RE: ACTION-98: Look at a list/matrix of the common formats (geojson, gml, rdf, json-ld) and what you can or can't achieve with it

Clemens et al.,

I am happy that generic container formats, like CSV or NetCDF, are excluded, unless there is a significant community developing with a geospatial version, as in geoJSON.

I think we should include SVG as a special case, as it does have a lot of non-spatial geometry that is often coerced to be geospatial. I have encountered at least two recent W3C efforts to standardise usage of WGS84 etc in SVG, however inappropriate some people think that is. 

<rant>I also would like to include it to stop geospatial people re-inventing 1980s computer graphics from scratch.</rant>

Chris

-----Original Message-----
From: Clemens Portele [mailto:portele@interactive-instruments.de] 
Sent: Wednesday, November 25, 2015 6:36 PM
To: SDW WG Public List
Subject: Re: ACTION-98: Look at a list/matrix of the common formats (geojson, gml, rdf, json-ld) and what you can or can't achieve with it

Some comments on candidates that have been mentioned in recent emails:

SVG: I see the rationale, although I see a risk that we make this too broad. There is also a difference between formats that describe spatial things / features / etc and others that are more geometries with an option to attach properties to geometries. In a way we could then add a whole range of CAD formats, too. But maybe SVG is appropriate here as it is a standard for the web?

WKT/WKB: I intentionally did not include it, as it is a spec for encoding geometry that will be used by other formats to encode the geometry part (like GeoSPARQL, GeoPackage, SF SQL). WKT/WKB is not enough in itself to encode spatial data. Without having a closer look, GeoHash probably is in the same category. Maybe a seaprate comparison is need how to represent geometry (this overlaps with ACTION-101, I think)?

Should we restrict ourselves to "open, non-proprietary, somewhat current formats": Whatever non-prioprietary and open means… For me the format should play a significant role in one or more target communities today. In addition, it have the documentation published for free on the web and with a license that allows and enables others to implement the format. Shapefiles, Mapbox Vector Tiles, GeoJSON all fall into that category, but depending on the definition of open or non-proprietary they may not fit.

Geodatabase: This would be another candidate, but I did not include this as I had Shapefiles already on the list (as I think there are more Shapefiles in use compared to geodatabases) and I did want to make the list too long. 

SpatialLite: Maybe someone can judge whether it adds sufficient value over GeoPackage so that it should be another column (I cannot).

Talk to you all soon,
Clemens



> On 25 Nov 2015, at 17:59, Andrea Perego <andrea.perego@jrc.ec.europa.eu> wrote:
> 
> +1 to including SVG.
> 
> Moreover, I think we are probably missing WKT / WKB.
> 
> 
> I take this opportunity to mention that the LOCADD CG started some 
> time ago a state-of-the-art document concerning (quoting) 
> "vocabularies/encodings for providing a geospatial reference for 
> resources, e.g. through coordinate geometries, addresses or 
> geographical names".
> 
> It was never finished, but it can stll provide some input. See, in particular:
> 
> - Existing standards
>  http://www.w3.org/community/locadd/wiki/SoA_Survey#Existing_standards
> 
> - Comparison of syntax encoding schemes  
> http://www.w3.org/community/locadd/wiki/SoA_Survey#Comparison_of_synta

> x_encoding_schemes
> 
> - Comparison of Vocabularies
>  
> http://www.w3.org/community/locadd/wiki/SoA_Survey#Comparison_of_Vocab

> ularies
> 
> 
> Note that in those sections we included also GeoHash, the geo: URI 
> scheme and Ian Davis's "WGS 84 Geographic Point URI Space":
> 
> http://vocab.org/placetime/2003/05/geopoint-wgs84-20030516

> 
> Can these be relevant here? They are examples of URI-based encodings 
> of geometries, and two of them can be related to the "linkage"
> requirement. IMO, we should address them in the BPs, at least to say 
> whether to use them or not, and when.
> 
> 
> Finally, one of the reason why the LOCADD draft was never completed, 
> is because we realised that the job we planned to do was already 
> carried out quite extensively by the GeoKnow project in the following
> deliverable:
> 
> http://svn.aksw.org/projects/GeoKnow/Public/D2.1.1_Market_and_Research

