1

I find it very annoying that one needs to explicitly inform TeX about matched parenthesis with \left and \right. Is there a clean and elegant way to make it the default? I guess I should be forced to inform TeX whenever the parenthesis are not matched, and not the other way around!

I guess the if one writes f(x), the correct (in a sense) should be f\left(x\right). In my humble and inexperienced opinion, we should write f(\int x^2) without \left and \right. And if someone intentionally does not want this behaviour s/he should say f\unmatched(\int x^2\unmatched).

Edit: I am not talking about typography. I am talking about semantics. I guess TeX mixes up those two concepts. For example David's answer to the thread just says that \left and \right implementation are, in a sense, broken.

Whenever it is important for typographical decisions to know the intended semantics, I think the user should be required to explicitly state it. Otherwise, the only reason I see for not letting the typography expert (not me) decide is the fact that implementation is too hard to be done or fundamentally broken to be adapted/fixed.

For the typographic enthusiast, I would recommend TeX to make good decisions and letting the typography professional fine tune at the end. It seems important to notice that the more "fine tunes" you have, the more rigid is your project, in the sense that changing the contents a little might imply all the "fine tune" work to be undone and redone.

I don't think most of people care about how the implementation of \left and \right is done internally. I guess they want to tell the typography professional that those two parenthesis are to be matched.

André Caldas
  • 111
  • 3
    If I understand your write-up, you appear to hold the view that matching pairs of "fences" -- e.g., pairs of round parentheses, square brackets, curly braces, etc -- should be auto-sized (say,via implicit \left and \right prefixes) by default. This proposition is of dubious validity, typographically speaking: Automatic sizing of fences frequently produces sizes that are not optimal from a typographic point of view. See Is it ever bad to use \left and \right? for a more in-depth discussion of the related issues. – Mico Nov 23 '19 at 12:35
  • Dear @Mico: I made an edit. – André Caldas Nov 23 '19 at 14:27
  • 1
    @Mico: I know it is very annoying having to explain over and over to people that they should not do what they are doing... thank you very much for your patience! I just want to make it clear that I am not complaining about anyone. – André Caldas Nov 23 '19 at 14:34
  • Left and right do not add any semantic value to the input, because opening and closing parentheses/brackets/braces are distinct characters. \left…\right is basically about sizing. So, leave the semantics of the brackets and let the typography specialist do the sizing in the end (e.g. with left/right). – TeXnician Nov 23 '19 at 15:17
  • @TeXnician: Nope. Left and right is about one being the left (opening) and the other one being the right (closing). The typographical consequence is about sizing. The implementation is about typographical decisions. TeX people tend mix up implementation and semantics. For example, warning you about mismatched parenthesis is semantics (as well as implementation). – André Caldas Nov 23 '19 at 15:32
  • 2
    no you get better spacing with f(x) than with f\left(x\right) only use \left \right if you need delimiters of arbitrary size, eg around matrices. – David Carlisle Nov 23 '19 at 15:32
  • 3
    @AndréCaldas \left and \right is just about size, if you want to mark open and close that would be \mathopen{(} and \mathclose{)} but they are the default anyway, but if you mark open intervals as )1,3( then \mathopen{)}1,3\mathclose{(} is what you should use (not \left\right) – David Carlisle Nov 23 '19 at 15:34
  • Isn't the default semantics of ( to be opening for you? It is for TeX as David said. What more do you want the character to hold? – TeXnician Nov 23 '19 at 15:59

0 Answers0