Item — HN Reader

> I've been using LaTeX for over ten years and I still wouldn't say that I "know" TeX in any meaningful way. I was not only productive but proficient in Typst in a day or two.

That's an interesting insight. Do you have an idea why this is? Do you have a CS background?

7 comments

That's also been my own personal experience, and I have a CS background and have written several papers and a PhD thesis in LaTeX. I have done many detailed figures in TikZ and done countless presentations using Beamer.

I have no idea how to implement anything in TeX on my own besides rudimentary macros. I did once try to implement a simple typesetting package for typesetting some particularly gnarly natural deduction trees, and I thought it worked fine for a few examples, but then it broke down in ways that was completely non-obvious to me.

The computational model of TeX is really, really opaque to me.

If you had opened and read the TeX book, you'd understand it.

I am the proud owner of all 5 five volumes and Digital Typography. I have been in front of a computer for almost every single day for around 45 years. I speak and write several languages, including Chinese and Japanese. I am regularly coding in several programming languages and consider myself fairly fluent in things like HTML and CSS. I have worked as a typesetter and assistant editor of two medical periodicals.

I could go on but what else in the world do I have to add to that so I can say "as for me, I opened the TeX book, read it, and understood it" or "I know how to use LaTeX, confidently"?

I don't know the answer to that question, but I do know the answer to the question "is TeX / LaTeX a well-designed, user-friendly, sane software?". The answer is no, no, and no.

Millions of people love to use LaTeX and use it daily. Your experience is just a personal opinion.

Millions of people love traffic jams which is why they're happy they can have them twice a day.

Your claim is just a claim.

Oh definitely, let’s do a PhD in Latex and its worthless toolchain before start writing the actual thesis.

Software should be invisible and not a hindrance to our ability to express ourselves.

Reading a single book is not equivalent to a PhD in LaTeX.

A single book on latex only covers you until the introduction of your thesis. Then you will need to add a table. That’s a different book. Then you will need to align your equations. That is another book. Then you will need a graph. This is a freaking library of books.

I mourn for the time I wasted on Latex. For nothing.

That's not even close to be true.

Oh, you mean those 500 pages of dense Knuth writing?

Dense? The TeX book is a total delight to read.

Not for people who are fun at parties

Isn't that everyone's experience? TeX is a nightmare. It predates all standards for sensible software.

Pure TeX makes MUMPS look well-designed.

Yep, I'm a CS researcher in a top-3 department. (Hate to be all credentialist like that, but I'm guessing it does mean something to this audience.)

I think a lot of the reason is that, ok, LaTeX is extremely complicated. We all know this. Its partisans tend to believe that this is because typesetting is a hard problem. Typesetting is a hard problem! We know this, too. But I think that a substantial fraction of the complexity of LaTeX is accidental complexity stemming, ultimately, from the inherently loosey-goosey nature of the system.

See, there is no real abstraction in TeX. There are no real mechanisms for encapsulation/information hiding/whatever. It's all just characters that eat characters and turn into other characters. Anything can do anything. Anything can be anything. As a result, the whole "theory" of what a TeX program "is" (like in the sense that Peter Naur used that word?) is conventional.

This means that even to reuse other people's code, you have to imbibe decades of convention that's maybe semi-documented, maybe folkloric, or maybe perfectly sound but requires you to read a 230-page manual. I can only speak for myself, but for me, it's no way to live.

Now, again, TeX-lovers tend to claim that its high degree of loosey-gooseyness is necessary, that it's what makes TeX "powerful." I don't know, that sounds to me like the same old "you can't handle writing assembly" story. Sure I can. I'd just rather not if I can help it.

Maybe it's the PL person in me.

Or could it be that the fact that LaTeX is a layer on top of TeX is a cause of additional complexity and that you still have to deal with TeX, so the abstraction is neither complete nor opaque? I assume in Typst they don't (yet) have this issue. Another reason could be missing static typing and thus less formal means to find bugs? You "missing abstractions" finding would point to that.

Yeah, absolutely, the sheer number of moving parts is related to this. There's TeX-the-language, there's LaTeX-the-format (a big layer of macros and conventions), the different LaTeX standards (LaTeX2ε, LaTeX3, whatever, I don't know the ins and outs). Then there's the implementation: all the different engines (pdftex, XeTeX, LuaTex), then there are the distributions, then there are build systems and wrapper scripts to handle all the mysterious compilation-cycle incantations that depend on all the other stuff (latexmk, etc.).

All these layers and alternatives have proliferated partly because of age and the necessity of new solutions to new problems. I also think they're partly coping with fundamental limitations, design decisions that seemed like (or very much were!) a good idea in 1978, that sort of thing.

And the TeX that we know is the "new" TeX, which was created after Knuth completely rewrote the "old" TeX in 1983.

I totally agree. I have written in LaTeX extensively and Typst is a game changer and a life saver. The community is also a plus. I love it and I will never go back to TeX.

TeX had many design goals. "Being a good programming language" was not one of them.

Indeed! It undeniably succeeded at setting the standard for typesetting quality. Big ups to Don K.

