Maple Vr4 Student Edition for the Mac Reviewed by Grant Keady, University of Western Australia (e-mail: keady@maths.uwa.edu.au) It needs a minimum of 8Meg RAM on your Mac (which is twice that the previous release needed). It works very well. It costs about 70 pounds. Read on if you need, or want, a longer review. The version used was 5.4.0b and interface(patchlevel); returned 1. No effort was made to check out if later patchlevels have been made. There is another review of Maple Vr4 - the full version, for PC - to appear in this Newsletter. Hence, I will, after a brief general introduction, focus my remarks mostly on matters which are either Mac specific or relate to the Student Edition. I will also assume that the readership of this newsletter is generally familiar with Maple, Mathematica, Mathcad. See the CTI book on Mathematical Software for older references and Web addresses. Maple is at http://www.maplesoft.com Student editions of CAS (Computer Algebra Systems) running on students' own computers, will be increasingly important. Already some textbook publishers are providing accompaniments (on disk, CD-ROM, or Web) in CAS - at least for the 3 Ma- packages above. A slightly dated, but generally accurate, account is given in [1]. The guidelines provided by the publishers specify that materials must run on student editions. The integration of word-processing capabilities, hyperlinks and so on, into these CAS is part of this process of integrating CAS into smoother exposition. It won't just be for maths texts: engineers will look up their "handbooks" electronically. See [2]. Reference [2] was written in Mathematica 3, but its general message is relevant to Maple Vr4 as another CAS with Sectioning, word-processing and hyperlinks. GENERAL COMMENTS ON Vr4 The review copy is from Clecom in Birmingham. I was pleased to note that Vr4 behaves nicely on the test of "assume" that I and others in the audience at Clecom's Dec 92 launch of Maple Vr2 posed to Vr2: assume(n,integer); sin(n*Pi); cos(n*Pi); Sadly sin(n*Pi/2) was left unchanged, though I expect there is some way for users to add this functionality. With a bit more work, we would have a tool of use in engineering maths teaching in connection with Fourier series calculations and so on. Vr4 improves on earlier releases. There are still a few bugs and obscurities. However, in spite of several weeks teaching with Vr4, I still haven't found one not already reported in Heck [3] or on the maple group. I'm sure there will be more on this in the main review, so here is just one example of obscure behaviour (from [3], p260). k:='k': Sum(3^k,k=1 .. infinity); evalf("); Maple Vr4 returns -1.5 . Of course, when asked to do the sum exactly, it gets it right and the reason for the obscure behaviour is the algorithms used "improve convergence". (Compare Sum(x^k, ...) then subs x=3 .) There are educational advantages in some imperfection: it helps us, the lecturers, explain to students the importance of checking their work. MAC-SPECIFIC MATTERS I have written two reviews of Maple for the Mac before, both in Australasian Mac Universities Consortium Newsletter. The first, on Maple 4 for the Mac, was published in 1988. It was a teaching related review, and is the earliest review in CTI-Maths list of reviews of Maple. I thought it was marvellous in spite of a few bugs. My second review concerned the initial release of Maple V for the Mac, Maple V.0.1, probably Australian summer 1991-92. In this case, V.0.1 was a near disaster, the disasters being overwhelmingly in user-interface matters. We lost our "restart" - now back - and essential for machines with modest amounts of RAM. Ira Gessel's review of V.0.1 in Notices AMS 1992 was also critical. However, matters have much improved from the 1991/92 low point. As grades in some imagined "maths software exam", in 1988 I awarded Maple on the Mac 9 out of 10, in 92 about 5 out of 10 (though the criticism was just of Mac implementation) and in 96, 7 or 8 out of 10. The review copy of Vr4 was tested on a PowerBook5300/100 (with 56Meg RAM) and a Quadra 630 (with 20Meg RAM). The single most immediately useful change from Vr3 is that Vr4 Worksheets can be exchanged with absolutely no difficulty between Macs and PCs. Two factors combine to make the transfer easy. (i) From System 7.5, Macs can use DOS-formatted disks. (ii) Prior to Vr4 Worksheets written on Mac had to converted (in the Save As operation) to be read by PC and Unix machines, but Vr4 corrects this. Vr2 and Vr3 improved from the low point reached with V.0.1. Indeed Vr3 "Export to LaTeX" was brilliant. Another hurdle was when PowerMacs came into use. The 2nd line of the following causes Vr3 to bomb PowerMacs. (When reported to Maple Waterloo they said: "know it, will be fixed next release". It is.) _EnvExplicit:= true: n:= 4: p.n:=x^n + n*x - a: solve(p.n=0,x); %2; subs(a=-(n-1),%1); simplify(subs(a=-(n-1),%2)); solve(subs(a=-(n-1),p.n),x); Works fine in Vr4. There is a change of behaviour in deleting a whole output region, i.e. a change between Vr3 and Vr4. I found that in Vr4 it was best to click right next to the terminator of the input line and then shift-click at the end of the output to select, then delete. In the process of discovering this I discovered that confusion can be created - copies of the output region - with various combinations of RETURN and DELETE with the cursor in an output region. (I guess one can imagine factors like the PC's mouse having 2 buttons and cut-and-paste code that might be causing this.) Vr3 to Vr4 Worksheet conversions. I'm not happy with the tools for these. I wanted all my text regions to translate to Normal text regions. The translator however is obsessed with trying to take font and style information over, and so one is faced with things like "R3Font0" and so on to convert to something tidier. Where Vr3 cells have multiple lines of different output, in the translation some can appear ahead of the input lines which produced them. We, the lecturer users, do need better tools. (Look at the Eng. Maths materials at http://maths.uwa.edu.au/~keady , following links to Maths 252, for an example of what just one lecturer has to convert.) Once, when an already converted Worksheet was dropped onto the conversion tool, it complained and suggested a disk error. It seems possible to generate System Bombs by translating .ms files, saving as .mws, and