Open prushforth opened 8 months ago
PeterParslow
commented
8 months ago Initially I thought that the August meeting was cancelled due to European summer holidays.
But then I see at https://www.w3.org/groups/wg/sdw/ that the charter has already expired but the group is not yet listed as "cancelled".
And at https://www.w3.org/groups/wg/sdw/calendar/, this November meeting appears as cancelled.
Is there a plan to re-charter? How can OGC members contribute to re-chartering at W3C?
situx
commented
8 months ago In the last meeting, as far as I recall, the consensus was to let the SOSA work finish, and then there would be no new items on the original charter. Surely, if there is more to do, one might think about rechartering I guess, but maybe @lvdbrink is the best person to answer that.
lvdbrink
commented
7 months ago The last meeting was 14th September 2023. We announced that we wanted to meet to discuss the expiration of the charter and ways forward. There was little attendance at the meeting, but at least representatives from OGC and W3C staff were there.
In W3C to create / continue a group you have to make a case there's interesting stuff to work on AND enough interest. Options that were mentioned were:
@bert-github said he would talk to W3C staff about an extension for the time being, and @ogcscotts and @rob-metalinkage would talk to W3C about the way forward beyond that. But, as far as I know there haven't been significant developments yet.
That said - while we wait, we can continue the work as if the group was still active. We should get the extension and still be able to publish work through W3C, for the time being.
Meanwhile if people feel there's more work to do, it would help if they make their case - but it would help even more if they can show interest (i.e. show we'd have active group members going forward) and leadership (e.g. someone willing and able to chair would also make a difference).
PeterParslow
commented
7 months ago "more work to do": there are currently 83 open issues. Could we look at those and decide which should be dealt with over the next six months & which can be parked/left? Of course, that requires continued resource (people / time) input to the group.
chris-little
commented
7 months ago @lvdbrink I have not been very involved because the meetings usually clashed with other meetings that I viewed as at a higher (short term) priority at the time. And I was spread too thinly.
ldesousa
commented
7 months ago Dear Linda, dear all,
I formally joined this working group last week. Before reacting to this inquiry I should probably present myself. I have been working on the application of the Semantic Web to the Soil domain since 2018, the result of which is the GloSIS web ontology [0]. From that experience I produced a manuscript introducing the Semantic Web to geo-spatial scientists and practitioners [1]. I am the technical contact of my employer with the OGC and have been involved with a few working groups in that venue.
I am still getting to know my way around, so forgive me if these thoughts do net fully align with the current work in this group. From my perspective there is a general deficiency with CRS expression in GeoSPARQL (and the wider Semantic Web) that needs to be addressed. There is yet too much ambiguity to allow for automated interpretation of CRSs. Linked to that is the emergence of DGGSs as replacement for both literal types and CRSs.
Looking forwards to contribute. Best regards.
[0] https://github.com/glosis-ld/glosis
-- Luís Moreira de Sousa Mastodon: @@.*** URL: https://ldesousa.codeberg.page
Sent with Proton Mail secure email.
On Tuesday, November 14th, 2023 at 2:59 PM, Linda van den Brink @.***> wrote:
The last meeting was 14th September 2023. We announced that we wanted to meet to discuss the expiration of the charter and ways forward. There was little attendance at the meeting, but at least representatives from OGC and W3C staff were there.
In W3C to create / continue a group you have to make a case there's interesting stuff to work on AND enough interest. Options that were mentioned were:
- extend the charter (for a maximum of 6 months, W3C won't authorize a longer extension) and try to get the SOSA work done in 6 months - we'd still have to find a solution for any other work afterwards.
- recharter as a joint group - but we noted that this feels unlikely given the number of people in attendance (not only at that particular session - attendance has been structurally low)
- use the OGC/ISO Joint Advisory Group model for cross org publication + small number of invited experts. This will work from OGC perspective.
@.(https://github.com/bert-github) said he would talk to W3C staff about an extension for the time being, and @.(https://github.com/ogcscotts) and @.***(https://github.com/rob-metalinkage) would talk to W3C about the way forward beyond that. But, as far as I know there haven't been significant developments yet.
That said - while we wait, we can continue the work as if the group was still active. We should get the extension and still be able to publish work through W3C, for the time being.
Meanwhile if people feel there's more work to do, it would help if they make their case - but it would help even more if they can show interest (i.e. show we'd have active group members going forward) and leadership (e.g. someone willing and able to chair would also make a difference).
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>
jvanulde
commented
7 months ago I've changed roles at NRCan so haven't been able to be as active I would like but agree that much work remains.
situx
commented
7 months ago From my perspective there is a general deficiency with CRS expression in GeoSPARQL (and the wider Semantic Web) that needs to be addressed. There is yet too much ambiguity to allow for automated interpretation of CRSs. Linked to that is the emergence of DGGSs as replacement for both literal types and CRSs. Looking forwards to contribute. Best regards. [0] https://github.com/glosis-ld/glosis [1] https://linked-sdi.com/
@ldesousa I have been working in similar directions and would also like to have a CRS ontology/vocabulary/JSON-LD context published. Right now the work seems to have somewhat stalled, because of several reasons, but I would be eager to pick it up again.