It also brought typesetting to academic research, which means it's doing something that wasn't necessary before it arrived.

I was one of the last holdouts from a bygone era. I finished my dissertation in physics, in 1993. It's neither typeset, nor even in a computer readable form. Some fellow students were already using LaTeX by that point (mostly high energy physics, the slowest to graduate of the physics specialties) but I wasn't going to change my already obsolete tech stack within mere months of finishing.

I also have my parents' chemistry theses. They took handwritten manuscripts to a typist who banged out 4 copies at once using carbon paper. And then they entered their equations and figures by hand. (My thesis is hand corrected too). And their theses were short.

LaTeX did a lot of things for my fellow students, but it didn't make them finish quicker.

You wrote an analog thesis about 31 years ago, analog31? You've been playing the long game with this account!

Appreciate the story, thanks for sharing. Can I ask what your thesis was on (in whatever way isn't too personally-identifying if that's a concern)?

I anticipated this moment. ;-)

It was at a time of extremely rapid development of laser technology, where you built your own lasers to be at the cutting edge, which meant optics, electronics, and in my case, computer control plus data collection. I developed a method of increasing the signal-to-noise for a class of spectroscopy measurements by a factor of roughly 1000. This opened up a number of possible experiments involving otherwise weak effects.

My experiment produced some token physics results, but the method that I developed was really the point of it. A couple other labs used my setup, substituting more modern lasers as they became available. Meanwhile, I went into industry, and still work on measurement instrumentation today.

Good stuff, nothing but respect for that kind of work. AMO people are a different breed. Spending all day interpreting power spectral densities and tracking down ground loops takes a certain je-ne-sais-quoi. Literally. Last time I used a microwave amplifier I plugged it in with the wrong polarity and blew it up!

Fellow PL grad student—not a top-3 tho ;)

I’ve looked at Typst and it looks attractive. I’m not quite willing to invest in it until, say, PACMPL lets me use it there. Any idea if that will happen any time soon?

I wouldn't count on it!

For me, it's been invaluable for notes and memos (and my thesis). When it comes time to submit a paper, I transcribe everything to LaTeX. It's a pain, but it's a lesser pain than transcribing from handwritten notes scattered across three notebooks and several crumpled-up sheets of printer paper in the bottom of my bag. That's my own peculiarity; YMMV.

(Edit: also, ugh, I feel like such a goof for having written that. Thanks for being nice about it.)

> top-3 departement

I’m curious- what does this mean?

Ugh, unfortunately it means "please note that I'm not an unserious person." Very often when you have opinions on software on the internet, trolls will jump out of the hedgerow and shout, "who are you to have an opinion?" And, sadly, that same sort of person will often be mollified somewhat by prestige.

Feels kind of icky to have written that, though, and maybe I regret it slightly.

I think it's good, because it will enable other people to be less scared to ditch Typst

Honestly, I consider it embarrassing that it took so long for LaTeX to be dethroned. Academia is supposed to be about thinking for the long term. Instead everyone is too busy publishing and not perishing to rid themselves of LaTeX (and predatory journals outside CS, etc. etc.).

Given that, it takes credentialed "winners" to change the system. Everyone else is too often too afraid to play the game in the lowest risk manner. Sad.

Thanks for your posts on on Typst.

It also belies an ignorance that some schools have specializations. Utah wouldn't come to mind as a "top-20" school by the snobs, but they basically created computer graphics.

It narrows things down to about 20 departments. ;-)

I have a CS background, I have used LaTeX for over ten years, and I also don't know TeX in any meaningful way. I feel like I'm copy-pasting snippets of code and trying random stuff until I get the outcome I want

From experience, you don’t need to not have a CS background to find TeX/LaTeX difficult or obscure to customize or extend.

Here’s a challenge to try (I hacked together something similar that only works in specific cases): write an environment where you get N columns in portrait page mode, with a vertical ruler, say, 12pt to the left of each column; respecting different paper sizes.

The amount of \makeatother, and what does that even mean??, will blow your mind

> The amount of \makeatother, and what does that even mean??

I’m not a TeX lover, but this part is just a snipe. I’m guessing you know and this was a joke. Apologies that it annoyed me enough to write this comment.

If you care enough to read this, you could learn the purpose of \makeatother [1]. It’s not tricky.

[1]: https://tex.stackexchange.com/questions/8351/what-do-makeatl...

Apologies, my comment was poorly written.

I meant “\makeatother, and a lot of other things whose meaning is not clear”

I used to use LaTeX pretty frequently and although I was reasonably productive I also never felt that I "knew" it. It was a whole lot of googling "how to do x in latex" and pasting in something that was too arcane for me to remember for next time (and/or copy pasting from my own prior documents)

In all honesty, I only ever did it to achieve the "LaTeX look". These days I write in markdown or (gasp) wysiwyg editors like google docs.

Thanks to pervasive use of LLMs I predict the future is mostly markdown. It’s the defacto output format for every LLM. It’s simple, plaintext and easy to version control.

Hell I wish I could submit my resume as markdown.

What's stopping you?

Ease of use can't have been top of mind when creating Tex, that's all.