TNT Is Not TeX

I like the broader question this inspires: Why are the technologies prevalent in the 60's still prominent today? The correct, standard answer is that they're good enough. Good enough often beats out better solutions. But that's only half of the story. Embrace that theory too much, and we'll find ourselves in a world of good enough solutions.

I don't know about you, but I don't want the world to be "good enough". I want it to be better. We could accept all of what we have today, and take it for granted, and we'd live with it forever. But that would suck.

Even the most entrenched technologies can be displaced, or even replaced, with enough ingenuity, effort, and a superior technology. Most changes are happening slowly, as they must when we have billions of "installs". Most won't replace their predecessor completely, but erode it slowly:

IPv4 to IPv6.

Windows to OS X.

SOAP to REST.

XML to JSON.

HTTP to SPDY.

IE to Chrome.

Flash to Canvas.

Those changes didn't happen automatically; people pushed to get those new ideas out there. Truly better software will make it, but it won't happen if you accept that the old guard is here forever.

Forever is a long time.

One more thing: Many things that are really old — Unix, for instance — aren't the same as they were before. They've evolved, changed, and grown as developers evolve the software. Emacs today is tremendously different from its first release. Software evolves, and often dies when it doesn't.

Unix is still the same (in some ways) as before because Unix had a remarkably simple mental model, and the ways we try to replace them don't. Unix didn't last forever because it had features, good god! No, it's because it was so robust and flexible as to serve even a modern system. TeX, and Unix, and all these things, were hackable. (If you don't agree about TeX, read the TeXbook and some traditional TeX, not this LaTeX crap that's grown around it.) That doesn't mean it these systems are irreplaceable. (Hell, I'm working on a project to replace the Unix shell...) It just means you need to put more effort into devising a very good mental model. Something that actually captures all of the variation you want. So, don't replace TeX because you don't like TeX. Replace TeX because you have something better, something obviously better, something that you could explain to a five-year-old.

This would seem a problem that could be solved by a cloudish sort of solution, though -- have an iPad app which hands actual typesetting of a document off to a server somewhere, sending TeX sources, and receiving a PDF in return.

A simple starter implementation might require your TeX to be a single file, and could just run TeX in a mostly-readonly chroot jail (bearing in mind that TeX can shell out, read/write arbitrary files readable by the executing user, and so on), and serve up a PDF, all wrapped in a web service of some sort.

This wouldn't be terribly fast, but given TeX's performance, the actual typesetting is likely to be the bulk of the time spent, even considering the round trip.

Okay, just for

dextorious

and

jakobe

and starting with my

http://news.ycombinator.com/item?id=3334504

I'll try again:

(1) Issue 1.

In that issue, the author had:

"It’s comforting to think that all the TeX documents I’ve written over the years will still be readable a century hence. But Mind Two reminds me that in practice I have trouble maintaining TeX documents even for a few months, much less decades or centuries."

but here his concern is NOT about "TeX documents" but about LaTeX documents.

They are not the same.

(2) From dextorious.

About Issue 1, dextorious wrote:

"And that is just a pedantic excuse, that no one cares about. People don't use TeX itself (and, to prevent another pedantic detail, I mean the vast majority of people), people use LaTeX+TeX."

No. The claim:

"People don't use TeX itself".

is false: I use TeX itself, a lot; I'm a Ph.D. applied mathematician and, thus, use TeX for math. In fact, for math, I know of no reasonable substitute except for LaTeX which I don't like as well as just TeX.

Then, since I have TeX and know how to use it, it is also my main high quality word whacking software; I use it for letters, technical papers, foils, and even business cards. I essentially never use LaTeX. And I have a version of Word and nearly never use it either. E.g., even for just simple letters, I much prefer TeX to Word.

That I actually use TeX itself is not "another pedantic detail" because one of the main reasons I won't use LaTeX is JUST the objection of the author that LaTeX is not stable: I want the stability of TeX. E.g., I have over 100 macros in TeX. Nearly all of these were written in the 1990s. When I changed to Windows XP, I copied over all my TeX macros, and all of them have continued to work just as before just as Knuth guaranteed. GOOD.

(3) TeX Doesn't Slice Bread.

There is no one tool that can do everything, and it would be foolish to hope for such a tool.

Knuth was extraordinarily careful about saying just what TeX was, providing some exemplary documentation, and then essentially freezing everything for all time.

The author criticizes TeX for not being what the author wants. This is not fair or even reasonable.

It would be fair to criticize TeX for failing to do what Knuth promised. But it's not fair to criticize TeX for not being something it was not intended or promised to be.

It's not fair or even reasonable to criticize a Stradivarius violin because it is not a Steinway piano, French food because it is not Chinese food, a car because it is not a truck, etc.

