Geospatial XG Teleconference

11 Dec 2006


Josh_Lieberman, Andrew Turner, Chris Goad, John Goodwin, Peisheng Zhao
Josh Lieberman
Josh Lieberman


OWL definition of GeoRSS and use in tagging HTML

[12:32pm] ajturner: Allan brought up some points on the GeoRSS blog about diff't uses for featuretype - more formal "mountain", "restaurant", etc. but also things like "hallway", "cafe", etc.

[12:37pm] joshli: Introductions: John Goodwin - OS ontology modeling

[12:37pm] joshli: Andy Turner: developer of geoblogging software and georss contributor

[12:38pm] joshli: Chris Goad: works for Platial and also worked on rdf

[12:53pm] ajturner: sorry - I guess my phone died

[12:54pm] joshli: Peisheng: works at GMU on Web Services composition

[12:59pm] joshli: Discussion of whether we need a "placeholder" feature object to hang the georss properties on.

[1:08pm] joshli: Action: Chris and John Goodwin will collaborate on an OWL /RDF proposal for GeoRSS

[1:08pm] joshli: Issue of microformats: three proposals for use, no real consensus yet.

[1:15pm] joshli: Discussion: basic ontology components (review)

[1:33pm] joshli: Discussion: SOCoP collaboration, working on foundation ontology components form which to run a domain ontology pilot project

[1:36pm] ajturner: what MF proposals are there?

[1:37pm] joshli: 1. Use span and div and stuff coordinates into attributes like "title".

[1:37pm] joshli: Ugllyyy[1:37pm] ajturner: how do those fit w/ the current geo mf?

[1:39pm] joshli: it is a little unclear

[1:39pm] joshli: if you look at http://microformats.org/wiki/geo

[1:40pm] joshli: there is an example which renders coordinates: e.g. <span class="latitude">37.386013</span>

[1:41pm] joshli: and another one which doesn't have to:n <abbr class="latitude" title="37.408183">N 37° 24.491</abbr>

[1:41pm] ajturner: right - either are acceptable

[1:42pm] joshli: Putting the coordinates into "title" is not pretty, however.

[1:42pm] joshli: Allan and I looked at <object>

[1:42pm] joshli: I have somewhere an example. It lets you put a data string in, a class, and also reference a component such as a plugin.

[1:43pm] joshli: 3. Use <meta content="" /> and <link /> , but in the body not just the head of the page

[1:44pm] ajturner: but I don't believe Object tags are rendered?

[1:44pm] joshli: This is the RDF-A proposal for annotating XHTML. It isn't currently valid XHTML, however, most browsers simply ignore it.

[1:44pm] ajturner: MF has the focus to keep it simple and as historically has been used and is simple to use in editors/current devs, etc.

[1:45pm] joshli: Re: Object - yes, that's the idea. A regular browser without some form of intervention has no useful rendering for geotags.

[1:45pm] joshli: But it makes the coordinates available for javascript or some other scripting language or plugin to make use of.

[1:46pm] ajturner: JS can access the current MF's very well too

[1:46pm] ajturner: in fact, it's rather easy now - can just getElementsByClassName("geo")

[1:47pm] joshli: Yes, and my only objection is the use of "title". It's a useful hack, but still a hack in terms of widespread usage.

[1:47pm] ajturner: true

[1:48pm] joshli: That said, I don't think these are mutually exclusive, it would just be good to have a strict usage in each case.

[1:53pm] ajturner: one thing to defintily keep in mind, which the current MF's don't do - is associate the geo or adr w/ the context

[1:53pm] ajturner: so you have those coords, but don't know specifically what they're referencing

[1:53pm] ajturner: the only way is to use an hCard or hEvent

[1:57pm] joshli: That is a good point, in that we can use several tags in RSS to be more specific, but the XHTML elements only allow one tag each.

[1:58pm] ajturner: right - RSS is at least forceably enclosed contextually - one item has a date, author, description, tag, geo reference

[1:58pm] ajturner: although - a big shortcoming of GeoRSS Simple is the limitation of a "single point"

[1:58pm] ajturner: and no possibility of a collection of points

[1:59pm] joshli: There is always the GeoRSS route for annotating anchors in a Web page. Then you can reference a context document to place the reference on, etc.

[2:02pm] ajturner: well, referencing another element would be sufficient

[2:02pm] ajturner: the problem is w/ MF, as with many standards bodies, there is a sol'n -b ut that gets dragged out forever in commmittee, pragmatic waiting, and adoption

[2:21pm] stakagi joined the chat room.

[2:26pm] joshli: Here is an <object> example:

<span>Tagged content<object class="georss:where" classid="georss:point" data="45.256 -110.45">

<param name="georss:crs" value="epsg:4269" /> </object></span>

The param element is a little wordy, but precise.

[2:27pm] ajturner: but - that object isn't rendered by the browser, right?

[2:28pm] joshli: No, in fact you can refer to what code is supposed to deal with the object.

[2:28pm] ajturner: well, but that's an extra layer of complexity - by default you do that work for no benefit

[2:28pm] ajturner: it requires that someon install some plugin to use that data

[2:28pm] joshli: It could just be javascript.

[2:30pm] ajturner: but what if I just want to display the coords?

[2:30pm] joshli: The concept is to geotag or "featurize" some part of a Web page, so that it is ignored unless appropriate, so that it has a standard meaning when appropriate, and so that it doesn't conflict with other usage in a page.

[2:30pm] ajturner: why can't you do all that just in a div or span insstead of an object? you don't like the coords in the title?

[2:31pm] ajturner: b/c I agree the MF could include more useful things like projection

[2:31pm] ajturner: name, tags, featuretype, etc.

[2:31pm] joshli: Displaying the coordinates is ok, but I don't think that would normally be appropriate.

[2:32pm] ajturner: hrm, depends - many people now display coords, but I agree that perhaps not really useful to most people

[2:32pm] ajturner: address/name is most useful

[2:33pm] joshli: Ultimately we'd like to have geotags and refs which the browser does something specifically useful with, but in the meantime they are just annotations which help find pages or help additional code to do something useful.

[2:34pm] ajturner: well, but "in the meantime" they could also be displayable - and there are lots of utils that already do something with the geo mf[2:35pm] ajturner: display maps, GPX, aggregate, etc. using FF extensions

[2:35pm] joshli: sure, that doesn't require display however.

[2:53pm] stakagi: I think it desirable that Geo Vocabulary Update is performed by the addition of a qualifier vocabulary. It is a modularization like Dublin Core.

[2:53pm] joshli: I am not sure what you mean.

[2:57pm] stakagi: it means this spec (http://www.w3.org/2003/01/geo/) itself is not updated but making a new specification. Based on Geo Vocabulary Update.

[2:58pm] stakagi: As qualifier

[2:58pm] joshli: Yes, I mean to update the namespace with a new specification, not really update the old vocabulary itself

[2:59pm] stakagi: yes

[3:02pm] stakagi: It is still under construction. > http://www.w3.org/2005/Incubator/geo/Wiki/LatitudeLongitudeAltitude

[3:04pm] joshli: good work, now we hope for an OWL/RDF contribution to make use of it.

[3:05pm] stakagi: ISO6709 tools for Java is http://blog.svg-map.com/2006/12/iso6709tool_for.html

Summary of Action Items

[NEW] ACTION: Chris Goad and John Goodwin to work on defining GeoRSS as Geo RDF/XML: revision of model and development of definition.

[End of minutes]

Minutes formatted by hand