In search for the perfect academic (math-heavy) software: Markdown, remarkjs, Deckset and IA Presenter
I became bored with LaTeX Beamer a long time ago. The temptation to simply throw in a bunch of math and have an insane number of bullet points is something that everybody faces when using LaTeX Beamer, and there is no way to get around it. Don’t get me wrong, I have seen many skilled academics do amazing things with it, using various color coding schemes to highlight and emphasize parts of the equation so that the ideas stand out, using figures and animations to explain the intuition behind the mathematical techniques. So it can be a powerful thing once you know what to do with it. But the amount of effort to get all of this going is fairly high, and in the end, because it is not a visual medium, it ends up being very painful to see some presentations. The best thing to do is to switch off from looking at the slides and focus on the speaker. And don’t get me started on the backslashes and Latex environments — they are painful to write and read in source mode.
A couple of years ago, there was a tool called LaTeXiT, which would help you typeset math equations and drag them as an image or a PDF in presentation software like Keynote and PowerPoint. Fortunately, Keynote started to offer the typesetting of LaTeX equations directly, and this actually works very well. So I have been using Keynote for a while, and to a large extent, I think it is a decent enough presentation software. But again, I feel there is a temptation to just throw in a bunch of bullet points and hope that they do the trick, while the speaker just ends up reading these bullet points. I don’t think there is a way to get around the issue that, at least for technical presentations, we do need to have some textual content in there.
Of course, it depends on what kind of technical presentation because science is so broad. I can imagine that a paper that discusses empirical methodology could really be designed purely with plots and some descriptions, making it visually appealing, but this approach may not be applicable to a paper that introduces some aspects of set theory, logic or probability theory.
So in the context of exactly those kinds of talks — set theory, logic and probability theory — I often wonder what good software for presenting would be. With the typesetting of LaTeX, especially directly in Keynote, it is a good option, but I find that Keynote files are way too heavy and monolithic. So every time you want to use some ideas from a previous presentation in a new one, it is a very painful process — you have to open up files many megabytes in size, scroll through slides, copy material from one slide, and paste it to another. I think, in general, a slide becomes a kind of monolithic structure that you have to work around, and so in this way, I don’t think keynote or powerpoint is a good approach in general either because what we want to capture are ideas, and we cannot simply rely on the partitioning of ideas in terms of slides to make a good presentation.
So, I’ve been on the search for good presentation software, and I’m kind of pleased to note that I’m clearly not the only one thinking along these lines, and there seem to be two attractive and interesting candidates that are offering a worthy alternative to LaTeX Beamer but are a lot more flexible than Keynote or PowerPoint.
Firstly, I should note that using the markdown package for latex, you can write slides in markdown and get a beamer pdf out of it. I’ve done this, and can share examples for those interested. Secondly, there’s pandoc that converts markdown to beamer or remarkjs. This may work for many, but I find beamer outputs a bit boring, and find the styling of remarkjs to be a little off, can’t put my finger on it. Both beamer and remarkjs (and its derivates) are highly customizable, so there are workarounds, but couldn’t be bothered.
I have been toying with Deckset and the new software from IA: IA Presenter. Both of these are very interesting because you write everything in Markdown. Basically, the header styling of Markdown (hash symbols) can become the title of your presentation, and you use Markdown lines ( — ) to separate slides, so it becomes very easy to read and move material around in a single document without going through and shifting around a bunch of slides to extract what you need.
Now there are differences between the two, and at this stage, I’m not entirely convinced and sold on IA Presenter. Let me explain.
In IA Presenter, apart from the headers, the text you write becomes speaker notes. If you want to include actual content on the slide, you have to indent it. This is an interesting approach, and the examples they show on their website looks very attractive but completely impractical for a technical presentation because presumably, when you want to introduce technical and mathematical concepts, you end up indenting a lot of things, and this is just not fun and not very organic. In contrast, in Deckset, everything you write is actually displayed, but you can have speaking notes using the carrot symbol (e.g., ^ notes). This obviously means that when you write a lot of things, there is a temptation that this all ends up getting included in your slide. In contrast, in IA Presenter, these things would be hidden. The philosophy is really then of minimal amount of text by default.
Deckset allows you to edit source in an editor of your choice, and by using typora, I get real-time math display because of mathjax. I could use obsidian too.
Ultimately, I have chosen to use Deckset because of its traditional center-justified formatting for titles and text. IA Presenter’s two-column format looks fantastic, but also feels like a lot of space is left unused for some of the content I tried. Might change my mind later. However, I recommend trying both and choosing the one that best suits your personal preference. Compared to everything else, these are elegant and powerful. I know many use lots of animation (e.g., graph coloring) — I don’t really know if anything of that sort is supported with these, but I could be wrong.
