvdXQgdGhlIHdvcmsgaGUgaGFkIGRvbm= Ugd2l0aCB0aGV0YSBmdW5jdGlvbnMuJm5ic3A7IFJlYWxseSBpbXByZXNzaXZlLiZuYnNwOyBFd= mVuIGhhZCB0aGUgY29kZSwgYnV0IHNvIG11Y2ggb2Ygd2FzIGluIGxpc3Agb24gcmVhbCBNYWNz= eW1hIHNvIEkgbmV2ZXIgZmlndXJlZCBvdXQgd2hhdCB0aG9zZSBmdW5jdGlvbnMgYWN0dWFsbHk= gZGlkLiZuYnNwOyBJIHRoaW5rIGhlIGFsc28gc2VudCBzb21lIG5pY2UgYWxnb3JpdGhtcyBmb3= IgY29tcHV0aW5nIDJGMSBvdmVyIHRoZSBlbnRpcmUgcGxhbmUuJm5ic3A7IEJhcnRvbiBoYXMgc= 29sdmVkIHRoaXMgdXNpbmcgYSBkaWZmZXJlbnQgYXBwcm9hY2gsIEkgdGhpbmsuPHVsIGNsYXNz= PSJkcm9wZG93bi1tZW51IHRleHRjb21wbGV0ZS1kcm9wZG93biIgc3R5bGU9ImRpc3BsYXk6IG5= vbmU7IHBvc2l0aW9uOiBhYnNvbHV0ZTsgei1pbmRleDogMTAwMDsiIGNvbnRlbnRlZGl0YWJsZT= 0iZmFsc2UiIHBvcG92ZXI9ImF1dG8iPjwvdWw+">&#8203;</div> <br> <fieldset></fieldset> <br> <fieldset></fieldset> <pre>_______________________________________________ Maxima-discuss mailing list <a href=3D"mailto:[email protected]" target=3D"_blank">M= [email protected]</a> <a href=3D"https://lists.sourceforge.net/lists/listinfo/maxima-discuss" tar= get=3D"_blank">https://lists.sourceforge.net/lists/listinfo/maxima-discuss<= /a> </pre> </blockquote> </div> _______________________________________________<br> Maxima-discuss mailing list<br> <a href=3D"mailto:[email protected]" target=3D"_blank">M= [email protected]</a><br> <a href=3D"https://lists.sourceforge.net/lists/listinfo/maxima-discuss" rel= =3D"noreferrer" target=3D"_blank">https://lists.sourceforge.net/lists/listi= nfo/maxima-discuss</a><br> </blockquote></div> --0000000000009d3d5306570b0022-- --===============3984330098473409032== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline --===============3984330098473409032== 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 --===============3984330098473409032==--