Hacker News Reader

Comment by Ralith | Hacker News Reader

So if I use LaTeX and it gives me the results I want and I don't find myself spending inordinate amounts of time tweaking things, does that make me an intermediate user? I guess I had better stop learning.

View on HN

I think you hit a local extremum, so yes, I'd advise you to stay there :-)

The day will come when you will need those two figures side-by-side, together with a caption positioned just right. Or a table that spans multiple pages and has multi-column cells. Or, heaven forbid, you will need to submit your bibliography in a byzantine format invented by someone with nothing better to do ("we take pride in the fact that we place a period after author's names").

LaTeX also has another problem: even if you know it, getting from a document.tex file with 0 bytes to a document that looks good takes much longer than it should. By "looks good" I mean doesn't use Computer Modern, and overall doesn't look like a thesis from the 70s.

If you don't like Computer Modern it's only a \fontstyle{times} change away. It's humerus that you claim Latex doesn't look good as it's beautiful output is the reason many people put up with it's warts.

> The day will come when you will need those two figures side-by-side, together with a caption positioned just right. Or a table that spans multiple pages and has multi-column cells.

Not sure if you mean to say it is difficult to do or not in LaTeX.. but I had to do just that last week (10+ pages long tables, side by side figures with fancy captions, etc.).. and it was just a matter of searching for "LaTeX $problem" and I was in business..

This just goes to show you the shabby general state of word processing software. I've been looking for something better than OO/Libre/MSWord/etc for years and avoided using LaTeX because the workflow is _so_ 80's. However, having recently decided to give up, and switched to using LaTeX, my worksheets, course notes, exams, and presentations have never looked better, and take about the same time, or less to create. I'll also be able to do things like randomize the order of exam questions and create individual grading keys in a way that makes sense to me.

To be fair, I would advise almost no one to do the same. Something better is needed[1], but I think that most potential competitors know better than to attempt to challenge MS Word.

[1] And, indeed Google Docs is usually passable and offers some advantages.

> avoided using LaTeX because the workflow is _so_ 80's.

I've never understood this way of thinking. Either it works or it doesn't work -- the age should have nothing to do with it.

I don't like drawing circuits in LaTeX. Here, a simple circuit example:

  \item   
  \begin{circuitikz}[american voltages]
  \draw 
    (0,0) to[battery, l=12V](0,3)
    to[R, l=$150\Omega$ ](3,3) 
    to[D,l=$D_1$](3,0)
    to[short](0,0)
    ;
  \end{circuitikz}

It makes perfectly nice looking circuits, so it passes a definition of "works" but it is an archaic way of describing circuits. if I had to do serious electronics work in that manner, I might not bother. Yes, I can include drawings created by other software, and do so for more complex circuits.

I can understand that as a reason. In this case it sounds like there is a newer, more efficient format that has superseded the old one. (I'm no expert in circuits, so I'll take your word for it.)

If some new technique/workflow/whatever is newer and more efficient, then I'm all for it. I just don't like the attitude that "newer == better" that has seemed to become more and more prevalent in recent years. (I'm not saying you were doing that, but the phrasing you used is something I typically see in that situation.)

Didn't you know that everything is style and nothing is substance?

Has no one here tried Lyx? http://www.lyx.org/

I love LaTeX, I wrote a physics PhD thesis in it and many papers as well. It works phenomenally well up until the point it doesn't work at all. And when you need to go fix something it is a deep and frustrating rabbit hole in my experience.

The point where something is unfixable in LaTeX is much more complex than the point something is unfixable in Word, where I seem to jump down the rabbit hole on relatively simple documents.

That is a matter of opinion. I think the lack of source code is a problem of incredible complexity, in that overcoming that hurdle would require one to somehow force MS to fix something.

I think LaTeX starts marginally easy, gets complex and fiddly but then pretty much just stays at the same level of complex and fiddly. Word starts very easy, gets complicated quickly and eventually presents un-fixable problems, no matter what your skill level.

That's where version control helps. Save and commit frequently. It is much easier to go back to just before the problem and insert smaller chunks. Found many an error this way. Selective cutting of sections and stashing them in a text editor works as well.

Me too. I'm currently using it for the second time for a large multilingual dictionary. Both dictionaries are the first produced for these languages. People are thrilled at the results. It is so easy to create something beautiful with LaTeX.

I think LaTeX simply doesn't live up to its promise. Yes, its a big improvement over Word, but I still find myself manually adding \newpages, rewriting paragraphs and shuffling text around because figures are placed in completely awkward places or you have pages with just one line of text and a sea of white.

That is in addition to the horrors you unleash once something stops working and you need to dig deeper. dvi2pdf..

There are about 5 parameters that adjust latex's pickiness when doing page layout with figures. Things like proportion of the page taken up by a figure. Tweak these and these problems should mostly go away. You may still need to move a figure ahead of a block of ordinary text (only do it at the final edit!) but you won't be tearing your hair out.

Over the last couple of years I have migrated my workflow away from MS Office products (Word/Excel/VBA/PowerPoint) to LaTeX, R, Sweave/knitr.

These tools play well with git and make it easy to automate an analysis. I am a strong believer in literate programming - embedding the code in the documentation where feasible. There are case where it is not (long Monte Carlo simulations) but I can at least pull in the results and have the source code that generated them in version control and the report automatically uses the latest version.

Microsoft kept changing VBA and each release would break things. We had third party tools that had locked VBA modules and ours just wouldn't permit processing data in these workbooks.

I can extract all the data with R and am much happier. The Open Source community seems to better support their products than Microsoft. Go figure...

Reply on news.ycombinator.com