INSPIRE view services – why it sucks

21.03.2013 – 9:00 am

Yah, several years, we know about INSPIRE principals. Where ever I was going, I was telling people, how great the life is gonna be: everybody will share the data on the same basis – holy Web Mapping Service 1.3.0 will work literally everywhere. There is common and agreed set of projections, which will be again supported by every public authority and so, you will be able to display seamlessly data from different sources across Europe! I was really looking forward to the bright future of web mapping!

Now, according to roadmap view services should be set and operable. Everything should work. Technical guide (which I’m not able to download since last year), says something about WMS 1.3.0 (with some minor modifications). The WMS should support pretty large set of coordinate reference systems (BUT, they are not defined with their EPSG code, you have to find them…). So where is the problem?

First problem is, that the data are not harmonized yet. You can not get comprehensive topographical map between two states. Good news is, people are working on it (works for One Geology Europe for example). I can understand that.

From the perspective of web mapping application developer, I need two data sources for background map, which are expected to be in every map application:

  1. Topo map
  2. Aerial map

Both are available in Google or Bing maps. They are cool, looking great, no problem. But for some reasons they can not be used in every mapping application. Or I simple do not want to relay on those proprietary data providers (nothing against them in general). For the topo map, everybody knows OpenStreetMap nowadays. Problem solved (at least for Europe, the data are good enough)! Their tail service is cool – works like a charm, fast and furious.

But for the aerial photography, there is no such user-oriented movement yet. Nobody takes his/her private plane and goes to make aerial photography just for fun. (Note that OpenAerial is down for some years, but it was never really usable.) But we have INSPIRE, everything should be so easy!

Today, I wanted to create such seamless web map with ortophoto background layer and topographical map. As said, I took OpenStreetMap for basis, that implies EPSG:3857 coordinate reference system. The tiles are working fast, the tiling system is well known and described. OpenLayers is used to them – no problem.

Than I wanted to cover it with Ortophoto and BANG – there is no easy way, how to do it. I went to several German providers, as well as to the national geoportal of Poland. I had enough: IF there were WMS services available, there did not support the worlds most used coordinate reference system used for web mapping. The system is not used in INSPIRE, so why to support it? In most software packages “to support” some system is equivalent to “and single note to configuration file”. Result is, you can not combine it with most available web mapping data sources which are around (again: OSM, Google, Bing, …), without heavy server or client reprojections.

Another point: WMS is great, but WMTS is event better. Basically I’m all for cashing of raster data, it makes sense for web mapping. WMS is usable for special layer types or special devices. But not for huge background layers. I’m asking, why WMTS or similar caching scheme is not part of INSPIRE viewservices specification? And why the public administrations do not support it by themselves, just so, because it makes sense (to them as providers as well as to us as users)? User (not me, but user of my web map) is not gonna wait five minutes, till the image is generated, tiles have to jump out fast.

There is ONE single exception I’m aware of: CZECH OFFICE FOR SURVEYING, MAPPING AND CADASTRE (COSMC), they are really publishing the WMS services for all data also with EPSG:900913 (3857). Just have a look – cool, you can combine it with other data sources.

What is even more cool: they are offering WMTS as well and it’s really working fast, as desired. I know, it no big deal from technical point of view to configure service like that – but I’m glad, they did in at COSMC.

So, this is, what INSPIRE should bring to me and I’m missing: data in usable services and projections. I do not care, if some processionals do not like the projection from some, maybe even rational reason (from their point of view). Huge step was already made towards paneuropean interoperability, data openness and their real-life usability. But this is not the end. Data distribution must be done in a way, it’s usable for the enduser: fast, and so we can combine it with other popular datasources. I say: WMTS + EPSG:3857 would solve the problem. WMS is important – but not suitable for background data.

Share Button
  1. 12 Responses to “INSPIRE view services – why it sucks”

  2. I fully agree. WMTS and pre-rendered or cached tiles is the way to go for background maps.

    BTW for CUZK there is now also TMS: http://tms.cuzk.cz/ made with MapTiler.

    We have published an open-source project which allows distribution of OGC WMTS compliant (including compatibility with ArcGIS Desktop clients and QGIS) services from an ordinary LAMP (=PHP) server – which is very fast and great for background maps rendered with http://www.maptiler.com/ or alternatives.

    Check: https://github.com/klokantech/tileserver-php/.
    To be finished including UI a la: http://klokantech.github.com/tileserver-php/tileserver/ by summer and announced on ICC2013 in Dresden.

    By Klokan Petr Pridal on Mar 21, 2013

  3. I can not fully agree.
    First point is, that WMTS IS part of INSPIRE TG on view services for years… And there is a TileMatrixSet defined on the base of GoogleCRS84Quad named INSPIRECRS84Quad.
    You lined out correctly, that there is somehow a lack of usability regarding the different INSPIRE specifications. But I think the win INSPIRE makes possible is even greater. It is interoperability and availability! Never before there was a that widespread initiative which declares that OGC standards have to be used for providing geodata. It will just be a matter of time untill there is a huge coverage for most geospatial data provided by administrations all over europe.

