LWN.net Logo

On text documents

On text documents

Posted May 10, 2012 23:42 UTC (Thu) by mrons (subscriber, #1751)
In reply to: On text documents by iabervon
Parent article: Who owns your data?

Of course if you had used TeX 20 years ago, the document would look exactly the same today, even down to the line-breaks, and be in an easily editable form.


(Log in to post comments)

On text documents

Posted May 11, 2012 2:34 UTC (Fri) by iabervon (subscriber, #722) [Link]

20 years ago, I only had a dot-matrix printer, and TeX isn't really set up to deal with extremely limited positioning granularity. Certainly everything I've written in the last 15 years has been in TeX unless it's been in HTML or something which renders to HTML. But TeX can actually be kind of problematic: you have to modify the file in order to avoid generating recto and verso pages, which are inappropriate for e-readers (or, really, any presentation form which doesn't involve dead trees). And TeX documents actually often make a lot of assumptions about the form of the result which means that there isn't machine-readable available to produce other presentation reasonably.

(Not to mention that building TeX requires implementations of at least two language dialects (WEB and \ph) which aren't used for anything else on any modern system; it's easier to make an emulator for the computers that Wordperfect ran on than to make a compiler able to build TeX, although people have done both.)

On text documents

Posted May 11, 2012 4:33 UTC (Fri) by eru (subscriber, #2753) [Link]

(Not to mention that building TeX requires implementations of at least two language dialects (WEB and \ph) which aren't used for anything else on any modern system; it's easier to make an emulator for the computers that Wordperfect ran on than to make a compiler able to build TeX, although people have done both.)

Huh? Most Linux distributions provide a TeX package. I believe it is build using a portable C implementation of WEB (web2c), which is a source to source translator. So just C is required for that part. Browsing the READMEs of a recent TeX for Linux implementation (http://www.tug.org/svn/texlive/trunk/Build/) there certainly are also other dependencies for building and auxiliary programs, but that is stuff that typical Linux implementations already provide. Of course bootstrapping TeX for a very different computer and OS from scratch would be a lot of work, but at least it is possible, thanks to the good documentation of teX and its source.

On text documents

Posted May 11, 2012 16:03 UTC (Fri) by iabervon (subscriber, #722) [Link]

web2c (plus a C compiler) is an implementation of WEB and \ph, and is of comparable portability and complexity to dosbox, which will run your old word processors. TeX's source does contain a lot of documentation about the expected Pascal dialect and the preprocessor; but there's even more documentation about the x86 and DOS. Your old DOS programs don't come with an extensive description of the platform they run on, but they're also not the only things that use that platform, so they don't have to.

On text documents

Posted May 11, 2012 17:07 UTC (Fri) by eru (subscriber, #2753) [Link]

Still have to disagree here. I'm pretty sure I could port web2c to a new platform in an evening or two, provide it has a decent ANSI C compiler (which is now a very common piece of infrastructure and can legitimately be assumed). Porting DOSBOX would be a much larger task, unless the new target is very similar to some of the existing ones. Yes, there is more documentation about x86 and DOS, because a lot more is needed to describe the complicated and ugly interface, and it is still incomplete...
I have found bugs in DOSBOX, which I currently use to support some legacy cross-compilation tools at my workplace. Also used DOSEMU+FreeDOS for the same task, and found it has some different bugs... I could work around the problems for the limited set of programs that were needed. But the fact is the only thing that is completely MS-DOS compatible for all programs still is the original MS-DOS.

the computers that Wordperfect ran on

Posted May 22, 2012 21:15 UTC (Tue) by Wol (guest, #4433) [Link]

Although the CURRENT WordPerfect file format is about 18 years old (it was created, as a result of MS breaking the old 5 format) in 1994.

To the best of my knowledge, WordPerfect files are both backwards AND forwards compatible between v6.0 (released in 1994 as I said) and the latest version.

So incompatibility like this is a deliberate or accidental vendor choice, not something that is inevitable ...

Cheers,
Wol

the computers that Wordperfect ran on

Posted May 26, 2012 19:06 UTC (Sat) by mirabilos (subscriber, #84359) [Link]

What influence has MS had on Wordperfect, which, like Wordstar, was something entirely different from MS Word and Winword.

On text documents

Posted May 11, 2012 4:49 UTC (Fri) by eru (subscriber, #2753) [Link]

Of course if you had used TeX 20 years ago, the document would look exactly the same today, even down to the line-breaks, and be in an easily editable form.

I mostly agree from personal experience. I have some large LaTeX documents that were started that long ago, and which I still maintain now and then. Not quite pure LaTeX, because they contain diagrams that were done with xfig (but that also is still available, and quite good for simple diagrams). Some changes in LaTeX (mainly the transition to 3.x) required minor changes to the source, but these were limited just to the macro settings at the beginning of the document. Also I started to use some PostScript-related font packages for much improved PDF output, which slightly changed final layout. But the bulk of the text has not needed any changes attributable only to the formatting tool evolution. Supposing I had not been maintaining the documents for 20 years, suddenly getting them formatted with current versions of the tools might be slightly more work, but not much.

On text documents

Posted May 11, 2012 5:07 UTC (Fri) by Cyberax (✭ supporter ✭, #52523) [Link]

Except that you've probably used some form of TeX macro library (MikTeX, LaTeX, etc), not raw TeX. In which case you have to hunt down all the dependencies and pray that they work.

On text documents

Posted May 12, 2012 15:56 UTC (Sat) by jengelh (subscriber, #33263) [Link]

RTF was an option to save in using Microsoft products back then, and shares the readability of TeX (in principle—WYSIWYG editors like Word had a tendency to not collapse redundant formatting statements, so that font name/color info was repeated for like every paragraph and bullet point).

Copyright © 2013, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds