2

A most respected moderator repeatedly advises against the automation of \left and \right in math mode. (Among others, here is one.) There must be some good argument behind that. I am just curious, why?

blackened
  • 4,181
  • 1
    For one thing, \left...\right cannot break across lines, whereas \big etc. have no such limitations. I would guess the other argument is that one should deliberately choose their delimiter sizes, rather than trusting them to an algorithm. – Steven B. Segletes Aug 14 '17 at 13:59
  • 1
    Try $a(b+c)$ and $a\left(b+c\right)$ – egreg Aug 14 '17 at 14:00
  • 2
    Often scaled to be too high. Excessive use (where there are no scaling being done) makes the math code a lot harder to read. – daleif Aug 14 '17 at 14:01
  • @egreg's no moderator:-) – David Carlisle Aug 14 '17 at 14:04
  • 2
    see for example https://tex.stackexchange.com/a/173740/1090 (I suspect this question is a duplicate of that) – David Carlisle Aug 14 '17 at 14:05
  • 1
    adding new questions on to existing ones doesn't really work (especially after the question has been closed, but even if not) the site works best with one question per post. (but yes stay away from commath:-) – David Carlisle Aug 14 '17 at 15:37
  • @David After these comments what I really want is to delete the question, but then I get the warning that deleting is not favored (or something like that). – blackened Aug 14 '17 at 15:39
  • 1
    you don't have to follow warnings:-) but without the extra "new" question it was fine, and just a normal duplicate of an older question, so no need to delete duplicates are just a natural part of the site function popular subjects get a lot of duplicate questions which help people find them as increases the search terms. – David Carlisle Aug 14 '17 at 15:42
  • @Mico I ask for this question to be reopened because here we are asking why when the other question asks takes it for granted that it is wrong. There is a real difference in the requests. I have read the answers to the other question and I still don't know why. I didn't quite understand all the answers. – AndréC Jan 27 '19 at 05:07
  • @Mico He knows that it is considered bad and wonders if it is still the case. As a result, the answers are technical and do not explain in depth why it is wrong. – AndréC Jan 27 '19 at 05:30
  • @AndréC - I wouldn't say that the answers to the linked query "take it for granted" that using \left and \right can be wrong. Instead, they provide specific examples to show that \left and \right can produce suboptimal typographic outcomes. E.g., in the answer I provided, I showed that using \left and \right can lead TeX to create "fences" that are either too small or too large. If this reply isn't satisfactory, may I suggest that you post a new query, in which you state what types of information you would consider satisfactory? – Mico Jan 27 '19 at 05:34
  • @AndréC - Just that I understand better what type of answer would satisfy your search for "why". I take it that you agree with the proposition that a primary objective of good typography is to convey the typeset material clearly and legibly. Take, then, the case $\left(\left(a+b\right) \left(c+d\right)\right)$. The outer parentheses are, mathematically speaking, not required; their only justfication them is typographic: to provide some visual grouping. Since \left( and \right) do not create larger parentheses, they fail the typographic objective. Is this enough of a "why"-type answer? – Mico Jan 27 '19 at 05:47
  • @Mico You give a typographical reason by illustrating the random side of left and right. Egreg tells him that the big ones are not good, he prefers the smaller ones. How do we know that typographically it is better to use this or that type of parentheses? – AndréC Jan 27 '19 at 06:08
  • @AndréC - The previous postings and associated answers on why it's not preferable to always use \left and \right to size "fences" automatically have evidently failed to address adequately a specific typographic issue you wish to resolve. At this point, I can only suggest that you post a new query, in which you ask explicitly for typographic, as opposed to merely TeXnical, answers and explanations. – Mico Jan 27 '19 at 06:56
  • @AndréC - Just to be clear about one thing: TeX and LaTeX do absolutely nothing randomly. Thus, I don't know what you mean by "the random side of \left and \right". For sure, the actions of \left and \right are utterly deterministic. Of course, "deterministic" does not imply "typographically correct"... – Mico Jan 27 '19 at 07:12
  • @Mico It seems to me that this question corresponds to this request to know why, which I understand as typographically why. Do you agree with me that the present question makes it possible to answer this typographical question? – AndréC Jan 27 '19 at 07:19
  • @AndréC - This posting’s overall style, and especially its why question, are IMNSHO not all that clear. Previous answers have attempted to shed light on how the indiscriminate use of \left and \right various fails to satisfy various typographic principles, but without stating what these principles are. You seem to be far more interested in understanding the principles themselves. – Mico Jan 27 '19 at 08:59
  • @Mico Yes, absolutely, but as a result, I will be told that my question is not related to TeX, whereas TeX is designed to respect typographical conventions, right? – AndréC Jan 27 '19 at 09:15
  • @AndréC - Don't be too pessimistic. Just explain in your query that you'd like to know if there's a structure to the list of reasons for not blindly using \left and \right. E.g., some reasons could be purely TeX-related (e.g., the prohibition against having \left and \right span rows of an align environment), some reasons could encompass a mixture of typographic and TeX-related concerns (e.g., the impossibility of line breaks in a \left...\right group), and a remainder of reasons which address purely typographic issues (e.g., spacing & correctly-sized "fences"). – Mico Jan 27 '19 at 10:30

0 Answers0