search

PreTeXtBook / pretext

PreTeXt: an authoring and publishing system for scholarly documents
https://pretextbook.org
Other
268 stars 209 forks source link

Simplify math markup #1365

Open davidfarmer opened 4 years ago

davidfarmer commented 4 years ago

The distinction between me and md can be determined by whether or not it contains an mrow. Therefore one of them can be eliminated.

The choice of whether or not to number a display equation currently made by the author, who chooses between me and men. I suggest this is at least partially a publisher choice. And if the author wants to make a suggestion, that can be done with an attribute.

I call it a publisher choice for two reasons. A low level book generally does not number most equations, because it is intimidating to the reader. So, if not a publisher choice, at lease it is a global choice that can be made at the time of processing.

My other reason comes from an experience writing a research paper many years ago. I was mentioning a paper by someone else, and I wanted to refer to an equation in that paper. That equation was not numbered, so I had to say something like "the second display equation after Theorem 4". I hold the publisher at fault, and not the author, for failing to number every equation in that paper. Multiple times I have had to point out the flaw in the argument "I just number the equations I plan to refer to." Another example of LaTeX dropping the ball by giving the author too much to worry about.

I think it is possible to temporarily make the code backward compatible, deprecating me, men, mdn, and providing an automated way to upgrade.

This issue based on a comment by @rbeezer in this thread: https://groups.google.com/g/pretext-support/c/Kj3egFnlVjA/m/Ks8F7CWzBAAJ

kcrisman commented 4 years ago

I call it a publisher choice for two reasons. A low level book generally does not number most equations, because it is intimidating to the reader. So, if not a publisher choice, at lease it is a global choice that can be made at the time of processing.

That could be a reasonable argument.

My other reason comes from an experience writing a research paper many years ago. I was mentioning a paper by someone else, and I wanted to refer to an equation in that paper. That equation was not numbered, so I had to say something like "the second display equation after Theorem 4". I hold the publisher at fault, and not the author, for failing to number every equation in that paper.

I disagree on that, for two reasons. 1) Nowadays what publisher actually checks this stuff? Sad but true. 2) It is very annoying to read papers where every single equation is numbered when there are many. Sometimes you just use display mode (or at least I do) so that it is visible. Maybe you would count that as presentation, not content, but many times I've had to make precisely this move in order to ensure that inline math didn't obscure some nested root or something.

So I would recommend this ticket be for the (reasonable) long-term identification of me and md, and then open another issue for discussing exactly who, and when, should determine numbering, because I suspect that it would be a little more contentious. I personally would recommend that the default remain author-specified (whether with the current syntax or an attribute), but would not be averse to a publisher switch that either suppressed all or numbered all (though then there is the problem of inconsistent numbers between different publishers, whoever they are).

rbeezer commented 4 years ago

The technical analysis here is spot-on. I suspect a hard-nosed deprecation will cause too many people too much pain. :-(

I'm going let the publisher responsibility percolate...

On September 17, 2020 7:46:32 AM MDT, "David W. Farmer" notifications@github.com wrote:

The distinction between me and md can be determined by whether or not it contains an mrow. Therefore one of them can be eliminated.

The choice of whether or not to number a display equation currently made by the author, who chooses between me and men. I suggest this is at least partially a publisher choice. And if the author wants to make a suggestion, that can be done with an attribute.

I call it a publisher choice for two reasons. A low level book generally does not number most equations, because it is intimidating to the reader. So, if not a publisher choice, at lease it is a global choice that can be made at the time of processing.

My other reason comes from an experience writing a research paper many years ago. I was mentioning a paper by someone else, and I wanted to refer to an equation in that paper. That equation was not numbered, so I had to say something like "the second display equation after Theorem 4". I hold the publisher at fault, and not the author, for failing to number every equation in that paper. Multiple times I have had to point out the flaw in the argument "I just number the equations I plan to refer to." Another example of LaTeX dropping the ball by giving the author too much to worry about.

I think it is possible to temporarily make the code backward compatible, deprecating me, men, mdn, and providing an automated way to upgrade.

This issue based on a comment by @rbeezer in this thread: https://groups.google.com/g/pretext-support/c/Kj3egFnlVjA/m/Ks8F7CWzBAAJ