Knowing what to look for (and at)
It has been a problem of our digital edition for about a year now, that the new interface, as cute as it is, is conceived for a much bigger amount of corpus than we now have. And whether with the amount we have now or with a bigger amount of texts, it does not provide for any kind of overview or orientation. I thought this was not to be helped, considering we have reduced the displayed information to a minimum that is, in my understanding of it, also a maximum.
But then Marian Dörk, whom I have been thinking about visualization of our data with, had the great idea to hand over to one of his students datasets from our edition. The result can be viewed here and is interesting on several accounts. (first, let’s all agree on the basic statement that this is an experiment, not a final product, right?)
– The visualization of the encoding layer that gives information on the network inherent to the correspondence is totally convincing, whereas the rendering of all the genetic phenomena (corrections, erasures, elements above or under the line) was not to be solved by a snap of the fingers.
– The frame elements providing for orientation (how to situate this specific correspondence in the whole edition, this particular letter in the correspondence and in the whole edition) would need almost no improvement to be implemented in our edition, I think.
– The whole thing offers a real reading comfort, which I thought was not really conceivable online.
The discussion I had with Jonas Parnow (the student who realized the visualization) when he presented his results to me was also very interesting. His visualization choices were completely independent from the research question that underlies our edition. When he came to ask “But why wouldn’t you…” in order to convince me to pick his visualization option over ours, I had to roll out the hours and hours of discussions we have had over the two last years, about what feels like every single pixel in the page. There was none of his objections where I wasn’t led to lecture him for (very much too long) minutes about what we are doing and why we are doing it this way. Not displaying the manuscript is not an option, because displaying it is a statement about what kind of edition we are doing. And we can’t display bibliography, biography or other content notes in the pop-ups, because it is the only way we have to found to render the variations in the hands, which are also kind of our trademark. To sum it up: you cannot separate the visualization choices from the general scholarly orientation of the data you are working on, and the data don’t always by themselves tell you everything about the expectations they are supposed to answer to.
Which also means that, if you want someone to realize a visualization, you have to tell them what you are looking for. That speaks for another meeting with Marian…