search

APEXCalculus / APEXCalculusPTX

In-progress conversion of APEX Calculus to PreTeXt
6 stars 20 forks source link

Some global housekeeping tasks #25

Closed sean-fitzpatrick closed 2 years ago

sean-fitzpatrick commented 5 years ago

A few things have come up in the last few pull requests that can/should be fixed globally:

These should all be easy to handle with a bit of global find/replace (hand editing needed for the first one). I need to do a build of Chapter 5 to double check my work there. But if someone has time to look it over, we should merge #20 this week, and then I'll do a round of editing.

This should get us to a working Calculus 1.

davidfarmer commented 5 years ago

Additional global issues:

davidfarmer commented 5 years ago

Consistency of th (and st, nd, and rd) as exponents vs inline. (I think inline was chosen as "official" in the online discussion, but we can have that discussion again in the context of this book.)

davidfarmer commented 5 years ago

Size of the partition should be a macro so that \lVert and \rVert are not hard-coded in the source.

davidfarmer commented 5 years ago

Always put a 0 before the decimal point for a decimal number between -1 and 1.

davidfarmer commented 5 years ago

Consistent usage of \fp and f'. See for example lines 174 and 175 of sec_deriv_inverse_function.ptx . Let's discuss this before making any changes, because there may be some issues that not everyone is aware of.

Later we should turn (most of) this issue into a checklist for authors/publishers who are proofreading their work, to be posted on ptx-announce and put into the guide.

sean-fitzpatrick commented 5 years ago

I'll look these over. We have a \norm macro for \lVert and \rVert. This wasn't used by Greg in the original LaTeX though, so some will have snuck through.

I did notice that in some cases the third derivative is set as \fp'' and in other places as \fpp'. We should try to make that consistent. Not sure which is better though.

sean-fitzpatrick commented 5 years ago

We also need to make a design decision on exercises with images. Right now we have a lot of exercises in Chapters 3 and 5 that are encoded as WeBWorK problems, but they also include one or more TikZ graphs. (There's some discussion on the forum between @Alex-Jordan and I.)

Eventually TikZ in WeBWorK might be a thing. But it isn't now. We can either: (a) remove/comment out the lines that make these into WW problems, and keep the TikZ graphs. (b) comment out the TikZ graphs and replace them by pg graphs.

The pg graphs generated by WeBWorK do not look nearly as nice as TikZ graphs, and it would be extra work to create them. But keeping TikZ means these become static problems, and it certainly would be nice if everything was a WeBWorK problem.

I'd like to get an opinion from @APEXCalculus on this.

sean-fitzpatrick commented 5 years ago

Another thing to watch out for, because it causes some serious display errors: During the last conversion, the \lz (Leibniz) macro was introduced for derivatives in Leibniz notation. But it's not always implemented correctly.

See for example the section on implicit differentiation. Things are entered in the wrong place, and we get output like: $\frac{d}{dx^{1/2}}(=)\frac12x^{-1/2}$

(I guess there's no math support in GitHub markdown.)

sean-fitzpatrick commented 5 years ago

I am wondering if perhaps one size does not fit all for WeBWorK graphs. Right now, almost all graphs are defined with size 400 x 400. Many are loaded without width attribute. See, for example, problems 5-12 in the section on one-sided limits

So these are being scaled to 600 x 600. Probably too big. But maybe 47% width is too small if students are being asked to extract numerical information (like limit values) from the graph. You can click on a graph, but the pop-out image doesn't get any bigger. Maybe 400 x 400 was reasonable, and we just need a width="67%" to match it.

Edit: the second on continuity has some webwork graphs with width set to 50%. I think they look OK.

sean-fitzpatrick commented 5 years ago

One other global task: we need to strip seed="1" from WeBWorK tags. For pages with many graphs, like the one linked above, we run into the bug observed in Portland, where each new graph replaces the ones before it.

sean-fitzpatrick commented 5 years ago

Looks like WeBWorK images will need to be wrapped in a sidebyside. (Or the mbx script needs tweaking.) Explanation to follow on pretext-dev

sean-fitzpatrick commented 5 years ago

Done tonight:

Next up: ndash and WeBWorK images.

sean-fitzpatrick commented 5 years ago

I didn't manage to find many instances of punctuation problems with display math. I expect I was looking for things like

<me>
  (some math).
</me>

instead of

<me>
  (some math)
</me>.

I'll watch for it though. If you saw specific examples, let me know. It could be my search terms for this weren't quite right.

sean-fitzpatrick commented 5 years ago

Fixed use of \ dx instead of \, dx. (Plus all other iterations, like ds, dS, dm, d\theta, etc.) Can't guarantee that I didn't miss some cases where there is nothing in front of the differential. But there shouldn't be many. Greg's LaTeX was generally pretty good about this.

sean-fitzpatrick commented 5 years ago

I think the only things left are usage of <ndash/> and the f' vs \fp issue. The former will require some care. I didn't tackle the latter because:

sean-fitzpatrick commented 4 years ago

I think most of these have been dealt with. Going to go <ndash/> hunting now

sean-fitzpatrick commented 2 years ago

I am pretty confident that everything in this issue has been dealt with.

Further formatting problems can go into a new issue.