    I can understand your point regarding EPSG:3857. But you re wrong in saying, INSPIRE does not specify which SRS to be used. Implicitly specifications say, EPSG:4258 has to be used. And this ETRS89 based SRS does not need big reprojections to others. UTM, the most used projection type is also based on ETRS89 and reprojections are lightweight.

    By Sebastian Goerke on Mar 21, 2013

  4. @Klokan Petr Pridal:
    You are saying you are providing an OGC compliant WMTS? I cannot believe that, as WMTS is not ready for compliance certification and there are no existing OGC certificates on WMTS yet…
    So I think, you should better formulate “we are implementing OGC WMTS standard claiming compliance” ;-)
    I will have a look at you project, as I am very interested in WMTS implementations (I am an active contributor of the deegree project)

    By Sebastian Goerke on Mar 21, 2013

  5. Thanks/Danke Sebastian, for your comments!

    By Jachym Cepicky on Mar 22, 2013

  6. For Flanders (northern part of belgium), you can get orthophotos tiled in EPSG:3857 http://www.agiv.be/gis/diensten/?artid=1671
    Check The tms and wmts.

    By Johan VdW on Mar 22, 2013

  7. I was not able to keep my hands still with such a provocative title ;-)
    And it is very interesting to get some input from a more or less users / web mapping view because INSPIRE is mostly discussed between specialiced geo-experts from the data providers. I see myself somehow inbetween both “worlds” and of course, your main issues stay as points of critiscism, especially reagarding usability!

    By Sebastian Goerke on Mar 22, 2013

  8. Oh yeah, the title was cheep and – as you practically tested on yourself – it worked ;-)

    But the blogpost is just my $0.02 to the INSPIRE discussion. Or, as Arnulf pointed out, to the INSPIRE dialogue.

    By Jachym Cepicky on Mar 22, 2013

  9. Ok, that sucks, but that’s not fault of INSPIRE. That’s fault providers . INSPIRE is very benevolent in many ways and in many DS there’s only written: “you should provide it, when you can”. It’s logical. How could I provide 3D buildings, when I don’t have the data?
    I think that a lot of providers have problems (not enough time or money) with basic implementation of INSPIRE GML data and they’re not even thinking about web services yet.

    By Michal Med on Mar 22, 2013

  10. Well, it’s always people’s fault, if something does not work. INSPIRE is as good, as people have created it. Concrete example is GML used in INSPIRE, which was not implemented anywhere in the time, it was published. Why was it made so complicated? KISS!

    WMTS is just another distribution of the same data, compared to WMS – but in some cases more effective. As I said: it takes not more than a day to set it up (for the student in his first year at university). And it would be great in general, to support real-life use case.

    This is really to be read as real-life example: I wanted to display ortophotos from some (all?) European countries on top of OpenStreetMap – is it really *that* extreme requirement?

    By Jachym Cepicky on Mar 22, 2013

  11. Yes, I know that you would like it in kml format and honestly, that’s not so bad idea. Unfortunatelly, it’s too late to complain about it. Maybe there could be added into INSPIRE specifications some possibility of providing light-weighted version of data, but honestly, that’s not so likely. In JRC are they quite busy at the moment (in my opinion).

    You’re right about WMTS. As I said: it’s fault of providers. And maybe you’re little bit overestimating first year’s students :)

    Big problem of INSPIRE nowadays is cooperation over borders. Accurately that there’s nothing like that. On the other side: INSPIRE should contain really HUGE amount of data. Some of the data isn’t available yet, some data are stored by thousands of providers in one country and INSPIRE is still in the beginning. Problems must be solved in sequence.

    By Michal Med on Mar 22, 2013

  12. I personally do not push KML as ideal format. Do not push KML at all. But it really is so simple, that it should be considered for INSPIRE purposes.

    Too late to complain? C’mon – it’s never too late. Better late, than later.

    I agree, that INSPIRE is at the beginning. It’s a huge project and certainly it will be better and it will change some components by the time. It should remain flexible for such proposals, otherwise, it will never be usable – accepted by large community and will remain “just another EU evil”.

    By Jachym Cepicky on Mar 22, 2013

  13. Jachym, regarding your described orthophoto use case, don’t you think it is a bit early for actually doing this with INSPIRE? Orthophotos are part of Annex II which has according to the original road map still not to be provided through those schemas.
    So your issue is not that much an INSPIRE one ;-)
    It is more about how providers configure their services which they also want to be used as their INSPIRE services.
    I totally agree, that interoperability is still the big issue and this has also meeting points with usability.

    Regarding the GML issue I would say, it has to be that complicated as INSPIREs original aim was to harmonize all geodata touching environmental themes. And harmonizing always means bringing different models together and create the least common factor of them. So GML is the best choice here.

    By Sebastian Goerke on Mar 22, 2013

Post a Comment