2011 (old posts, page 2)

Ancient Toponym of the Week: Krokodilopolis

Known to its inhabitants as Shedet, this settlement was known to Herodotus and other Greeks as Κροκοδείλων πόλις, the "City of Crocodiles" [1]. According Smith (via Wikipedia), Krokodilopolis was a center for the cult of Sobek. The Greek practice of naming foreign cities after a totemic animal also produced toponyms such as Kynopolis (dog), Latopolis (fish), and Lykopolis (jackal).

[1] Getzel M. Cohen, The Hellenistic settlements in Syria, the Red Sea Basin, and North Africa, 56: http://books.google.com/books?id=RqdPcxuNthcC&pg=PA56

REST in six lines?

Duncan Cragg makes REST easy:

Here's how easy it is to be RESTful - "Level 3" - in just six lines:

GET /resource/1 HTTP/1.1
Host: restful.com

HTTP/1.1 200 OK
Cache-Control: max-age=100
Content-Type: application/json

{ ..,  "next":  "/resource/2",  .. }

Pretty simple, huh? If you're surprised by my use of JSON, hold on a minute, I'll get to that.

I still think that get our heads around REST is the most difficult (and largely unaccomplished) part. The other obstacle is our frameworks: ones with REST baked in are rare and unpolished, mature ones often make REST difficult.

New KML for Pleiades

I'm happy to announce improvements to KML representations of collections and search results served up by Pleiades. New features include: snippets, richer descriptions, and my favorite: better representations of roughly located places. I'm going to illustrate the new features using Pleiades places that reference the Periplus Maris Erythraei, a sailing and trading itinerary covering the Red Sea, Persian Gulf, Indian Ocean (see also http://pleiades.stoa.org/places/39290), and their coasts. The collection of those places has HTML and KML representations and can also be displayed directly in Google Maps.

I've moaned before about KML's name, Snippet, and description elements, but Snippet is what we have to use to get the information we want to surface like an Atom summary or RSS description. The new Pleiades snippets combine a list of feature categories (such as "Settlement" or "Station") with a list of time periods from the Pleiades vocabulary ("Roman", in the PME case).


The same categories also show up in the new popup balloon style. "Tags" would be other user provided categorizations (such as material types for mines, deities of temples, etc). The logo, for what it's worth, is by me and the Gimp.


Last, but not least, we've turned our old representations of roughly located places inside-out: we map what we know exactly – the bounding boxes that contain the places – and link to the places in the KML descriptions. It's a bit like a clustering approach, but with semantics that are rare in GIS or neo-geographic applications.


A permanent link to the search results containing those same places closes out the new aggregate bubble style. One of the things I'd like to do after we see how many more matches we can squeeze out of the DARMC dataset is add an "edit" link so that users who have the precise locations (and evidence in the form of references and links) can address themselves to the "fixable" places.

The vultures return to Fort Collinstrano

It's almost time for our neighborhood's annual miracle, the return of Cathartes aura. I expect the turkey vultures to majestically descend on 920 W Mountain Ave while I'm out of town. If you record the event, could you send me a link to images online or post links to them in a comment?

Update (2011-03-29): I found the vanguard in their tree this morning.


"Write for yourself" and other good tips

My favorite bits of advice in Tim Bray's meta-blogging post: "write for yourself", "be intense", "lighten up", and "math is hard". I remember when I used to get hung up on how much I was getting read or being linked, but I'm mostly getting past that and just enjoying writing and getting ideas out there to flourish (quickly or slowly) or fail (the faster the better).



Author: Kurt Schwehr

Good advice! I stopped bothering with tracking about 5 years ago. The best feedback I get is from people who talk to me because of my blog, about what from my blog made a difference to them, and the collaborations that resulted.

Beyond the PDF and Scholarly HTML

Says Peter Sefton on Scholarly HTML:

If we can get this right we can help the web be what it was meant to be from the start, a platform for scholarship.

I've had my head down in Pleiades for a while and am belatedly becoming aware of the crucial efforts labeled #beyondthepdf and #scholarlyhtml. Open access to PDFs or Word docs, while a big cultural step, has never impressed me all that much; semantically marked-up HTML could make so much more possible – search indexes of hypotheses and methodologies, for example, or a traversable web of citation. Sebastian Heath, too, is working this ground for ISAW Papers. Hopefully, I'll get to represent and discuss that a bit next week, and consider what I can do to make the HTML of Pleiades more scholarly and interoperable.

Ancient Toponym of the Week: Ad Fines

Ad Fines is Latin for "at the border", and we've got 13 of these in Pleiades [HTML, KML]. By the second century AD these were well inside the borders of the Empire, but the name suggests that they were founded as frontier outposts.


Re: Ancient Toponym of the Week: Ad Fines

Author: Tom Elliott

Quite likely they might have been road stations or other stops coincident with provincial boundaries (internal to the empire) or on the transition from the extra-urban territory of one ancient city to another.

DH/Geo/LOD Workshop March 24, London

An upcoming workshop that I'd previously mentioned now has a page for signup.

The Pelagios workshop is an open forum for discussing the issues associated with and the infrastructure required for developing methods of linking open data (LOD), specifically geodata. There will be a specific emphasis on places in the ancient world, but the practices discussed should be equally applicable to contemporary named locations. The Pelagios project will also make available a proposal for a lightweight methodology prior to the event in order to focus discussion and elicit critique.

The one-day event will have 3 sessions dedicated to:

  1. Issues of referencing ancient and contemporary places online
  2. Lightweight ontology approaches
  3. Methods for generating, publishing and consuming compliant data

Each session will consist of several short (15 min) papers followed by half an hour of open discussion. The event is FREE to all but places are LIMITED so participants are advised to register early. This is likely to be of interest to anyone working with digital humanities resources with a geospatial component.


Perl+RDF Hackathon the week after

Author: Kjetil Kjernsmo

The week after, I'm organizing a

"Semantic Web with Perl hackathon" in London.

If any of the participants on the Pelagios workshop would be interested in contributing to advance the state-of-the-art of Perl tools to work on LOD, please let me know.

My favorite Pleiades update yet

I'm extremely happy to have finally improved the coordinates for Pleiades around SW France and NE Spain.


The year my family and I spent touring around these regions was amazing. One of my biggest regrets is never having stopped in Narbonne; I've heard that the ancient Via Domitia is exposed right in the centre-ville.

I've got a KMZ of these places the curious: http://dl.dropbox.com/u/10325831/Updates%202011-03-04.kmz. It has network links to 3 Pleiades collections.


Re: My favorite Pleiades update yet

Author: KaCeBe

You are right. You can see the Via Domitia in Narbonne, but you're not allowed to walk on it...


Python and WFS?

Searches for "wfs python" are driving visitors to my blog all of a sudden. I'm sorry that there's not much here on the subject in the last 4 years or so. In theory, Python should be a great fit for OGC web feature services. The language allows runtime creation of types, and so the descriptions of feature types discovered in a WFS can be turned into actual Python classes on the fly. All the benefits of the Python type system without the brittleness you'd expect from compiling code against the WFS capabilities and feature type descriptions. But we don't see much of this in action. Python's XML libs are adequate – the lxml interface to libxml2 is killer, even – so there's no lack of a foundation. Is it a cultural issue? Has WFS become a technology that's only used by "viewer" type clients? Is potential WFS client work going into Javascript or desktop (and this means .NET/C++/etc) clients instead? Maybe the searches mean the situation is changing?