On 3/7/07, Karl Crisman <[EMAIL PROTECTED]> wrote: > Dear Prof. Stein,
> This is the "completely optional" email to let you know of a SAGE download > and use. I have been using SAGE (2.2?) for about two months on a Mac OSX.4 > Powerbook G4. The only major problem I have to report is that the notebook > server is extremely slow (which seems to be a frequent topic on the lists I > browsed) and that it usually hangs and then stops altogether, even if I use > "Interrupt" and then type (for instance) 1+1 and try to evaluate. > I imagine > this is a fairly universal concern, but I won't hold my breath, since it > probably has to do with my machine more than anything else - not exactly a > mainframe. Often the people that find the notebook to be slow are those using Powerbook G4's or 600 Mhz Linux machines. The notebook is reasonably usable on a modern intel-based laptop, especially if it has good RAM. I would like to figure out why it is so slow on Powerbooks. In your case it could have something to do with memory requirements -- you're running OSX 10.4, whiich in my experience requires at least 1GB of RAM to work well -- maybe your laptop has < 1GB RAM? There could be some other reasons why the notebook feels slow, e.g., there are some parameters that could be tuned that determine how frequently the notebook and server talk to each other, which might be relevant. I'm not suggesting you do anything here, just that I am concerned about this problem. > It seems your group is trying to get a sense of who is using SAGE, so I'll > give you a snapshot below. This became much longer than I intended, but We are. Many thanks. > perhaps it will be useful for you as the development of SAGE continues, > especially if it truly does pursue a growing edge for non-research > applications, which I believe could have a very profound impact (along with > similar projects) on education. I think SAGE will continue in this direction. My personal primary focus for my work on SAGE is going to be mainly improvements related to my number theory research. However, there are numerous students who work on SAGE employed by me or as part of projects, and they are often intensely interested in educational applications. One student For example, Bobby Moretti, has done a huge amount of work on improving SAGE's symbolic calculus and manipulation capabilities -- this will be rolled out in SAGE in probably 2-3 weeks, hopefully. In any case, I am going to continue to encourage student work on educational applications of SAGE. > Thank you (and your collaborators, and the NSF) for making this tool openly > available to all, even my students. > > Sincerely, > Karl-Dieter Crisman > Asst. Prof. of Mathematics > I am a professor at a small Christian liberal arts college near Boston, and > have been looking for appropriate software for this term's Number Theory > course for over a year now. I mean appropriate ala your contributor David > Joyner's phrase "insanely easy", and in the sense of the quote below from > the lists: > > "Looking at the actual problems that people might want to do is a good way > to get started. But again, I'd caution against taking too much of an ad-hoc > approach. I think that calculus is a great place to start. First of all, we > do have Maxima. Also, having an easy-to-use calculus package that displays > its output prettily is a must for any self-respecting CAS. With jsMath and > Maxima, we have a real potential to blow away the competition here. And > finally, calculus is kind of a lowest common denominator among people who > are using a CAS. It's probably the first thing that any non-mathematician > would try... first impressions are important." > > Perhaps something similar could be said of number theory, which really > should be accessible (and taught to) people who would never dream of taking > calculus. Or, in my case, math, physics, and CS majors of a wide range of I, of course, totally agree! > ability when it comes to interpreting/coming up with computer code. First > impressions are mixed but overall favorable. As it seemed allowed, I have > set up a page (which I periodically change for different activities) on your > server for this class at > http://sage.math.washington.edu:8101/MA338 > which has been slowly getting more and more used as the semester goes on. Yep, this is certainly welcome. Note that currently if multiple people use the page at the same time very bad things will happen. Their individual calculations might seem to work, but things will get out of sync. It's best if people make their own pages, e.g., http://sage.math.washington.edu:8101/MA338_joe It's actually "insanely easy" to make a new page -- just visit a new URL and the page is automatically created. > Very nice, though for insane ease there is a strong need for more > computer-phobe-friendly documentation. > > I want to thank you and your community very, very much for developing a > tool which comes much, much closer to what I had envisioned than any of the > other free or non-free options out there. I sometimes am amazed at the > paradigm shift the open-source movement has brought - a long way from the > closely guarded secret of the cubic's solution! But I know that for many of > my students, this