> _Overview.pdf
> 
> 
> Andrea
> 
> 
> On Wed, Nov 25, 2015 at 4:43 PM, Frans Knibbe <frans.knibbe@geodan.nl> wrote:
>> 
>> 
>> 2015-11-25 15:49 GMT+01:00 Ed Parsons <eparsons@google.com>:
>>> 
>>> I agree Spatialite is a interesting case, a database that is also a file..
>>> esri's file geodatabase would be similar..
>> 
>> 
>> Yes, geodatabase could be anotother candidate. Or do we only want to 
>> describe open, non-proprietary, somewhat current formats? At least 
>> that would help to keep the matrix manageable. Otherwise we would 
>> need to add more formats (take a look at the list of OGR formats for 
>> example)
>> 
>> And how about adding SVG? It was discussed before, so it makes sense 
>> to add it.
>> 
>> And if we add a row 'supports topology', then TopoJSON is an 
>> interesting candidate too. Well, I think the topology support 
>> criterion is good to have anyway.
>> 
>> I did some quick websurfing to find out if GeoPackage and SpatiaLite 
>> could be lumped together. I'm not sure, but it could be they differ 
>> in some respects. For example, it seems that SpatiaLite allows 
>> multiple geometry columns in a table while GeoPackage does not (I 
>> have not tested this). That would mean they would differ in the row 
>> 'supports multiple geometries per feature'.
>> 
>>> 
>>> This is great work Clemens alhough I share your thoughts around 
>>> JSON-LD we are premature calling a spatial format at the moment.
>> 
>> Yes, JSON-LD is an odd one in the matrix. It is not intended as a way 
>> to encode spatial data, it is someting on another level (an RDF 
>> format, or a document annotation format).
>> 
>> Regards,
>> Frans
>> 
>>> Ed
>>> 
>>> 
>>> On Wed, 25 Nov 2015 14:06 Frans Knibbe <frans.knibbe@geodan.nl> wrote:
>>>> 
>>>> 2015-11-25 14:30 GMT+01:00 Clemens Portele
>>>> <portele@interactive-instruments.de>:
>>>>> 
>>>>> Hi Frans,
>>>>> 
>>>>> in general I have only added formats that I have some familiarity with.
>>>>> 
>>>>> In the SpatialLite case, isn’t it more an implementation than a format?
>>>>> The linked wikipedia page states: "SpatiaLite supports several 
>>>>> open standards from the OGC and has been listed as a reference 
>>>>> implementation for the proposed GeoPackage standard."
>>>> 
>>>> 
>>>> It can be considered a vector geometry format, like it says on the 
>>>> wikipedia page: 'Being a single binary file, SpatiaLite is also 
>>>> being used as a GIS vector format to exchange geospatial data'. It 
>>>> is possible to see a SpatiaLite file as a smart Shapefile.
>>>> 
>>>> On the other hand, one could say that OGC Simple Features is the 
>>>> actual format used by SpatiaLite. It depends on what we understand 
>>>> 'format' to mean exactly.
>>>> 
>>>> Greetings,
>>>> Frans
>>>> 
>>>> 
>>>> 
>>>>> 
>>>>> GeoPackage is in the list.
>>>>> 
>>>>> Best regards,
>>>>> Clemens
>>>>> 
>>>>> 
>>>>> 
>>>>> On 25 Nov 2015, at 14:25, Frans Knibbe <frans.knibbe@geodan.nl> wrote:
>>>>> 
>>>>> Hello Clemens,
>>>>> 
>>>>> Have you considered adding SpatiaLite to the collection of formats?
>>>>> 
>>>>> Greetings,
>>>>> Frans
>>>>> 
>>>>> 2015-11-25 14:16 GMT+01:00 Andrea Perego
>>>>> <andrea.perego@jrc.ec.europa.eu>:
>>>>>> 
>>>>>> Hi, Clemens.
>>>>>> 
>>>>>> Just to say that an option would be to create a page in the SDW wiki.
>>>>>> 
>>>>>> Andrea
>>>>>> 
>>>>>> 
>>>>>> On Wed, Nov 25, 2015 at 2:03 PM, Clemens Portele 
>>>>>> <portele@interactive-instruments.de> wrote:
>>>>>>> Looking at
>>>>>>> https://lists.w3.org/Archives/Public/public-sdw-wg/2015Nov/0052.

>>>>>>> html the table structure seems to be lost after the email is 
>>>>>>> processed by the list software, so I will make the table available somewhere and send a link.
>>>>>>> 
>>>>>>> Clemens
>>>>>>> 
>>>>>>> 
>>>>>>>> On 25 Nov 2015, at 13:57, Clemens Portele 
>>>>>>>> <portele@interactive-instruments.de> wrote:
>>>>>>>> 
>>>>>>>> Dear all,
>>>>>>>> 
>>>>>>>> below is a first attempt at such a matrix for vector data only.
>>>>>>>> 
>>>>>>>> Beside a review (I am not sure that everything is correct or
>>>>>>>> adequate) this would need
>>>>>>>> - additional explanations in text,
>>>>>>>> - more work to align the terminology with the rest of the BP to 
>>>>>>>> make it understandable for the different target audiences,
>>>>>>>> - links to the specification for each format.
>>>>>>>> 
>>>>>>>> But before we work on this, I think we should have a discussion 
>>>>>>>> whether
>>>>>>>> - this is what we were looking for in general,
>>>>>>>> - the list of aspects is complete, too much, or missing 
>>>>>>>> important aspects (e.g. time support, closely coupled APIs / 
>>>>>>>> service interfaces, etc),
>>>>>>>> - the list of formats is ok or whether we need to remove / add some.
>>>>>>>> 
>>>>>>>> I hope the table is still readable once it passes the W3C list 
>>>>>>>> software :) … Best regards, Clemens
>>>>>>>> 
>>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> --
>>>>>> Andrea Perego, Ph.D.
>>>>>> Scientific / Technical Project Officer European Commission DG JRC 
>>>>>> Institute for Environment & Sustainability Unit H06 - Digital 
>>>>>> Earth & Reference Data Via E. Fermi, 2749 - TP 262
>>>>>> 21027 Ispra VA, Italy
>>>>>> 
>>>>>> https://ec.europa.eu/jrc/

>>>>>> 
>>>>>> ----
>>>>>> The views expressed are purely those of the writer and may not in 
>>>>>> any circumstances be regarded as stating an official position of 
>>>>>> the European Commission.
>>>>>> 
>>>>> 
>>>>> 
>>> --
>>> 
>>> Ed Parsons
>>> Geospatial Technologist, Google
>>> 
>>> Google Voice +44 (0)20 7881 4501
>>> www.edparsons.com @edparsons
>> 
>> 
> 
> 
> 
> --
> Andrea Perego, Ph.D.
> Scientific / Technical Project Officer European Commission DG JRC 
> Institute for Environment & Sustainability Unit H06 - Digital Earth & 
> Reference Data Via E. Fermi, 2749 - TP 262
> 21027 Ispra VA, Italy
> 
> https://ec.europa.eu/jrc/

> 
> ----
> The views expressed are purely those of the writer and may not in any 
> circumstances be regarded as stating an official position of the 
> European Commission.

Received on Wednesday, 25 November 2015 18:50:32 UTC