garry@batcomputer.TN.CORNELL.EDU (Garry Wiegand) (02/12/86)
Having spent a week (last month) learning Knuth's "TeX" (I swore
I never would!),
Given the restriction that it is intended to be a slow batch-mode
type-setter (not, in addition, a graphics language, figure editor,
page editor, indexer, spelling-checker, WYSIWYG screen editor, etc etc),
Then I have two basic observations:
1) It will indeed (eventually) generate good-looking output for almost
any situation, and
2) It has the ugliest and sloppiest syntax of any language I've ever
encountered.
And my question is:
Is anyone out there working on a tool which will combine the
functionality of TeX with a decent human interface?
Send replies to me or the net ...
garry wiegand
garry%cadif-oak@cu-arpa.cs.cornell.edu
PS - The "any language encountered" above would include Snobol, Teco, NRoff,
BAL, Turing machines, and a certain professor I once had... :-)rentsch@unc.UUCP (Tim Rentsch) (02/14/86)
In article <242@batcomputer.TN.CORNELL.EDU> garry@batcomputer.TN.CORNELL.EDU (Garry Wiegand) writes: [about TeX] >Then I have two basic observations: > > 1) It will indeed (eventually) generate good-looking output for almost > any situation, and If you think TeX generates good-looking output, you are really living in a fool's paradox. For one thing, the fonts are terrible. For another thing, most TeX output shows how difficult it is to use TeX to produce good looking output -- see Ullman's Database book for a prime example. Saying it is possible to use TeX to produce good output doesn't cut any ice; certainly a system which allowed each individual pixel on a page to be turned on or off could be used to produce good output, but no one would consider using such a tool. > 2) It has the ugliest and sloppiest syntax of any language I've ever > encountered. What do you expect for a language designed using the same principles of macroprocessing languages of the early 1960's? >And my question is: > > Is anyone out there working on a tool which will combine the > functionality of TeX with a decent human interface? Why would anyone want one? If you want nice looking output and a good user interface, I would suggest a Macintosh. It is now in vogue to use TeX to generate incredibly bad looking books, papers, and tech reports. Some things, however, must transcend fashion -- such as producing written words to fulfill their intended purpose: to be read. Disclaimer: Any opinions expressed are my own. The above is intended *not* to reflect on Garry Wiegand, but only to comment on the TeX system and on some of the atrocious output it has produced. cheers, Tim
crm@duke.UUCP (Charlie Martin) (02/14/86)
You're really down on TeX, aren't you? Have a bad experience with it in the past? Seriously, the examples I've seen of bad output from TeX seem to have had two characteristics: first the book design was bad; second, the "camera-ready" seems to have been prepared on a *feh* laser-printer. The book design problem is not one that TeX can be blamed for -- no language can overcome a bad specification. The problem is that the publishers (to reduce costs) let the authors get away with doing their own book design. And not many of the people writing the books have much of an idea of book design. But Knuth's volume II 2nd ed loos pretty good -- perhaps because he was working with a book design by a real book designer. The second problem is also one of cost, but seems to me to be a little less justified. The problem here is that laser printers of the conventional sort have not enough resolution. And while there was some joking about typesetters with their jewelers loops in this group a little while ago, there is a reason: the eye is quite good at noticing that kind of extrememly fine detail in black and white. People *can see* the quality difference between 300 dpi and the sort of awesome crispness that the best phototypesetters can produce. -- Charlie Martin (...mcnc!duke!crm)
garry@batcomputer.TN.CORNELL.EDU (Garry Wiegand) (02/18/86)
In a recent article Tim Rentsch ('>') writes in response to me ('>>') :
>> 1) It will indeed (eventually) generate good-looking output for almost
>> any situation, and
>
>If you think TeX generates good-looking output, you are really
>living in a fool's paradox. For one thing, the fonts are terrible...
We're using the Postscript's fonts, not TeX's. It's a shame good computerized
fonts are so hard to come by in the world, especially for free :-)
>> Is anyone out there working on a tool which will combine the
>> functionality of TeX with a decent human interface?
>
>Why would anyone want one? If you want nice looking output and a
>good user interface, I would suggest a Macintosh.
