search

w3c / sdw

Repository for the Spatial Data on the Web Working Group
https://www.w3.org/2020/sdw/
148 stars 81 forks source link

W3C LBD reach out - 2D and 3D CAD geometry for building projects #1143

Closed mathib closed 4 years ago

mathib commented 5 years ago

Dear members of the Spatial Data Group,

This is a message of the Linked Building Data Community Group. We are defining web ontologies (OWL) and methods to model buildings, including building topology, building products and properties. To make this approach more attractive for industry, we need to pay attention to geometry. However, geometry in a Linked Data context is currently considered out of scope as it is not only used in the construction industry, but also by mechanical engineers (product modelling), scientist, artists, spatial planners, chemists, etc. For this reason, we try to reach out to this group (and potential other related groups), to achieve a more uniform approach towards geometry in a Linked Data context.

Currently, based on a quick scan through the output documents of the SDW group, we mostly see (geo)spatial data used in applications for sensors, efforts regarding a time ontology, CityJSON, etc. Construction industry partners currently use a lot of CAD applications to create drawings and 3D models in different formats (DWG, OBJ, Collada, glTF, etc.). They hardly use WKT and GML and more and more of the data is 3D, so GeoSPARQL on its own cannot be used directly for common building projects trying to adapt Linked Data technologies. Another concern is that the proposed methods to use geometry in a Linked Data context should be straightforward enough (and modular), to increase the adoption by industry.

Is it currently within the scope/interests of this group to work around this kind of geometry? If so, it might be beneficial to connect between our two W3C groups to exchange ideas and use cases. I see some potential overlap with https://github.com/w3c/sdw/issues/1095 and https://github.com/opengeospatial/geosemantics-dwg/issues/9, but maybe SDW/OGC members can give better feedback here?

Best regards,

Mathias Bonduel for the W3C Linked Building Data Community Group

chris-little commented 5 years ago

@mathib Perhaps the approach that seems to be happening in environmental science, in particular weather and climate, may be a useful model. There was a peripheral discussion in the SDW WG involving @rob-metalinkage. In essence, the existing data silos stay in their local, compact, binary, efficient formats, surrounded by their domain-specific tool ecosystems and domain experts, because no-one is going to retro-convert that enormous amount of data, at great expense and disruption into more voluminous data serializations. But, the metadata describing that data, including how it can be queried and reasoned about, should migrate into the semantic linked web world. So I suggest an approach that does not expend effort on yet more geometry format conversion tools, or one format to rule them all, but work in the metadata/catalogue/vocabulary/ontology area to automate invocation of the existing tools correctly. HTH, Chris

mathib commented 5 years ago

@chris-little Interesting! In general, I think along the same lines in my personal research. Some members of our group were/are also actively investigating a more RDF-based description of geometry, e.g. in a context of parametric geometry where the parametrics have to be accessible as Linked Data.

This brings me to the following questions:

chris-little commented 5 years ago

@mathib Firstly, it is my personal opinion, but there is some consensus around it and @rob-metalinkage did develop a QB4ST (geo-spatial datacube) ontology restricting the generic RDF QB data cube. It is a W3C Technical Note. @Fransie proposed a project to develop a spatial ontology, similar to the OWL-Time ontology. See SDW Issue #1095.

FransKnibbe commented 5 years ago

Hi Mathias,

Although I am not in the Spatial Data on the Web Interest Group (SDWIG), I would like to share my view. Chris mentioned a 'one format to rule them all' approach. What I propose is a variation on that theme. In condensed form it could be called 'one model to underpin them all', with 'model' meaning a universal model for spatial data, and 'them' meaning storage formats, exchange formats, functions, software libraries, software, etc.

