- From: Krzysztof Janowicz <janowicz@ucsb.edu>
- Date: Tue, 26 May 2015 12:12:14 -0700
- To: <Kerry.Taylor@csiro.au>, <L.Svensson@dnb.de>, <eparsons@google.com>
- CC: <public-sdw-wg@w3.org>
> Krzysztof, why is Java such a hot bed of linked data?!? Good question. If you look at the detailed map here http://stko.geog.ucsb.edu/pictures/lstd_map.png you will see massive data errors. Based on our 14+ million sample, it looks like about 10% of all linked geographic data has some issues related to it. In many cases those issues were systematic and we notified the data providers, e.g., DBpedia. For instance, the dense block-like areas in China are in fact places from the US east coast. I will look into the data from Java and will let you know if I find something interesting. Anyway, the systematic list of errors may turn out to be useful for our sdw group as it illustrates what typically goes wrong. Best, Krzysztof On 05/18/2015 07:39 AM, Kerry.Taylor@csiro.au wrote: > WGS84 is certainly widely used for linked data in practice, probably heavily influenced by this http://www.w3.org/2003/01/geo/, commonly called "geo". > > Oddly, perhaps, schema.org seems not to care about CRS at all: http://schema.org/GeoCoordinates > > Can we take inspiration from the former one (geo) and admit alternative CRSs that must be identified by virtue of the ontology (and therefore namespace, assuming a 1-1 relationship) that is used? We could perhaps develop a couple ourselves (perhaps a WGS84-like one, and another for a relative 3D system), and then allow any other to be used by virtue of reference to the intended vocabulary (as our best practice advice)? > > Maybe this is a cop-out but it is a way of dealing with the common cases blindly, yet requiring a CRS to be implicitly identified, and also enabling the use of more complex or niche CRS whenever needed. We won't stop people making mistakes, whatever we do. > > This could do for *referencing* a CRS without ever needing a "default". For the *description" of a CRS, I would vote to defer that to the OGC by its existing methods, and I see no reason why that description needs to have a linked data representation, beyond an ontology that permits its use. > > > > > Krzysztof, why is Java such a hot bed of linked data?!? > > Kerry > > >> -----Original Message----- >> From: Svensson, Lars [mailto:L.Svensson@dnb.de] >> Sent: Monday, 18 May 2015 9:44 PM >> To: Ed Parsons; janowicz@ucsb.edu >> Cc: SDW WG Public List >> Subject: RE: UCR issue: phrasing of CRS requirement(s) >> >> On Monday, May 18, 2015 12:24 PM, Ed Parsons wrote: >> >>> In most cases I don't think they actually do mean WGS84 as in the >>> ellipsoid and datum. >>> >>> I would guess it is usually shorthand for the the full spatial >>> reference system defined by EPSG4326 or more likely on the web >>> EPSG:3857 >> My fear is that in some cases the data providers don't really know what >> their coordinates mean in terms of ellipsoid, datum and reference >> system. They have some coordinates taken from geonames, Wikipedia or >> some other source and haven’t really thought of that (geographic) >> coordinates are not just coordinates but that there is a context to >> that, too. To what extent we can assume that they mean CRS84, I don't >> know. >> >> So I think I'm on the same page as Linda on this. >> >> Best, >> >> Lars >> >>> On 16 May 2015 at 04:02, Krzysztof Janowicz <janowicz@ucsb.edu> >> wrote: >>> right, so how can they be sure they mean WGS84? >>> >>> Here is a funny example how this can go wrong and went wrong in the >> past: >>> http://stko.geog.ucsb.edu/location_linked_data (See the Copernicus >>> crater) >>> >>> Best, >>> Krzysztof >>> >>> >>> >>> >>> On 05/15/2015 04:27 AM, Peter Baumann wrote: >>> right, so how can they be sure they mean WGS84? if I copy-past >>> coordinates from web info about Germany then in the past this used to >>> be Gauss-Krüger, and several strips = sub-systems. Now let's talk >>> about height and SI vs imperial units etc - what default could we >> agree on? >>> With a default, all coordinate info out there on the Web (flat, >>> height/depth, time, pressure, ...) will often be interpreted wrongly. >>> IMHO we should rather encourage, for m2m communication, that we >>> achieve informational completeness. >>> >>> my 2 cents, >>> Peter >>> >>> On 05/15/15 13:21, Linda van den Brink wrote: >>> Hi all, >>> >>> OK, that could be the consensus within OGC, but the GeoJSON spec does >>> describe a default CRS and I can understand this very well. Non- >> experts, i.e. >>> people from outside the geospatial domain who are using or want to >> use >>> geospatial data, often have no idea that there even *are* multiple >>> coordinate reference systems. >>> >>> Linda >>> >>> Van: Peter Baumann [mailto:p.baumann@jacobs-university.de] >>> Verzonden: vrijdag 15 mei 2015 13:01 >>> Aan: Linda van den Brink; Frans Knibbe; SDW WG (public-sdw-wg@w3.org) >>> Onderwerp: Re: UCR issue: phrasing of CRS requirement(s) >>> >>> Hi all, >>> >>> FYI, there has been a vivid discussion in OGC on default CRSs on the >>> occasion of JSON coming up with such an idea, and