with Maple Vr4 and the Conversion Tool open, putting the old .ms files in the Trash. The facilities for linking to Excel (available via OLE on PC) do not appear to be available in the Mac version. The F1 key for getting Help on selected word is also not available on Mac. As Cmnd-H did this in Vr3 but appears not to do so in Vr4 this route through to Help is missed. The Export to LaTeX in Vr4 worked but some .sty files in the /etc directory had each line of text interleaved with blank lines and, as a consequence, LaTeX-ing the files did not work. We expect that this is a problem specific to this release of the Mac version. STUDENT EDITION MATTERS I am still not totally clear about the extent of differences between the student edition and the full version. My impression is that the Student Edition is powerful enough for staff and postgraduates. By setting out to give the machine a big sum (but one which had been run in a full version of Vr3 on Mac), I obtained: "Error, (in expand/bigprod) object size cannot exceed 8000 words in student edition" (I also came across a limitation to 100 float Digits, and a message that this was a limitation of the Student Edition.) Another conspicuous feature is the STUDENT> prompt. I did not notice a way to make this shorter and tidier. (At least the interface(prompt ...) function did not remove the word STUDENT.) The word STUDENT does seem to be removed when read in on a full version. Some of the Example Worksheets have not been tested properly. E.g. COMBSTR1.MWS with treesize:=100 gave the error "too many levels of recursion" Running with smaller tree-size works. There are a few "broken" hyperlinks. (Probably not just in Student Edition.) Is the package unnecessarily big - in some respects - for students? The previous "one Worksheet active at a time" would be adequate for a Student Edition - and offers less chance of confusing students. In practice the students will be doing sums, and while they need to be able to read the word-processed parts of pre-written Worksheets, they may not need all the features to write, in text fields, polished mathematical discussion. Students need to be given useful URLs. In particular, they need to be told where to get the latest patches. Probably this is a job for the publishers who sell Student Maple: their Web sites can provide a page with relevant pointers. The start-up screen (which did have Springer's e-mail) would be a place where the URL could be put. MINOR CRITICISMS The book "Maple V Learning Guide" (Springer) isn't great. However the Example Worksheets, e.g. NEWUSER.MWS, and the on-line help are enough to get started. Returning to criticisms of the book. (i) There are numerous examples, e.g. pp201-202, of squalid problems - tediously long squalid output. To save trees, please suppress long and hideous output. (ii) The book would be totally inadequate as a sole reference. Fortunately this wouldn't be the case as there are better materials up on the net to support students. For example, the book does nothing on Maple Programming. There is not even a reference to for or to if in the the index. However, for Maple Programming, Monagan's notes in the Share Library are quite good. (When I last looked they were still at Vr3 but the most obvious change that needs to be made is in a few places like proc(n:: integer) i.e. the use of :: for Vr4 type-checking declarations. Undergraduates are much more aware of looking around on the net than are many older academics, so they will find the other items. However, it remains a criticism, as initial impressions of the software are probably more important to them than to older academics who know that the software can do useful things. I would much prefer less on the wordprocessing in Chapter 1 of the book. Yes, there needs to be something on user interface matters, but "writing polished documents in Maple's word processing system" is far less important to our undergraduate students than learning mathematics. A few comments in the code, and, if the work is printed out, extra neatly hand-written annotation and explanation is enough. The fancier points on the word-processing belongs, I think, in a final chapter, not the first. GENERAL WORRIES Perhaps Maple is getting too big? I like the idea of using a small set of orthogonal software tools. Maple has had some spectacular successes for the working mathematician, such as the "Export to LaTeX" facility in Vr3, and continued into Vr4. (This is available in Student Edition and would be useful for postgraduate students.) I think that I would prefer Maple to link with other software in various configurations. I wonder if greater development of products like the following might not be a more appropriate way for Maple-related development to proceed than building one huge package: (i) the MathOffice link with MSWord (for MSWord users), (ii) Scientific Workplace (for TeX-using mathematicians), (iii) Mathcad (principally for engineers), (iv) Matlab Symbolic Toolbox (for numerical scientists and engineers). At the risk of sounding like a boring older academic not wanting to learn lots of new computing tricks, please Maple Waterloo, leave it to (i) to (iv) and to Mathematica to be the big all-in-one packages. Clever combination with other tools - Export to LaTeX, OLE to Excel, C-code generation, etc. - would be a marvellous forte for Maple to have. All this is is relevant to Student Editions too. In practice, now, students often do not have the RAM on home computers to run, say, Mathematica 3. Acknowledgement I thank colleagues at Queen Mary and Westfield College, London, both for the opportunity to teach with Maple Vr4 and for worthwhile conversations on its potential. For some Mac-specific matters, I particularly thank Wilfrid Hodges who noticed some of the faults - e.g. in the LaTeX .sty files - reported above. References [1] N. Fowkes, G. Keady and J. Ward. "Interactive maths texts from Computer Algebra Systems: use in teaching engineering students". Pp137-142 of Vol 2 of Proceedings of the Australian Computers in Education Conference, Perth, July 1995. (Eds R.Oliver and M Wyld). [2] G.Keady and P. Abbott. "Tables of stress and strain- the elastic torsion problem in a CAS". Pp387-392 of Proceedings of the Australian Engineering Mathematics Conference, Sydney, July 1996. (Ed. P. Broadbridge). [3] A.Heck. "Introduction to Maple". (Springer, 2nd ed.: 1996).