Scientific Communication As Sequential Art | Hacker News Reader

Scientific Communication As Sequential Art

worrydream.com

46 pointsrileyt15 comments

Looking at this, it makes me realize just how much harm the journals are doing by limiting the number of figures per paper (and sometimes charging per-figure as well). If it really meant smaller, simpler, papers it might be a reasonable restriction, but in my field figures with 10 or more panels -- reproduced at a tiny size which makes them pretty much useless when reading offline -- are commonplace. They end up completely divorces from the flow of the manuscript.

Putting all the figures at the end (or separately) when the manuscripts are sent out for review just makes matters worse.

I haven't yet seen any charging for figures, only charging for colour figures (which probably really significantly increase cost). But since I don't know anybody who reads papers on paper it doesn't really mater anymore. And you might make the point that making sure that your figures loose as little information as possible when viewed on greyscale also improves accessibility.

I really want to like (and buy) this "publication as a richer experience" as we need to find a way of getting beyond are book-like mentality in building documents that will essentially be read on a computer anyways. However, I don't really understand how to apply this technique in a way that could be successful and fairly generalized (i.e., it could become my new writing style). For better and worse, static linear prose allows the author to control the story that the reader is perceiving completely, it is quite efficient especially if the author can actually write! Anything else, like figures with verbose captions or lists, is often not very effective and a paper smell (like a code smell).

We'll need many more experiments like what Bret Victor is doing here before we can successfully evolve what is basically a 7000 year old format.

I use more verbose captions lately. It makes the figures be comprehensible on their own. The running text is shorter; The passage is mostly copy-pasted into the caption.

The caption approach often doesn't work very well in prose. The reader doesn't really want to divert their attention to a long caption, which interrupts their reading of a paragraph, if they bother they might have a hard time going back to where they left off after they've taken in the caption. This is why captions should be fairly short, especially when your reader is a reviewer who is not specifically interested in reading your paper. I actually prefer inlining my figures into the paragraph text, especially if they are short or math oriented. Reference becomes harder, but the reader's flow is better preserved.

Now, we can avoid that problem by moving to an illustrated comic book style (as done in Bret's example), where the figures become the story and panels are read in lieu of prose. But its very difficult to do a whole paper like this.

Reader flow is not a problem, if the figure+caption is used to make a point, which can be skipped for flow. Example from a recent publication [0], where the actual figure is not necessary, so the reader's flow is not broken:

> The libFirm API provides serialization and deserialization support for the intermediate representation graphs. As can be seen in Figure 3, it is therefore possible to compile the C part of the standard library with cparser, stop before actual code generation and instead export the intermediate representation to a file.

I believe most readers (e.g. me) look the figures first, so I want the figures to be comprehensible on their own.

[0] http://digbib.ubka.uni-karlsruhe.de/volltexte/1000028112

I'm willing to bet that any revolution in this field will involve Victor; one of the most fascinating thinkers in all of HCI.

I am really blown away by this man's innovation and productivity. A true artist, low key, strike high.

http://worrydream.com/ClickShirt/

This is how scientific publications should be at our technological level. Not a multiple-pages-based pdf.

However, the details still need some iterations. The design breaks when the browser windows gets narrow. Does it work on a phone or tablet?

This looks very similar to the popular conference/workshop poster presentation format. Many disciplines have these sessions at meetings where authors display an A0/A1 poster summary of the paper and discuss it with attendees who wander around the room with a list of posters like art-gallery visitors. It can be very effective, with crowds around interesting posters and some debate. Often the visitors have a glass of wine in hand. In any case the posters look just like the posted page, though this is a very elegant example.

The authors don't mention the existence poster format, but it is clearly very closely related. Notice how politely I put that.

To see how common this is, Google for "conference poster" images. Many of the results returned are ugly and overdone - the idea is often executed poorly.

Conference posters are not considered a "real" publication, and the ones that I have seen are not as carefully designed as this example.

Surely, indeed I made your second point in my comment. Some posters are beautiful, most are not. My complaint about the OP is that it doesn't acknowledge the existence of posters, which are in exactly the format presented.

Then I did not get my point across well: I have never seen a poster in any approximation of the form the author presented. Hence, I find it reasonable that he did not mention them, because he likely has also never seen them.

Unrelated to the content of the article, but the light grey text is almost entirely illegible. The figure captions are particularly bad, and I am surprised Bret Victor chose such an awful color scheme. This page is basically a perfect case for http://contrastrebellion.com/

Reply on news.ycombinator.com