This idea was partly inspired by trying to use the same spatial data in a GIS, BIM and Computer Graphics (CG) contexts. As you may have also experienced, that is not easy because those domains use different ways of working with spatial data. But although the domains work differently on the surface, if you drill down to underlying principles the foundations are the same. At the core, working with geometry is purely mathematical, so if all the different domains could be based on a single model with a mathematical core, it should be possible to achieve much better interoperability across the many ways in which software can manipulate spatial data. In this line of reasoning, such a model does not really have to be a web ontology, but I think building it as a web ontology makes a lot of sense, because the semantic web by nature is a cross-domain effort and space by nature is a cross-domain phenomenon. Also, making it a web ontology should lower the threshold for people and organisations with different backgrounds to collaborate, thus increasing the likelihood of coming up with a truly robust and universal model. And yes, I think also this will increase the likelihood of the model being as simple as possible. Furthermore, I imagine such a model to be helpful for use just within a single domain. At least within the GIS domain, working with spatial data is much harder than it needs to be.

I don't think there is an official SDWIG standpoint on this matter, but I am glad it at least seems to be under consideration here and in the OGC Geosemantics DWG. Anyway, it seems to me that intensifying the sharing of ideas between the SDWIG and the LBDCG should be helpful anyway.

Regards, Frans

Op wo 30 okt. 2019 om 12:44 schreef mathib notifications@github.com:

@chris-little https://github.com/chris-little Interesting! In general, I think along the same lines in my personal research. Some members of our group were/are also actively investigating a more RDF-based description of geometry, e.g. in a context of parametric geometry where the parametrics have to be accessible as Linked Data.

This brings me to the following questions:

  • is this a general view endorsed by the SDW group or rather a personal opinion?
  • did you or other members of the SDW/OGC group developed approaches/ontologies/... to implement this view? Or is there in the SDW/OGC group a will to develop something here?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/w3c/sdw/issues/1143?email_source=notifications&email_token=AAGNDTHPQJBJE3NQ4YYTNZTQRFXRDA5CNFSM4JDRZ7W2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOECT3NTI#issuecomment-547862221, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGNDTCQZUNHDN3ZPZ3Z3N3QRFXRDANCNFSM4JDRZ7WQ .

lvdbrink commented 5 years ago

Mathias; it would be very useful to at least discuss the possible synergy between our groups. The SDWIG has a face to face meeting coming up in Toulouse, France next month. We could fit in a discussion on this topic, if it is possible for you and/or others from the LBDCG to join either physically or remotely. Would that be an option?

Linda Co-chair of SDWIG

mathib commented 5 years ago

Hi Linda, I pass this information to the other members of our group. Personally, I would like to join the meeting remotely. I'll let you know if other members of the LBD group will join

lvdbrink commented 5 years ago

Good! I've added an agenda item for this. Keep an eye on the meeting page for remote participation details!

lewismc commented 5 years ago

@chris-little ...but there is some consensus around it and ...

Yes I agree with that.

@mathib over in the ESIP Semantic Technologies community we have been developing and are the stewards of the Semantic Web for Earth and Environment Terminology (SWEET) ontology suite which does contain relevant resources worth mentioning here, specifically (but not limited to)

N.B. both of the above resources (like all SWEET resources) are served as linked data via the http://sweetontology.net/ URI e.g. http://sweetontology.net/reprSpaceGeometry. Content negotiation is also available via usual means.

The reason I mention there is is NOT to say that everyone should use them, but it is to say that these representation constructs in SWEET could be used as the framework from which the LBD effort builds out relevant implementations specific to your use cases e.g.

...RDF-based description of geometry, e.g. in a context of parametric geometry where the parametrics have to be accessible as Linked Data

I was discussing some related issues with @pipauwel and am somewhat aware of the work going on over at LBD group. This is a really interesting effort.

chris-little commented 5 years ago

@lvdbrink Should we leave this issue open to capture the outcome of the face-to-face meeting in Toulouse, or is having it on the Agenda enough? I am happy to close for the the lattter.

lvdbrink commented 5 years ago

Let's keep it open to capture the outcome of the meeting!

chris-little commented 4 years ago

I did not have time to to attend any F2F meeting with the LBD people. Was there one? has enough been done? Can we close this issue?

lvdbrink commented 4 years ago

SDWIG had a discussion with the Linked Building data group in November. See https://www.w3.org/2019/11/19-sdw-minutes.html#item06