As an experienced LaTeX user, I was less impressed with how good LaTeX did than the spectacular results he was able to get out of Word and Writer. It isn't that LaTeX failed to impress (it did) -- I was just amazed at how much better the Word and Writer versions looked than the defaults.
Also, his characterization of LaTeX as nothing more than "a set of finely tuned defaults" is absolutely correct. Generally speaking, one wants to think hard before starting to twiddle too much with the default settings in LaTeX, mostly because Knuth already did, and it shows. One also doesn't want to get caught up in the trap of paying more attention to formatting than the text, which is easy to do when using WYSIWYG environments. There's a good reason that the smallest unit of "finely tuned defaults" available in LaTeX is the documentclass -- because producing documents is ultimately what it's all about, not formatting text.
I think the real lesson here is that, for the fixed cost of learning LaTeX once, you get a significant benefit in improved rendering and layout (and that's before even accounting for formulas) on par with what would be rather significant incremental effort in each Word or Writter document you created.
I also love using LaTeX because I can grab a set of template and style files from whatever publication I might be writing for, and I can be sure that the outcome will be in line with what they were expecting. True, I could also get a Word template, but I've always found that with Word templates I hit backspace at some inopportune location and ruin the layout. "A set of finely tuned defaults" indeed!
“Yes, but the smaller caps should have thinner lines.”
That’s not true. Small caps should be drawn so that their line widths about match those of capitals. They should not just be smaller versions of capitals. Both Word and Writer fake small caps by making capitals slightly smaller. That’s just wrong.
“Instead the larger font is thinner.”
As it can be. The thinnest lines of the capitals should in most cases be thinner than the widest lines of the small caps. That’s just normal.
Shrinking would give the right results, with line thicknesses proportionally smaller than regular caps. Keeping the same weight and squashing vertically would also give results that, although skewed, wouldn't have fatter lines that the originals.
Whether you label those techniques as 'fake' or not does not change that LaTeX and/or Metafont's displayed results are poorer than Word and OpenOffice in this aspect.
The article missed the biggest advantage of TeX. The fact that it doesn't change.
I can take a document written any time in the last 25 years in TeX and print it today. It will still print the same way that it did then, modulo available printing technology, down to the visible wavelength of light. I can then tweak it and print the edited version, without any loss of formatting. Using WYSIWYG word processors and random formats that have existed over that time, not so much.
It isn't just word processing formats. The same is also not true of other printing formats such as postscript and pdf. (Witness in this thread how pdf documents that printed fine for one person not printing correctly for another. That with a pdf created this year.)
This trait is very important for anyone who needs to archive documents. Such as happens all of the time in academia...
Maybe in TeX, but my experience with LaTeX had been a little different because the packages and interdependencies between them changes with the time.
I mean, if you take a non-trivial document using several packages from several years ago, and try to compile now, a likely result is the compilation fails with some obscure error or the original result can be different in unexpected ways.
Usually that can be fixed with the use of a new option or calling a different command or little things like that, however the point of compatibility through many revisions of many packages (and its interactions) in LaTeX is something to be warned about.
Because unlike Word, LaTeX can easily be rebuilt from source. And it should work on future versions of the OS. Though, of course, there's no guarantee of that either. So might be best to archive the OS too, just in case.
LaTeX is really, really good. In fact, it's too good for the common people who don't spend a lot of time writing and/or reading books and publications, and instead spend their time not expecting publishing quality. The sad truth is that Word is more than good enough for almost anything.
If you buy a cheap power drill and you use it maybe twice a year, it still probably lasts your whole life and you really don't even need to know why a professional power drill might be much better. Unless, that is, you're a hi-fi junkie and want to do your 100 holes with the best drill you can find, or you're a professional who drills 100 holes a day.
I've been using LaTeX to write my MSc thesis. Anyhow, one of the reasons I used is was that it could be plugged into git quite easily, changes to partials would appear in the source control and could easily be tagged. Instead of changing one huge word or OO document, place a chapter in a partial and only change that one file. All changes could also be tracked by my lecturer. Quite easy tbh.
No, Helvetica's stature as a high point of type face design was not. In fact, Helvetica was the Arial of that time, and The People Who Are Always Right would snicker at it and instead go for Univers, the well-designed alternative that the riff-raff didn't know about.
People often try to make Word do things that require a tool like LaTeX. Most of this is because very few people know about LaTeX, and fewer still know that the quality of LaTeX output makes it worth learning anytime you have a complex document or one that needs to look good.
To recycle an analogy: Word is a hammer, and complex documents are screws. When your only tool is a hammer, everything looks like a nail.
Ultimately, Word and TeX are just tools for producing formatted documents. It's silly to cling to Word because it's WYSIWYG when it can't actually do the job you need done.