Handling of \frametitle and \framesubtitle

[josephwright]

The optional-group syntax in beamer is not one we want to carry forward. That leaves us with the question of how to handle frame (sub)titles

\begin{frame}{mandatory-frame-title}

or

\begin{frame}
  \frametitle{Optional frame title}

or something else. Subtitles make this more tricky: it's hard to see they could be mandatory, but they are likely wanted.

[samcarter]

Personally, I've always been a fan of the

\begin{frame}
  \frametitle{Optional frame title}

syntax. In beamer it made the frametitle a lot more flexible to use, e.g. for overlays.

---

Just to throw in a different approach for brain-storming: how about key-values in frame options, e.g. \begin{frame}[fragile,title={Quack},subtitle={Breakfast}]

[Skillmon]

Wanted to throw in the key-values as well.

I don't think that changing the \frametitle depending on an overlay is a good idea, that should be an entirely new frame (also in view of tagging in the future this will make things easier).

[FrankMittelbach]

I have quite a number of decks where the frame title varies depending on the overlay inside and I don't think that makes it several frames.

But I never used \frametitle because of that: I always altered the title argument using \only or \visible inside, ie the flexibility is there without the need for a separate command

[u-fischer]

From the view of accessibility and reading order you have in such a case several frames, simply because you can't in sensible way collapse that into one "handout" frame. E.g. the Question/Answer titles @samcarter's example obviously can only be read if you view them as two frames, regardless about the technical implementation.

[josephwright]

Indeed - whilst I'm not keen, I suspect we may end up having to allow more than one handout page per frame for this reason. /That said, for e.g. questions and answers, I'd expect them to simply be in one page intermixed, as I'd show the questions 'fill in the blanks' style and simply have the answers in the right places in the handout.

[FrankMittelbach]

From the view of accessibility and reading order you have in such a case several frames, simply because you can't in sensible way collapse that into one "handout" frame.

In my terminology that still remains a single frame (but of course with several slides to put on top of each other).

Overlays come in two forms a) uncovering and b) rewriting+covering. In case of rewrites that certainly poses a question how to represent that in a different medium and yes in that case you may have to offer slide by slide unless you try to explain in words what gets covered, what replaced and what added. And I guess we agree that this would be very difficult to do and even if possible probably not "accessible" aka understandable in that case.

So in my opinion, tagging a deck could be done as simple handouts frames when all the overlays do is uncovering and the moment you do more "interactive" stuff the tagging needs changing and perhaps initially simply by representing each overlay slide on its own (which means a lot of repretition for the consumer, but ...)

[samcarter]

I don't think that changing the \frametitle depending on an overlay is a good idea, that should be an entirely new frame (also in view of tagging in the future this will make things easier).

I find things like

\frametitle<1>{Question:}
\frametitle<2>{Answer:}

very useful. This avoids a lot of code duplication compared to using two separate frames.

[josephwright]

Just to throw in a different approach for brain-storming: how about key-values in frame options, e.g. \begin{frame}[fragile,title={Quack},subtitle={Breakfast}]

I'd thought about that (should have listed it earlier), but I worry

• That for many people this makes an 'optional' argument mandatory
• That it heavily intermixes two very different types of data - typesetting material (titles) and control info (e.g. fragile)

[josephwright]

If we do decide the frame title is like a theorem name, then does that mean that subtitles don't generally make sense? I'm not seeing much push for subtitles (and we could have say that a split at some pre-defined token in the title denotes a subtitle - if needed).

[cfr42]

I do use subtitles, but I wouldn't put much energy into defending them.

[FrankMittelbach]

If we do decide the frame title is like a theorem name, then does that mean that subtitles don't generally make sense? I'm not seeing much push for subtitles (and we could have say that a split at some pre-defined token in the title denotes a subtitle - if needed).

I wouldn't say that they do not make sense; subtitles are just structuring a title. We don't have it but you could imagine that for theorems or section headings etc too. They certainly make sense in the "visual" world and as far as tagging is concerned you could encode that too in some way (or initially just turn the concatenation of title + subtitle into a single data object (loosing some information there but ...).

As to the interface on doc level, it could be key/val or it could be some token as you suggest. Not sure which would be a nice one though. Maybe even \\ but that would then prohibit using this for denoting a line break in a large title (and you would have to resort to \newline for that to avoid the subtitle

[josephwright]

I was thinking that if we decide that subtitles should be handled within a frame title, then we'd have a dedicated marker (a bit like \and in the author list). But at the moment this is all up in the air - we need real presentations with structure to explore.