is the only realistic way they are going to be able to do > experimentation on the computer. You're welcome. And I agree that the likelihood of students playing around a little with the online SAGE notebook is higher than the chances they'll install PARI, MAxima,Mathemtaica, etc., and really play around with them. I think SAGE still has a ways to go, but it is starting to feel to me like it is finally getting there, after over two years of very hard work. > For a mathematics computer program to work in a wide range of courses at a > typical smaller college which hasn't already invested heavily in the very > expensive M's, it has to have/be > 1. A "webby" interactive GUI interface with almost zero learning curve and > intuitive controls Yep. And my understanding is that bizarrely, even if you do invest in the expensive M's, you still don't get (1), since they are somewhat restrictive in how their web interfaces can be used. > 2. Easily accessible on most campus machines, or in one whole lab at least Makes sense. > 3. As close to free as possible, because of both institutional and student > (at least many of their) budgets That's definitely the reality of the situation. > 4. A lot implemented directly (e.g. SAGE's prime_range()) so the instructor > realistically has time to prepare for its use each class period I assume you just mean there is a huge library of useful mathematical functionality? I.e., the instructor doesn't have to write a prime_range function. > I should point out that the Mac OSX.4 Grapher program (in Utilities) also > satisfies these requirements for a fair number of calculus items, at least > if a school already has an easily accessible Mac lab; I use it for It is web accessible? Maybe it doesn't matter, since you all have Mac's. > multivariable calculus. However, it's not a CAS. If SAGE could have a > graphing/plotting component with the totally intuitive syntax and > (relatively) fast graphing, manipulation, etc. and options of Grapher, there > could be no competitor. Grapher is simply that easy to use. (Its > documentation is stygian, though.) We also have a faculty member who uses > Octave for teaching linear algebra, so there is precedent (and support) for > this sort of thing on our campus, which is always nice. That's good to here. SAGE's plotting is nice in some ways, but fundamentally constrained in others since SAGE works via a web browser, and (currently) doesn't use Java at all. We do plan to have a 3d java based graphing system, but it's unclear how long that will take. > There are still certainly issues I will be interested in learning more > about. Some are very technical, like how to get Python's Timer.timeit() to > deal with functions which call other SAGE functions, like today when I tried > to get it to time a trial division function I wrote but it couldn't find the > floor function; despite a lot of searching, neither Python nor SAGE > documentation seemed to cover this. (This was intended to compare different > factorization methods.) Unfortunately I don't really understand the question. Some potentially useful comments: (1) If you do sage: %prun a_command... then you'll get a profile of what functions were called in the course of running the command, how many times, how long, etc. (2) If you do sage: search_sage('floor') you'll see every line in the SAGE core library source code that includes the word line in it. > Some are very simple but important for user > friendliness; for instance, where did the little arrows below the the cells > go? That one threw my students for a loop, since in class I used the > arrows! Good question. The arrows vanishing is a mistake on my part, which I need to fix. > Possibly there should be a huge thing on each worksheet saying, use > "Shift-Enter" (and please don't use Internet Explorer). You're right -- this should be much clearer. > But in general each > time I use it, alone or in class, it works out better. > I am encouraging my students who are computer-capable to try out their own > things, and one in particular (who is doing a project for his Numerical > Analysis course re-implementing a number of Python math functions) is very > excited about the possibility of contributing in some way at some point. There are tons of possibilities for contribution if somebody is seriously interested. > Obviously Number Theory is the strong suit of SAGE proper, particularly via > PARI, but it does seem like there are some low-level things which could be > implemented (ala some of Keith Matthews' > http://www.numbertheory.org/gnubc/ bc number theory > programs, or even simpler things), so I hope he will think about it again > after the semester. He has also downloaded SAGE and used it a fair bit. Yep. > I hope this gives a good "case-study" sense of the kind of people starting > to experiment with using SAGE. I certainly intend to keep using it, and > perhaps to also contribute something someday as well, though I have no idea > what. Thanks again. Thanks! -- William Stein Associate Professor of Mathematics University of Washington --~--~---------~--~----~------------~-------~--~----~ 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/ -~----------~----~----~----~------~----~------~--~---