As software goes, LaTeX is relatively stable, but the author criticizes LaTeX on stability. So, (A), the author wants stability, more than LaTeX has. But, (B) the author wants more in user interface and likely languages, alphabets, fonts, color, cross references, maybe rotations and other transformations, likely spell checking and hyphenation in many languages, graphics, maybe with sound and video, etc. And we have to assume the author also wants the ability to do well setting math. Then in practice (A) and (B) conflict: That is, trying to provide the much greater functionality of (B) will mean years and years of revisions that conflict with the stability of (A).

Well, I have good news for the author: It's been a secret and ABSOLUTELY not to be distributed beyond this thread. So far known by no more than a very few people, the UN has funded, starting some years ago, a unique software development effort in a monastery in Tibet to develop UniTeX, word whacking software that will handle all of the Unicode alphabets and their languages with spell checking and hyphenation along with 48 bit color, 3D, resolution from 600 pixels per line to tens of thousands, execution on super computers down to cloud servers, desktops, laptops, palmtops, tablets, mobile phones, wrist watches, and magic decoder rings, many thousands of fonts, Haskell, for at least 1000 cores, as the macro language, all of the functionality of Distiller, PhotoShop, Final Cut Pro, and AutoCad, output on paper, PDF, AVI, Flash, SilverLight, and MP3, and much more. Version 1.0 will be released when all the functionality has been implemented and is at least as stable as TeX. Then version 1.0 will be the last version. The project will be open source, but the name UniTeX will be reserved for the one, only, genuine, guaranteed authentic, dyed in the wool, UN UniTeX. The release date for version 1.0 is currently set at year 2100.

In the meanwhile, TeX is what it was intended to be. I believe that TeX, just as it is, is terrific for all my higher quality word whacking.

The main target audience for TeX is writers of the most serious work in subjects with a lot of mathematical notation in math, science, engineering, and technology. For such writing, TeX remains quite good. E.g., the emphasis on English, black and white for the fonts, the prominence of the TeX Computer Modern fonts and the AMS fonts, the stability, etc. are all reasonably appropriate. In particular, it will not be easy to get the relevant academic journals to give up TeX in order to get a lot more in color, fonts, alphabets, etc.

Information technology entrepreneurship is a very active field, especially in Silicon Valley. So, someone who wants more than TeX could do a startup. My view is that getting a good combination of product stability, product functionality, product development cost and time, user adoption, and revenue with the good features of TeX and much more, that is, getting what the author wants, would not be promising. Or, Knuth worked hard on TeX for what, over 10 years; doing a lot more would take much more, would take too darned much.

Finally, there is another point about TeX: Before TeX, getting math typed was just GRIM. Typically the typing was much more difficult than the math. TeX made it quite reasonable to get math typed. GREAT.

TeX is so good that now the challenge is the math and not the typing.

For the main intended users and uses of TeX, e.g., Knuth and his writing, TeX is fine, and changing to something else just is not worth the effort.

In particular, Unicode, color, a graphical user interface (GUI), etc. are from not worth the bother down to very much not welcome.

Net, there was a big problem. TeX provided an excellent solution to the problem. The problem has not much changed and remains solved. Done.

The article has several issues of 'drama' regarding TeX. All of these issues are just contrived to have drama for the sake of drama; none of these issues is a serious comment about TeX.

I discuss six such issues:

Issue 1.

The article has:

"It’s comforting to think that all the TeX documents I’ve written over the years will still be readable a century hence. But Mind Two reminds me that in practice I have trouble maintaining TeX documents even for a few months, much less decades or centuries. What about those presentations done with the foils class that stopped working after an upgrade and that I’ve never bothered to fix? Or the articles using the pstricks package that won’t compile under pdflatex? TeX itself may be a fixed point in the software universe, but everything else spins dizzily around it."

So, here the author seems to have a 'dramatic issue' to grab the reader by the heart and/or the gut. Yes, 'writers' tend to do that.

Of course, strictly this "issue" is 100% total nonsense: The problems he listed are all about LaTeX which strictly just is NOT the same as TeX, that is, as in:

Donald E. Knuth, 'The TeX book', ISBN 0-201-13448-9, Addison-Wesley, Reading, Massachusetts, 1990.

with the macros there called Plain.

Or LaTeX is as in, say,

Helmut Kopka and Patrick W. Daly, 'Guide to LaTeX, Fourth Edition', ISBN 0-321-17385-6, Addison-Wesley, Boston, 2004.

and

Frank Mittlebach and Michel Goosens with Johannes Braams, David Carlisle, and Chris Rowley, with contributions by Christine Detig and Joachim Schrod, 'The LaTeX Companion, Second Edition', ISBN 0-201-36299-6, Pearson Education, Boston, 2004.

and in these two books we do not see Knuth's name.

So, can't blane the problems of LaTeX on Knuth or TeX.

So, his:

"It’s comforting to think that all the TeX documents I’ve written over the years will still be readable a century hence."

Good. So, with TeX, just TeX, that is, just Knuth's TeX itself and not LaTeX, FakeTeX, JunkTeX, or NotTeX, the author of the article of this thread can remain comforted.

Issue 2.

