by Anthony_Siegman » Wed Jul 14, 2010 1:00 am
Presentations, and especially the use of Mathematica for preparing presentations and publications, are controversial topics, on which I have strong and long-standing views, based on very long experience with teaching; lecturing, preparing and giving all kinds of presentations; preparing and publishing innumerable reports, journal articles, class notes, and books; and even doing a bit of web page creation. So I'll take the opportunity of this forum to express my basic views straightforwardly, and others can and, I expect, will differ!.
Main points:
1) Based on my experience, Mathematica is beyond question a superb (if at this point overly complex and overly expensive) tool for doing symbolic analysis; for doing numerical computations; and for preparing the "graphics" part of sophisticated graphics displays of these analytical and numerical results. Don't know of anything better (and I've used every version of Mathematica to date, and a fair number of other tools before that).
2) The Manipulate aspects of Mathematica in particular are "insanely great" -- what else can one say?
3) But, a real problem with Mathematica is excessive complexity of its interface, and the size of its command and options vocabularies, both of which seriously damage its usability for a very large fraction of its potential users.
Speaking for myself, if I could buy at a much lower cost a "Mathematica Lite" that provided maybe 300 to 500 of the most used of Mathematica's current 3000 commands, and then supplement this with various specialized, individually priced packages for more specialized areas, I'd choose this an instant. And, I think the science, engineering, educational, and academic and professional worlds would all be far better off, and Mathematica could serve a far wider user base, with this alternative -- but I don't expect this to happen.
4) Mathematica is, to say it again, fantastically powerful for preparing the "graphics" parts of graphics output: the points, the lines, the surfaces, the axes, the tics, the pixels. But it is, for multiple reasons, a far from optimum tool when compared to almost any WYSIWG-type and/or mouse-controlled graphics software alternatives such as (just for example) Adobe Illustrator or Photoshop Elements for handling the "annotation and labeling and formatting" aspects of a graphics display: the fonts, the annotations, the labels, the titles, the line weights, the point sizes, the colors, and so on.
And, there is no way around this. These "annotation and labeling and formatting" tasks are just inherently better and more easily done and, very importantly, the skills and techniques for doing them are more easily learned and remembered in conventional, heavily mouse-oriented graphics software programs than in a "command-line" interface like Mathematica.
And, when one is going to re-purpose or re-use a graphics output for many different purposes (in class notes, and in a formal lecture, and in a journal article), with the same points, lines and curves but with modified annotation and formatting, this basic distinction just becomes even stronger.
And, many users will in any case, want to have those "Illustrator and Photoshop and Acrobat type" skills for other purposes, whether they're also present in Mathematica or not.
And finally, attempting to get around this problem by adding WYSIWYG or mouse-driven graphics capabilities within Mathematica itself -- attempting to imbed an ersatz Illustrator within Mathematica -- just makes things worse (rapidly worse!). The already excessive complexity and unlearnability of the Mathematica interface gets worse; and the ersatz Illustrator commands just frustrate anyone who has even modest knowledge of Illustrator.
5) So, this basic situation leads me to two conclusions. The absolutely clear and crucial conclusion is that Mathematica should put top priority on making its graphics-related "Export" commands as solid, powerful, reliable, consistent, simple, and absolutely format-compliant as possible, for all of the major standard document-transmission formats -- PDF first and foremost, of course, and then JPEG, TIFF, and any others that merit similar treatment.
Mathematica should in particular create and structure its Exported graphics output files in ways that help subsequent processing by other imaging-editing and document-preparation apps and make this subsequent processing easier, rather than hindering such subsequent processing. [Minor but telling example: If Mathematica exports a label or some other text string in a graphics output in such a way that the string is formatted as individual characters rather than as a single PDF string -- or if it formats a label like "height -L in meters" broken into three PDF strings just because the "-" is from another font -- this makes subsequent font name or font size changes really painful.]
6) The second conclusion -- for me, anyway -- is that WRI should back off on its oft-state objective trying to make Mathematica a "single tool for all tasks" or at least tone down its focus on this. Concentrate on making it better, faster, more powerful at the tasks at which it's truly great -- symbolic analysis, numerical computation, Manipulation, and very sophisticated implementation of the "graphics" aspects of graphics displays.
But, restrain this urge when it comes to the "annotation, labeling and formatting" aspects of graphics display. Let these be good enough that Mathematica graphics output will be acceptable, if not great, for class notes, lectures, reports (academic and industrial), and for the great majority of routine academic and industrial presentations.
But if you want to go beyond that and prepare "publication-quality" documents and presentations (e.g., glossy advertising brochures or whatever), then do that by exporting from "basic Mathematica" into a separate publications tool, sold separately, and learned separately, primarily by sophisticated programmers. Trying to cram this capability into a "universal" Mathematica that's supposed to be usable by, let's say, high school students and teachers, or even college-level students and teachers, is just a recipe for making Mathematica a worse experience for all levels of its user base.
Those are my views, and I'm pretty confident in them. Others may of course differ . . .