In reading the article some thoughts I had. 1. The example about the paper on the infinite group seems like a very weak example to me. It seems the problem is not that the programs used are proprietary, but that the author of the paper gave no code. If she used SAGE but didn't give code, then the situation wouldn't be any better. And presumably if some code had been given it could have been run and verified with another Computer algebra system open or closed. This seems to be part of a discussion on how to present results/proofs that depend on computations, but not really necessarily relevant to open vs closed software. Realistically if she gave code for one computer algebra system, and you used that to write code in another computer algebra system and got the same answer, are you really going to believe the result is false even if they are both proprietary. The example about the fictional mathematician jane seems similar to me, if jane gave the code she ran, and you were skeptical, presumably you could verify its results in another system. Unless you doubt the computer algebra systems ability to carry out each line of her code, its really only her code that you need. (this may not be a valid counterargument if no other system could do the computation she did, I am assuming that another system could perhaps less efficiently do the same thing).
2. No real discussion of the costs of the proprietary systems are given. For me, this is really the primary issue. If it was possible to buy a full (non-student) copy of matlab or mathematica for $100 and have freedom to use it on whatever computer I wanted, however I wanted, I might not necessarily feel so strongly about open source. I don't mind paying for closed source software if the cost is reasonable. But it is the fact that as a non-student it would cost thousands of dollars, and I would be restricted to a single computer, etc that makes me feel strongly about free math software. The fact that it is closed would not be as troublesome to me if it was less restrictive financially and legally. 3. I think that the mathematica quote is aimed at people using mathematica to do things mathematica already does well. Most people using mathematica, mathematicians, scientists, engineers, probably don't need to understand how it calculates bessel functions or computes integrals to get their work done and knowing these things probably wouldn't help them. Knowing how it works is really only necessary if you need to extend it beyond its current capabilities. I think this is the real point to me, it is important for software to be open not so you understand how it does what it does, but so you understand how to make it do things it doesn't do yet. This is actually talked about in the paper, but I think there is more emphasis on wanting open source software due to skepticism of correct results, than on ability to extend. Of course we are curious people so we want to know how things work as well even if we don't need to. In fact often we just want to know that we could find out how somethign worked. As an analogy, how many people using linux actually read the source code. I for one feel good knowing that I could look at the source code and I think most people feel good knowing they could look at the source. The only kernel code I have looked at was a wifi driver that I was having trouble getting to work, and if that had worked perfectly I probably would have never looked at the source. In other words, I want to know that I could look at the source, but unless it doesn't do something I want it to do or something is not working, I probably don't need to even if I want to. Just my thoughts. Josh On Aug 5, 11:37 am, "Justin C. Walker" <[EMAIL PROTECTED]> wrote: > On Aug 5, 2007, at 01:00 , Robert Bradshaw wrote: > > > > > > > On Aug 4, 2007, at 5:58 PM, William Stein wrote: > > >> On 8/4/07, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > >>> Shoot, this came in as I sent that last message. I have much less > >>> to say about > >>> this version; it's very strong. Part of me thinks that the > >>> mathematica quote > >>> is too direct an attack... but it's from the horse's mouth, and I > >>> think that the > >>> mathematical community needs to know. > > >> It is indeed *their* statement on the matter and it is what they > >> strongly believe > >> to be true. It is critical to my argument that the reader be > >> convinced that > >> current closed mathematical software is nothing like the situation > >> with math papers -- with > >> closed math software you can't just pay $25 and read the source (like > >> you can with > >> expensive journal articles. No journal would *ever* dare make a > >> statement > >> like Mathematica does about the proofs of the theorems they > >> publish. Many > >> mathematicians probably don't know this or believe it about > >> mathematical software. > >> It is very important that they realize the truth before it is too > >> late. > > > Though we all understand the principle, I don't think many people > > will (instantly) see the connection between software and proofs. I > > don't want to be to aggressive, but I think we need to state that > > point directly. I like how you worded it in this email--perhaps right > > after the quote something like. > > > No journal would *ever* dare make a statement like Mathematica does > > about the proofs of the theorems they publish, yet (here/increasingly > > commenly) the software and algorithms used are an essential part of > > (Jane's/the) proof. > > I really like your argument here. I think it is precisely the issue > when discussing the use of closed source/proprietary software in > mathematics: "a theorem without proof is as useful as software > without code". Perhaps remove 'dare' :-} > > Justin > > -- > Justin C. Walker, Curmudgeon-at-Large > () The ASCII Ribbon Campaign > /\ Help Cure HTML Email --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