We started an OGC Whitepaper as a first start (https://github.com/opengeospatial/ontology-crs) and I tried to create vocabularies from existing databases (https://github.com/situx/proj4rdf/ , browsable in HTML here (https://situx.github.io/proj4rdf/data/def/crs/EPSG/0/9742/index.html)), which could be used to refine the model.
As far as GeoSPARQL is concerned we would like to link to definitions of CRS in RDF, but the creation of the appropriate vocabulary is separate from GeoSPARQL.
If people want, this could also be an activity to pursue within the Spatial Data On the Web Working Group, but I believe it would be crucial to find a group of people committing to this in a foreseeable timeframe.
rob-metalinkage
commented
7 months ago I note several references in the wild to the concept of "W3C maintenance groups", but nothing in the process document (https://www.w3.org/2023/Process-20231103/#WGCharter). something like this would be useful for SOSA but new work needs a new charter.
The other item I think that needs a home is for GeoSPARQL to be a viable target for JSON-LD annotation of GeoJSON (basis of OGC APIs) and FG-JSON extensions. There are several possibilities for this, via direct support for the same granular elements of geometry or via formal alignments to a GeoJSON ontology.
My conclusion is that there is a bigger architectural piece of work the OGC needs to complete to define and test ideas before settling on a scope for specifications, however perhaps a joint WG could be chartered that takes this into account. Given nature of W3C timeboxed specification WGs perhaps some sort of joint roadmap is appropriate instead however, in the context of a persistent Community Group.
oldskeptic
commented
7 months ago As far as GeoSPARQL is concerned we would like to link to definitions of CRS in RDF, but the creation of the appropriate vocabulary is separate from GeoSPARQL. If people want, this could also be an activity to pursue within the Spatial Data On the Web Working Group, but I believe it would be crucial to find a group of people committing to this in a foreseeable timeframe.
There are a few CRS's that would make my life easier in RDF. Sign me up.
PeterParslow
commented
7 months ago "definitions of CRS in RDF"
I would hope that these remain hosted in one of the existing CRS registers, rather than being yet another one.
For those unfamiliar with the 'big three CRS registers', this may help: https://committee.iso.org/files/live/sites/tc211/files/Resources/GuideToCRSRegistries3.pdf - a paper authored across OGC, TC 211, and IOPG (owners of EPSG).
It may make sense for a joint WG to define the RDF model that the register(s) should use (in the same way that OGC defined the GML schema & some combination of OGC & TC211 defined the WKT one, with input from EPSG)
ldesousa
commented
7 months ago @situx proj4rdf is a strong starting point, of which I haven't heard before. Intuitively, the easiest path would be to derive a web ontology from the CRS domain model underpinning WKT (or GML). That path would also make it easy for programmes like PROJ to adopt a RDF syntax. I am on board to work further on this.
But I see a more elementary issue with the bundling of the CRS with the geometry itself in GeoSPARQL. A CRS object property with Feature and FeatureCollection as domains would be more logical and easier to use. Forgive if this was already discussed in the context of version 1.1.
@prushforth A CRS web ontology and a CRS registrar are different topics. Although the latter is certainly important, it is far from being the ultimate goal of a web ontology.
situx
commented
7 months ago @ldesousa You probably mean the geo:inSRS property, which would connect a geo:Geometry to a CRS representation. This was not meant to be defined on a geo:Feature level, as a geo:Feature might be linked to many geo:Geometry in different CRS. However, we dismissed this for GeoSPARQL 1.1, with it being on the roadmap for the next revision. For GeoSPARQL, we do not need and will not define what the CRS representation looks like. We only define a property that can link to a CRS representation described by a URI in RDF.
So, defining a CRS ontology is a different project altogether.
For me, I see the following steps to be done:
Involved in the proj4rdf work, I worked on all of the things as mentioned earlier. What is missing is a critical review of people, endorsement, and, eventually, a place to publish it.
So, if you are willing, I would be happy to move this forward.
Originally posted by @lvdbrink in https://github.com/w3c/sdw/issues/1288#issuecomment-1794973280
This is surprising. There is a lot of work to do for maps on the Web, like: actually standardize them for interoperability, accessibility, internationalization and so on. The Maps for HTML Community Group is a forum that depends on the SDW in some ways, if even just from a community of practice perspective. It would be a shame for the spatial data community to lose interest, and a forum for expressing that interest, in the Web and especially Web standards.
Maybe we should engage in this discussion at a future group plenary call on the foundations of the group's existence, rather than letting the charter expire quietly.
Note that the plenary call from tomorrow (Nov 9th) seems cancelled.