[Skillmon]

I used subtitles extensively in the past, grouping a slew of frames under the same frame title (well, in my case \section really), and giving the individual frames different sub titles. Something like

• The Algorithm
  • The idea
  • Basic equations
  • Optimisation criteria and cost functions
  • Extending the algorithm: Kernel functions

So I'm definitely on the pro-subtitles team.

[Skillmon]

I don't think that changing the \frametitle depending on an overlay is a good idea, that should be an entirely new frame (also in view of tagging in the future this will make things easier).

I find things like

\frametitle<1>{Question:}
\frametitle<2>{Answer:}

very useful. This avoids a lot of code duplication compared to using two separate frames.

Don't get me wrong, I didn't mean it couldn't be useful (and I see the appeal), I'm just wary whether the advantages outweigh the implications. That's something you two are certainly better suited to assert.

[josephwright]

Looking back, I should have added I do wonder about fragile - we could reasonably simply handle all frames that way all of the time (yes, it's slower but possibly the balance here is more toward utility than speed).

[FrankMittelbach]

you could have both, i.e., all frames doing fragile handling by default but offer robust-frames=true/false on package load (with default=true) and those concerned about speed (and I have decks with 500+ pages) can set it to false and then need to add "fragile" whenever necessary.

[cfr42]

Compilation speed is a huge deal for me. I very rarely need fragile.

[samcarter]

One point which might be against \frametitle: the other approaches might make it easier for xbeamer to know already at the start of the frame if there is going to be a frametitle.

• the frametitle could be set in the header, which would make it easier to create layouts similar to the Berkeley theme
• a more meaningful textheight could be calculated

[josephwright]

As we collect the body of the frame, we can always simply do a search up-front for \frametitle and the following balanced text, and extract before we typeset the content. So in that sense it's not so important. On the other hand, that's obviously a bit of work we can avoid if we only use an argument to \begin{frame}.

[wspr]

Since this discussion has become somewhat broader, I thought it would be good to give a couple quick examples from my own course notes.

See first example image below -- the top four lines of the slide are section, subsection, frametitle, framesubtitle. Sections are used to delineate "topics" in the course, and subsections used to essentially mark where I end one video recording and start the next ("subtopics" or "key concepts"). I use the "lecture" subdivision to mark "weeks" (each week contains a small number of topics) but it would make more sense to me to just use \part for the higher level of heading than \section.

Each topic starts with a table of contents -- see second image below. The fact that almost all of this functionality comes "for free" in beamer is essential to preserve in my opinion. I think it would be unfortunate to lose the ability to create collated course notes from individual source files and structure the PDF in an appropriately hierarchical way.

I thought I used subtitles more than I actually do, based on a quick scan over this slide deck. I like their existence -- I don't mind Frank's suggestion to concatenate it with the frame title for tagging.

[wspr]

(please excuse the typo in the description of the maths, this is an old version of the slide deck I had on hand)

[josephwright]

Each topic starts with a table of contents -- see second image below. The fact that almost all of this functionality comes "for free" in beamer is essential to preserve in my opinion.

Yes, definitely we need that - I use a TOC per section myself in longer series of lectures, so it's on the 'draft, that's broken' list at the moment :)

[josephwright]

@wspr Very useful to see someone else's teaching content - thanks - I'm guessing the two columns are blocks?

[josephwright]

Or is this just two columns with free text?

[wspr]

@josephwright — just \begin{columns} \column{0.5\linewidth}... or whatever the syntax is :) I actually haven't used blocks much... something for me to keep in mind!

[josephwright]

I've been thinking a bit about frame titles, specifically as they look like we should tag them as Captions. That means that frames look quite like floats (an env with a caption), and that suggests to me that the \frametitle approach is likely the best for consistency with the kernel. [Yes, in many ways one with hindsight would do \begin{figure}{caption} or similar (cf. ConTeXt), but that ship sailed in the mid 1980s.]

Going with \frametitle allows for adding an overlay spec, so there are advantages. We can happily have \framesubtitle too here.

So unless there are strong arguments, I think we'll likely stick with the current model.

[cfr42]

I use subtitles quite a lot - probably similar to the way other people have described using them. Like most people's, my slides are now on a VLE behind a wall, but I have some older ones which aren't e.g. https://cfrees.wordpress.com/wp-content/uploads/2012/01/2-persons-h.pdf. Note these are all created with handout mode. In class, I'd use a version with overlays, but the use of subtitles is the same.