Yes, the 'writer' seems to like drama more than rationality since he also has:

"The skeptical Mind Two has another argument as well: Under Knuth’s edict it’s not just the TeX markup language that can’t change; it’s also the architecture of the system. Knuth created his flawless soufflés and dæmon diarrhœa at an ASCII terminal wired to a PDP-10, and the only way he could see the product of his labors was to walk down the hall and retrieve hard copy from the AlphaType machine."

But the "architecture of the system" has nothing to do with "walk down the hall and retrieve hard copy from the AlphaType machine". Instead, part of the "architecture" is the guarantee:

"His intent in freezing TeX is to ensure that the same input should always yield the same output."

and that guarantee continues to hold without a "hall" or "AlphaType" machine. How? Because the guarantee was about the output of TeX which was just a file of type DVI (which abbreviates 'device-independent file'). The TeX I continue to use as my most important word whacking software continues to put out a DVI file and appears still to meet the guarantee.

Issue 3.

The author goes on with

"But the core programs still run in batch mode, as they did in the Dark Ages. To make even the smallest change in a document, you still need to throw away all the existing output and run a whole file (or set of files) through the compiler tool chain."

Of course. He should expect something else? I have two responses:

First, what's wrong with "batch mode"? What else is to be done? As any Web site designer knows, due heavily to the 'sessionless' model of the Web, the server code for a Web page in HTML and CSS (but not necessarily Ajax) runs essentially in 'batch mode'.

Second, even a little reading of Knuth's book above shows that it would not be easy to be given a change in the input and then to get significant computational savings by determining what fraction of the output can be saved and not calculated again. So, we calculate all the output again.

Issue 4.

Next, the author has:

"Sometimes you have to do it twice."

Not due just to using TeX! However, TeX does permit writing to a file and reading from a file. So if want to use such file writing/reading to generate, say, cross references, then will need to run TeX more than once. But the effects of such writing/reading are due to the user and not to TeX. In particular, we can notice that in Knuth's book, there are no cross references that would require running TeX more than once.

Or, as we can see from reading Knuth's book, it is not easy to know exactly where will be the break betwoen one page and the next. Then it will not be easy to do cross references accurate to the page number. So, in Knuth's book, he didn't.

Sure, in my usage of TeX, I have some macros for cross references to the page number, but my macros can occasionally make a mistake and, thus, are not up to the quality of TeX.

I am sure Knuth saw this issue but wanted the high quality page breaking more than some automated cross references.

Issue 5.

The article has:

"By then we’ll just throw Moore’s Law at it: Automatically rerun TeX n times for every keystroke in the editor."

Here the author is implicitly assuming that, of course, a word whacking system SHOULD be based on an 'interactive interface' with 'what you see is what you get', one keystroke at a time. That's a very specific assumption that would be quite questionable at any time and otherwise would tend to 'date' TeX.

For me, I would reject that assumption. That is, absolutely, positively I do NOT want TeX to have an 'interactive interface', to be 'what you see is what you get', or to do word whacking for each of my keystrokes. I do NOT want such things.

In particular, I want to be able to run TeX in a 'batch script' with, "Look, Ma, no hands.". That is, I want TeX to remain something that can do its work without my being there, that is, part of the future where computers do work without a human giving full attention at each keystroke.

In particular, a good response to 'what you see is what you get' (WYSIWYG) is that with such an 'interactive interface' 'what you see is all you've got' and that's not very much.

What can there be that is not seen? Sure, can have a 'conception' for the document and parts of it and have some macros that contribute to this conception. Or, the macros automate parts of the conception. E.g., want a macro for the page headers. Then all the page headers look as in the conception for the page headers. Strictly with WYSIWYG, would have to type in the header for each page individually and, then, struggle to get them all as in the conception. Similarly for how to number sections, subsections, figures, tables, and equations and much, much more.

Issue 6.

I would agree with the author that TeX does not bake bread, drive a car, vacuum the carpet, recognize human speech, do the calculations for computer generated scenes in movies, etc.

Instead, TeX was mostly just for what Knuth said it was for, "beautiful books", especially ones with quite a lot of mathematics. Moreover, implicit was that tbe books would be essentially in English. E.g., the hyphenation was for English, and no fonts were provided for text in Cryllic, Fraktur, Arabic, etc. In particular, the characters were to be from 7 bit ASCII and not 8 bit whatever or Unicode.

Consider his "books": TeX was looking backwards in time, that is, was intended to be a computer version of traditional, beautiful typesetting for BOOKS, just such BOOKS, and not something different, better, for the future, etc.

Knuth didn't say 'beautiful PDF files', 'beautiful Web pages', or 'striking foils with voice over, background music, and multi-touch gesture recognition', etc.

So, there is a lot that TeX is not. Right.

There's no 'drama' in saying that TeX won't sweep floors, and the article has nothing that says that TeX fails in any of its promises or guarantees.

In particular, the article's 'failures' of TeX are for, say, LaTeX or something contrived about some printer down the hall.

There is more on what TeX is and is not in

http://news.ycombinator.com/item?id=2327490