[OHM] tagging for historic race tracks / NEH

[Rob H Warren]

Jo,

Given that XMLSchema dates (ISO8601) are the basis for most WWW date and time, I would venture a guess that it has been a direct help to anyone reading this message. And should you know of anyone looking at grant funding for Time OWL, kindly send them my way. A lot of people are desperate for a better "people time" solution and Time OWL is currently the only such solution. I use it.

To paraphrase a colleague "The nice thing about standards is that there are so many of them to choose from" (See also [1]) . This said for all of the many problems with WGS84 (aka EPSG:4326), it is not a stretch to say that it currently runs a big chunk of civilization.

I am not a fan of GML or KML, but a number of javascript libraries will draw on the screen in some reasonable way if you give them a string in that standard. GML is also the only spec that I have seen deal with precision and accuracy. This is a problem for me in the automated extraction of features and I insist it be something that  OHM should tackle, probably not with GML. Whatever solution we come up with will also be useful in automating some of the 2D rendering decisions that are currently hard-coded. 

Jerry's routing over historical data is a nice prototype and foreshadows cool applications with historical simulations of say, the barbarians sacking Rome or railroad goods distributions in Victorian England. But It's a prototype. Not a tool yet or a major piece of infrastructure that everyone is going to pound on. 

What makes that prototyping possible is that the 'core' tools have been solid and consistent enough for others to build around them. This thread is a necessary exercise to get us to that same point.

[1] http://imgs.xkcd.com/comics/standards.png

On Dec 29, 2014, at 2:47 AM, historic-request at openstreetmap.org wrote:

> Message: 2
> Date: Mon, 29 Dec 2014 07:47:04 +0000
> From: Jo Walsh <metazool at fastmail.net>
> To: historic at openstreetmap.org
> Subject: Re: [OHM] tagging for historic race tracks / NEH
> Message-ID:
> 	<1419839224.2474479.207519109.6AE599F5 at webmail.messagingengine.com>
> Content-Type: text/plain
> 
> My problem with a lot of this material is it's very theoretical basis in
> standards world. I'm seeing a lot of keywords for pure-standards
> initiatives without implementations and know that they never took off
> during the last decade. Who's using GML outside of the mapping agencies,
> as a thin transport layer? Has XMLSchema ever been of any direct help to
> you? Has Time OWL got any use outside of academic prototypes for more
> grant funding to work on Time OWL applications?
> 
> I look at SK53's prototypes for historical routing based on OSM data and
> I'm pleasantly amazed. What will the side-effects be of historical
> routing? And the adhoc, rapid-iteration approach of OSM allows you to
> work that out. So I don't really understand why you would be wanting to
> replay this decade of failpath standards development in order to model
> hypercorrect data that you'll almost never going to find.
> 
> Historical name search looks like an easy place to start for OSM. Think
> about a temporal aspect to Nominatim? Look around at some of the work
> that Herbert van de Sompel did with Memento and temporal WMS, and
> whether that worked out for them or not. Dig into the serious archival
> problems, how to keep the data consistently readable for the next 100
> years. This sort of thing?