Resources related to the Roman world between 800 BCE and 500 CE.
SIG Coordinator: Chiara Palladino
The Tabula Peutingeriana in Recogito
May 12, 2016 at 1:43 pm #971
We’re having a conversation about how to integrate the data from Talbert’s Tabula Peutingeriana and other available datasets (Vici.org and Omnes Viae) in order to have a referenced Peutingeriana on Recogito, among other things.
But as you say, the references to the actual map only come in the form of “segment grid” codes. (Assuming that’s a regular grid, we could obviously match that back to a scan, which could give at least some searchability?) I agree it would be good to loop in the Roman SIG!
May 13, 2016 at 2:26 pm #973
The grid codes of Talbert’s Tabula Peutingeriana are indeed not really practical to work with, but every item of his database does have a unique number, which is contained in the address of each page, e.g. http://www.cambridge.org/us/talbert/talbertdatabase/TPPlace398.html. With these numbers it is possible to create all the Pelagios links necessary, I suppose.
On the other hand, the Trismegistos database did incorporate already most of the Talbert references in TM Georef (only the nameless icons were left out), and each of these references has its own stable identifier, the TM Georef number (cf. http://www.trismegistos.org/geo/authors_allgeoref_list.php?tm=380). If you want, also these numbers can be used to create Pelagios links.
My best wishes,
May 19, 2016 at 2:02 pm #998
@herbertverreth, thank you so much for this useful answer. Trismegistos is always ahead indeed (and we should definitely talk about annotating more material…let’s just make May end, then I will dedicate my attention to it).
So the identifiers are stable I suppose. The problem is that their connection to the graphic image is not as precise as we would wish it to be, to make it useful in Recogito. On the other hand, it would be a waste of good work to start annotating it from scratch. What do you think @rainer? Any other comments?
May 19, 2016 at 2:13 pm #999
indeed, we have the correspondence between Pleiades and the Talbert database URLs. (See e.g. the CSV on Recogito). What’s missing is:
- a good image (I guess there must be a reasonable-resolution one that’s open for re-use?)
- the correspondence between the Talbert grid IDs (also in the Recogito CSV) and the (relative or pixel) coordinates on the image.
Assuming it’s just a regular, sensibly numbered grid of equally spaced rectangles, it should be relatively straightforward to generate annotations for Recogito (v2) that link a grid square on the image with the places contained therein.
Obviously, just a rough approximation rather than “exact” annotations, but could be a start. I guess we’d just need someone who’d have the time to do it 😉
- This reply was modified 2 years, 3 months ago by Rainer Simon.
May 21, 2016 at 12:02 pm #1015
This is a great idea
It think it’s probably worth reaching out to AWMC for the image – they definitely have the best hi-res scan. In terms of the locations, I suspect that identifying the bounding boxes for each toponym is a manual tsk but a relatively quick one (especially if we, e.g. used it as an example for tutorials, etc.). Thereafter, there is the task of mapping identifiers and transcriptions to the bounding boxes. As these are already linked together it’s effectively a single task, and we could reduce the difficulty still further, by limiting candidates to those with the grid squares.
One tool, I’ve thought would be extremely useful for this kind of task would be something that extracts, scales and rotates bounding boxes from an image (i.e. displays a list of topoinyms details from the larger image, and then gives the user the opportunity to match them to some other data series (in this case the entries from the Talbert DB). I suspect it’s a fairly trivial coding task – maybe something for the Resource Development call?
May 21, 2016 at 1:47 pm #1016
I started playing around with this (I really ought to be doing some other stuff), but it’s not possible without knowing the angle of the bounding boxes. That’s not currently in the CSV download.
Rainer – can we include it?
May 21, 2016 at 7:54 pm #1017
Yes, I’ll make sure to include it in Recogito 2. Speaking of which: I wonder what a good data format would be for this kind of information. Right now, Recogito records the location of the boxes in terms of X/Y (pixel) coordinates for the “anchor point” (where the user made the first click), then the length and angle of the “base line” (which is drawn with the mouse drag), and then the height. It’s not very standard, I guess. But convenient in the sense that it’s compact and retrains the original orientation of the box explicitely (as opposed to a polygon).
I assume we’ll want alternative annotation modes in Recogito 2 as well – e.g. simple point and rectangle selection. But the “toponym box” we currently have may still have its merits.
May 21, 2016 at 8:16 pm #1020
February 7, 2018 at 1:37 pm #3277
Did you resolve this issue for v2? I have a much more fine-grained database of Tabula Peutingeriana coordinates than the Talbert grid IDs offer. It is used to generate the Tabula Peutingeriana Animated Edition at http://www.piggin.de and could easily be supplied as a CSV. My plot is 15,850 pixels wide, so this gives very high resolution and unique coordinates for every place name and gloss. The Piggin database is cross-referenced to the Talbert database and will soon reference TM Georef as well.
February 9, 2018 at 9:19 am #3278
You must be logged in to reply to this topic.