I deal in manuscripts running up to several hundred pages, with figures,
tables of contents, indices, cross-references, special page layouts, etc
etc etc up the wazoo. I'd love to have a real CAD documentation system
for keeping track of all this information, but I can't afford the $$$$.
I'm not aware of anything on the Mac that comes anywhere near doing a
tolerable job.
And in a private mailing Andrew Reibman responds:
>Why don't you try Lamport's LATEX - I think you'll find it mutch more
>agreeable - plus you have access to all of TEX's funcitonality
Latex, I think, puts one back in a league with Nroff/Runoff etc, where
someone else has made a few limited choices about what the poor lowly
user is to be allowed to do. Mixing Tex into Latex requires considerable
luck and/or fortitude. And the Latex manual (at least the old version)
is a bit scrambled in the brains -- I've had to write notes explaining
Mr. Lamport's explanations.
What I was really asking is: with what can we replace this godawful macro/
macro/macro/what-do-I-do-with-a-space/mouth&stomach language? I admit
I'm puzzled about what a replacement might look like - the problem does
not neatly fall into any traditional computer language. Creative suggestions,
anyone?
garry wiegand
garry%cadif-oak@cu-arpa.cs.cornell.edu (arpa)
[Perhaps a new TeX front end would be a suitable project for a human-
interfaces or programming languages course... hint hint] rentsch@unc.UUCP (Tim Rentsch) (02/24/86)
In a recent article Garry Wiegand writes: > > We're using the Postscript's fonts, not TeX's. It's a shame good > computerized fonts are so hard to come by in the world, especially > for free :-) Great! I'm glad to hear *someone* is dissatisfied with the default TeX fonts and is using better ones. Now if only all the other people out there who use TeX would switch . . . . :-) > I deal in manuscripts running up to several hundred pages, with > figures, tables of contents, indices, cross-references, special > page layouts, etc etc etc up the wazoo. I'd love to have a real > CAD documentation system for keeping track of all this information, > but I can't afford the $$$$. I'm not aware of anything on the Mac > that comes anywhere near doing a tolerable job. The Mac won't do everything, I'll grant you that. And it is slow as well. Both of these defects are changing, though -- Mac Plus is significantly faster (I've heard), and 3rd party applications are more flexible than the standard MacWrite and MacDraw. My suggestion of the Mac was based on its superb user interface (the best I've seen). But if you consider functionality to be more important than user interface, why not consider office publishing systems (perhaps best exemplified by Interleaf)? The Interleaf system is quite flexible, extremely fast, has mixed text and graphics, and most of the goodies you say you're looking for, *and* is a WYSIWYG display editing system. (The user interface on the Interleaf isn't as good as the one on the Macintosh, but then no one else's is, either.) The Interleaf system is expensive, but consider: one, the recently announced educational discounts (quite substantial), and two, what you are buying and how much that is worth to you. If you are in the academic world, your papers and reports are important to your career, and whatever you can do to make those papers easier and more pleasant to read will make more people read them and advance your reputation correspondingly. > Latex, I think, puts one back in a league with Nroff/Runoff etc, > where someone else has made a few limited choices about what the > poor lowly user is to be allowed to do. Mixing Tex into Latex > requires considerable luck and/or fortitude. And the Latex manual > (at least the old version) is a bit scrambled in the brains -- I've > had to write notes explaining Mr. Lamport's explanations. Not having enough flexibility is certainly frustrating. But having too much flexibility is also frustrating, albeit in a different way. The best aphorism I've heard in this regard is generally attributed to Alan Kay: "Simple things should be simple; complex things should be possible." > What I was really asking is: with what can we replace this godawful > macro/macro/macro/what-do-I-do-with-a-space/mouth&stomach language? > I admit I'm puzzled about what a replacement might look like - the > problem does not neatly fall into any traditional computer language. > Creative suggestions, anyone? Hear, hear. > garry wiegand > > [Perhaps a new TeX front end would be a suitable project for a > human-interfaces or programming languages course... hint hint] This I would say differently: How can we take a usable, WYSIWYG system that also makes it easy to produce good looking documents, and add to that system (rethink that system?) so that it has the flexibility and elegance of concept (but not necessarily the same elegant concepts) that TeX has? cheers, Tim Rentsch [personal to Garry: as should be obvious, I enjoyed your note. Glad to see you took my posting in the spirit it was intended.]