OGC very much and >>> strongly agreed that this is not a good idea. >>> >>> In general, a coordinate tuple should have exactly one CRS referenced >>> which may include >>> - spatial horizontal (such as Lat/Long) >>> - time (possibly using different calendars) >>> - elevation >>> - anything else (eg, atmospheric sciences like to use pressure as a >>> proxy for >>> height) >>> - finally, planetary CRSs are more and more coming into play as well. >>> I sense that this is very much in alignment with the ideas that we >> are >>> discussing here. >>> >>> OTOH, it is indeed important to have one common mechanism of >>> describing CRSs. As mentioned earlier, OGC has such mechanisms in >>> place through CRS WKT plus the CRS Name Type Specification (maybe >>> quite misleading in its title, it allows to describe CRSs by >> composing >>> them from other ones, such as flatland >>> + time, flatland + pressure, flatland + depth, flatland + geological >> time). >>> So definitely supporting Linda's observation on referencing vs >> describing. >>> -Peter >>> >>> On 05/15/15 09:40, Linda van den Brink wrote: >>> Hi Frans, >>> >>> I noticed that a requirement related to this is in the spreadsheet >> but >>> not (yet?) in the UCR document. It is this requirement: >>> >>> “There should be a default CRS that is assumed when nog CRS is >> specified” >>> (s/nog/no) >>> >>> WGS84/lat lng is the de facto standard CRS for spatial data on the >>> web. Both publishing and using spatial data on the web should be easy >>> for non-experts, so this requirement of having a default CRS makes a >>> lot of sense to me. The most common cases become more easy that way. >> I think this should be added to par. >>> 5.6 of the UCR. >>> >>> In this light (i.e. usability for non-expert users), the best >> practice >>> should have information about how data owners should describe, how >>> users can recognize and what tools they can use to transform non- >> WGS84 >>> coordinate systems to the coordinate system they need. >>> >>> A second point I’d like to make is that CRS should be suitable also >>> for non- geographical reference systems (for non-Earth oriented >>> applications).I think this is covered by 5.14, but the text of that >>> paragraph is not completely clear to me. )“Standards for spatial data >>> on the web should be independent on the reference systems that are >>> used for data.”) >>> >>> Finally, to answer the question in the issue, as I read it, req A is >>> not replaceable by req B. Req A is about *referencing* a CRS, while >>> req B is about *describing* a CRS – i.e. the description you get >> about >>> the CRS when you dereference a CRS reference. >>> >>> Linda >>> >>> Van: Frans Knibbe [mailto:frans.knibbe@geodan.nl] >>> Verzonden: woensdag 13 mei 2015 14:20 >>> Aan: SDW WG Public List >>> Onderwerp: UCR issue: phrasing of CRS requirement(s) >>> >>> Hello all, >>> >>> I have raised an issue for the UCR document: ISSUE-10. >>> All help in getting this issue resolved is very welcome. >>> >>> Regards, >>> Frans >>> >>> >>> -- >>> Frans Knibbe >>> Geodan >>> President Kennedylaan 1 >>> 1079 MB Amsterdam (NL) >>> >>> T +31 (0)20 - 5711 347 >>> E frans.knibbe@geodan.nl >>> www.geodan.nl >>> disclaimer >>> >>> >>> -- >>> Dr. Peter Baumann >>> - Professor of Computer Science, Jacobs University Bremen >>> www.faculty.jacobs-university.de/pbaumann >>> mail: p.baumann@jacobs-university.de >>> tel: +49-421-200-3178, fax: +49-421-200-493178 >>> - Executive Director, rasdaman GmbH Bremen (HRB 26793) >>> www.rasdaman.com, mail: baumann@rasdaman.com >>> tel: 0800-rasdaman, fax: 0800-rasdafax, mobile: +49-173-5837882 >> "Si >>> forte in alienas manus oberraverit hec peregrina epistola incertis >>> ventis dimissa, sed Deo commendata, precamur ut ei reddatur cui soli >>> destinata, nec preripiat quisquam non sibi parata." (mail disclaimer, >>> AD 1083) >>> >>> >>> >>> >>> -- >>> Dr. Peter Baumann >>> - Professor of Computer Science, Jacobs University Bremen >>> www.faculty.jacobs-university.de/pbaumann >>> mail: p.baumann@jacobs-university.de >>> tel: +49-421-200-3178, fax: +49-421-200-493178 >>> - Executive Director, rasdaman GmbH Bremen (HRB 26793) >>> www.rasdaman.com, mail: baumann@rasdaman.com >>> tel: 0800-rasdaman, fax: 0800-rasdafax, mobile: +49-173-5837882 >> "Si >>> forte in alienas manus oberraverit hec peregrina epistola incertis >>> ventis dimissa, sed Deo commendata, precamur ut ei reddatur cui soli >>> destinata, nec preripiat quisquam non sibi parata." (mail disclaimer, >>> AD 1083) >>> >>> >>> >>> -- >>> Krzysztof Janowicz >>> >>> Geography Department, University of California, Santa Barbara >>> 4830 Ellison Hall, Santa Barbara, CA 93106-4060 >>> >>> Email: jano@geog.ucsb.edu >>> Webpage: http://geog.ucsb.edu/~jano/ >>> Semantic Web Journal: http://www.semantic-web-journal.net -- Krzysztof Janowicz Geography Department, University of California, Santa Barbara 4830 Ellison Hall, Santa Barbara, CA 93106-4060 Email: jano@geog.ucsb.edu Webpage: http://geog.ucsb.edu/~jano/ Semantic Web Journal: http://www.semantic-web-journal.net
Received on Tuesday, 26 May 2015 19:12:53 UTC