Discussion and development of gazetteers and related resources.
SIG Coordinator: Wolfgang Schmidle
April 14, 2016 at 10:51 am #811
I wonder whether we can use this as a platform, to begin a discussion about Pelagios and PastPlace.
I, for one, have had quite a few questions about the gazetteer. Many of them have been answered in the post above, but a few remain – I’ll post more about that later. If any others of you Recogito annotators have questions about PastPlace that Humphrey hasn’t addressed, then let’s hear them.
April 14, 2016 at 1:10 pm #817
I’ll post issues separately.
As I said, some aspects are already dealt with in Humphrey’s post. I’ll mention them anyway. Take what’s worth keeping, and then (as the poem says), with the breath of kindness, blow the rest away.
Wrongly located places. I’ve only come across a couple of these: Banbury (moved to Suffolk) and County Dublin (moved to Galway). In both cases there was a full and accurate description; they’d just been marked in the wrong spot. And for both the latitude is about right, so it’s probably a typo on the coordinates.
What’s the process for correcting these errors? Do we inform the Pelagios team, PastPlace, Wikidata, or all three? Could or should we make our own corrections to Wikidata? Will those corrections automatically feed through to PastPlace and Recogito? Might those corrections already have been made in Wikidata, but not yet fed through to us?
I suggested a Recogito gazetteer-in-limbo – perhaps just a plain text feature – where mistakes can be noted, and remain until someone decides what to do about it. The process needs to be simple, or there’ll be a temptation to let errors go unreported.
- This reply was modified 2 years ago by damien-bove.
April 20, 2016 at 5:06 pm #850
The Who’s On First gazetteer may also be an interesting resource to consider.
Detailed information ( and the actual gazetteer , source code for an application for exploring the data, etc. ) are available from the following URLs:
This particular project has several interesting characteristics:
- By establishing “cross-links” between multiple resources (and by now the authors are also considering WikiData, besides established resources such as geonames), the authors were able to derive polygonal geometries for the boundaries of most of the places that are described in the gazetteer. I think this is probably the most interesting feature in this particular gazetteer. Moreover, they also have temporal periods of validity for some of the features that are listed in the gazetteer. The detailed polygonal geometries are very useful, in the sense that they can better support the development of cartographic representations.
- The authors are using a simple JSON format for exchanging gazetteer records, which nonetheless considers features such as provenance of the resources, polygonal geometries, temporal periods of validity, concordances to other gazetteers, hierarchical/administrative relations, etc.
March 30, 2017 at 5:00 am #2531
Picking up this old thread – I wonder if anyone has been working with Who’s On First (mentioned above, lives here, made by Mapzen)?
@bgmartins – Do you have further experience with WOF? I’m intrigued by their efforts to play nicely with other gazetteers/records, and the potential for incorporating more historical data.
April 6, 2017 at 6:51 pm #2638
I’ve used WOF and I’ve been following the project.
The idea of having places associated to polygonal boundaries is really great, but I’m not sure how they are currently performing the alignments towards other gazetteers. The updates on the “regular” place types have also been somewhat slow, and it seems that they are mostly focusing on “venues” and on keeping a very detailed list of small “points-of-interest”. The data model can accommodate historical data, but they aren’t going on that direction.
Some interesting resources:
- There’s an updated description at : https://whosonfirst.mapzen.com
- Statistics on the number of places that they have : https://whosonfirst.mapzen.com/spelunker/placetypes/
- Statistics on the number of alignments that they have towards other gazetteers : https://whosonfirst.mapzen.com/spelunker/concordances/
April 7, 2017 at 1:18 am #2639
April 17, 2016 at 9:19 pm #840
No suitable gazetteer match available
This might happen where the place is created in the mind of the author. I’m not talking about those – though a gazetteer of imagined places is an interesting idea.
Some weeks ago I couldn’t find Birmingham (England) in the PastPlace gazetteer. It’s there now; perhaps I missed it before, or perhaps it is only recently added. In any case, a missing Birmingham would be a clear oversight that needed to be remedied.
But what about the smaller places? The Gough Map shows a place called Ogerston (a house owned by the Knights Templars. Edward I stayed there). Do places like this merit inclusion in PastPlace or Wikidata? If yes, what’s the best way to get them listed? If no, how are they to be dealt with in Recogito? It seems unsatisfactory to dismiss them as ‘unmatched’ – especially if the location is known.
EDIT: A gazetteer that included every royal stopover might be so bloated as to be unusable. This might be what Humphrey will deal with in his next blog post. From my narrow experience, it would help to have the places sorted into broad categories. Is that sort of thing possible or practical? I can’t quite think how… but when looking for a town, it would be easier not to have to scroll through ballet schools, battles and business districts. Of the following list, I needed only the first three.
- political geography, settlements etc
- physical geography
- individual buildings (castles, forts, palaces, abbeys, houses ???)
- historical events (is there any reason to mark these on Recogito?)
- modern addresses and businesses (not needed)
- This reply was modified 2 years ago by damien-bove.
You must be logged in to reply to this topic.