Optimizing markup for texml and LaTeX #165
Replies: 2 comments
-
In my opinion, it is stylistically better to avoid breaking equations---even if they are disturbingly long---for MathViewer. My reasoning is threefold:1
With that said, here are some of my observations. First,
or
or
Second, certain alignments are used for nothing more than getting a very long equation to fit on the page; again, see tran8916, Example 4.2. In cases such as
we can be almost certain that the original equation was simply too wide to fit on the page and needed to be broken for the print product. Conversely, in cases such as
we would not want to reformat this as a single line equation as the alignment has implied meaning (these equalities follow from each other and are related) and, visually (and unfortunately, only visually), it's easier to parse. In light of the first point, perhaps we could have TeXML automatically convert multlines to equations? Despite not being able to think of a good example, I could foresee this working flawlessly 98% of the time but the 2% of the time being enough of a problem that it would not be worth the trouble. For the second point, I think the best way to go about it would be to add some syntactic sugar (perhaps to the
and we are told to break it as:
If TeXML were to automatically flatten all alignment elements we'd end up with a spurious Footnotes
|
Beta Was this translation helpful? Give feedback.
-
Some thoughts #165 (comment) I generally agree that it's better not to break up equations. However, for very different reasons and with different consequences. The reason I favor to not unnecessarily break up long expression is that I hope we can develop means of enabling natural reflow in the resulting web-based outputs. Having hard breaks plays against that. A good example are very large inline expressions. We currently have to break them up manually. However soon won't have to because we will get that automatically downstream. Similarly, I hope that we can figure out ways to differentiate tabular layouts that are actually table-like (e.g., tables, matrices) from those used for alignment. Then we can consider downstream methods to reflow such content more naturally. I don't think flattening tabular environments is a good way to get there however. Alignment information is very valuable for good reflow, too. I'd rather explore ways to match some tabular environments to reflow-optimized markup instead of just tex-math. Finally, I want to comment one detail:
While that's true for content delivered via the open web, this is not true for the more limited world of epub. Here, we cannot make things scrollable (because most epub reading systems remove/override the relevant CSS). As a result, our only means is to shrink-to-fit equations. As a result, very wide expressions will become illegible on small viewports. (And sadly most epub reading systems have poor zooming features.) |
Beta Was this translation helpful? Give feedback.
-
Let's discuss how to handle cases where the optimal markup for LaTeX/print is not the optimal markup for texml.
To begin with, let's start listing the kinds of things you're running into, preferably with references to specific papers.
Beta Was this translation helpful? Give feedback.
All reactions