[Qgis-developer] [Qgis-user] A discussion: is qgis still affordable in Europe if it violate the Inspire directive ?

Andrea Peri aperi2007 at gmail.com
Sun Jun 8 04:15:51 PDT 2014


Hi,
The last response of Jef in the ticket explain clearly what it the
strateguy of qgis.The important is tobe compatible with other software GS.
Thie mean the QGIS-server don't metter for real wms compatibility and
interoperability.

I guess is not really clear that the XML response to GetCapabilities is a
protocol tocommunication between system.
It is not only for other gis softwar.
It is for every system can need to ask to the server wms.
The response need to be aboslutely compatible tobe really interoperable.

As example:
this simplest case:
http://www502.regione.toscana.it/geoscopio/servizi/wms/OFC.htm
Tis html page is product autocatically from an our system Not gis that call
the mapserver and with an xslt that parse the XML response produce this
html page periodically.

also thi
http://www502.regione.toscana.it/geoscopio/servizi/wms/INFRASTRUTTURE_E_PRESIDI.htm
and this
http://www502.regione.toscana.it/geoscopio/servizi/wms/CTR.htm
and so on

Also we could use the same mechanism to implement a B2B colloquial betwen a
metadato system and other systems or to the wms server.

This mean to be interoprability. Mean grant to work with every other system
that work with the same specs (WMS spacs in this case),
not only with other GIS software.
The server wms is not only a "maps-dispenser".

This page
http://www502.regione.toscana.it/geoscopio/servizi/wfs/html/comuni.html

is product from another xslt proc that again parse a WFS response from a
TinyOWS system (please notice also the wfs of qgis-server is not complaint
to specs)

This is an automatically parse of a wfs request to a tinyows that return
the list of cadastral parcels and propose this page.
http://www502.regione.toscana.it/geoscopio/servizi/wfs/html/elenco_foglicat.html?com=047001

This is the list of street of a single comunity
http://www502.regione.toscana.it/geoscopio/servizi/wfs/html/elenco_strade.html?com=047001

And clicking on one of them you can jump to the system webgis to see them.

Or you can have the civic number of every street.

http://www502.regione.toscana.it/geoscopio/servizi/wfs/html/elenco_civici.html?com=047001&codtop=RT04700110007TO

or jump to a webgis to see them.
http://www502.regione.toscana.it/geoscopio/fototeca.html?codtpn=-91175&idtpn=21203

All of this mean to be interoperability.

It mean to use many system differents but all understanding the same single
set of informations and capable to exchange to each other the same
information.
Is not only a question betwen three or four software GIS.

When I say arcgis or autoca try to explain that could be more other system
that could be n difficult to operate with a qgis-server that resonse a non
standard XML response at getcapability request.

But if the response of qgis group is
we see the main GIS software work and this is sufficient for us.

This mean the QGIS-Server is not planning to be WMS compliant.

I test the response with XMLSpy and see it is not compliant WMS because it
fail when use the official xsd from OGC.

Instead other product (mapserver and geoserver) will pass the test.

If I report this issue and the response is:
"is all okay for us".

This response is more clearly of every other.

Bye.

Andrea.


Il 08/giu/2014 09:07 "Mats Elfström" <mats.elfstrom at gmail.com> ha scritto:

> Hi!
> I have read this thread and I am puzzled. In my mind, standards and
> interoperability are good, and should be respected.
> Now tell me, is it a deliberate design choice not to make Qgis server OGC
> compliant, and if so, why?
>
> Hälsning / Regards
> Mats.E
>
> Skickat från min / Sent from my iPhone, Ursäkta att jag är kortfattad /
> Excuse my brevity.
>
> 8 jun 2014 kl. 08:51 skrev Andrea Peri <aperi2007 at gmail.com>:
>
>  Sorry Alex,
>
> As you certain know the
> The QGIS server is really far from be a good WMS Server.
>
> It has not the multy style rendering.
> It has not the "named groups" if don't allow to change the html response .
> It don't allow to return a GML response to the user in a GetFeatureInfo if
> don't active the WFS ption.
> This is obviously unacceptable.
> Because there is some layer that a Public Administration cannot put in
> download as they are.
>
> The GeoServer and MapServer are real WMS Server and they allow all of this
> and more other thinks.
>
> I don't speak of velocity where the coparing is impossible.
> But alsoin the rendering capable.
> I guess the Mapserver (the one that I know very well), has monay point of
> advantage on qgisserver.
>
> The ONLY question good of qgis is The project qgis is the the same in the
> server.
> But this mean only that the  qgis-server is good for fast prototyping, no
> more.
>
> When go in production and the users grow the qgis-server will become
> rapidly a bottle of neck.
>
> So the investiment need to enhance the qgisserver to have a good and fast
> , fully complaince and affordable wms server are massive.
> This could be ask to a public admiistration to fund.
>
> But if it is not  a compliant wms server it could not never choose to
> start.
>
> You think really that ths is a killer advantage that could put an
> administration to choose an incopmaptible with the WMS specs piece of
> software instead of a fully compliant with the WMS specs ?
> :))
>
> Is really ore easy to invest on an author for GeoServer of for Mapserver
> rather than waste public fund in a Incompatible piece of software.
>
> Because the laws deny to choose it.
>
> I guess you can also hope to sell your service to someone .
> But if it ask you:
> Hey your software is compliant with OGC WMS ? This is a mandatory need for
> inspire.
>
> You what say to your possible future client ?
>
> Yes (but you say that the real response is NO)
> or say No hoping that it say no problemI fund it to became, using the
> public money and rejecting other more compliance and also opensource
> soltion as mapserver or geoserver.
>
> As an exmaple:
> We fund really many enhancement in qgis.
>
> And refund the same think when pass from 1.8 to 2.0.
> After the our product was break in the pass from 2.0 to2.2 and now go to
> refund again to have them work in the 2.4.
>
> Now the qgis release every 3-4 month.
> These mean that a public administration should plan to refund all thei
> product on qgis every 3-4 months ?
>
> I guess this is an example clear of what mena affordable and reliability.
>
> I guess no one public administration could fund an incompatible piece of
> software to became compatible when there is other solution availables and
> compatibles.
>
> Returning to the incompatible qustion:
>
> Who put the GetPrint tags in the getcapabilities response ?
>
> It was the only to have a direct and immediate advantage. I dont know why
> do this, but can expected that it sell its product to someone.
>
> I say :
> ok you have your advantage, you have take your money, and now ?
> You have destroyed the credibility of qgis as an WMS Server OGC.
> You choose to eat the egg today instead of take the chicken tomorrow, and
> now ?
> Are you happy of this ?
>
> I guess who break repair.
> So I guess the solution should be put and fund by who involontary or
> volontary (i don't know) break the WMS specs compatibility.
>
> Or otherwise the qgis group say:
>
> we don't care about the wms compatibility so choose other software for
> this work.
>
> But I guess there many service seller that go around to sell service on
> qgis and say that it is a server wms compiant with WMS specs.
>
> It is a really bad
> "ok qgis is not usable for an OGC WMS infrastructure."
>
> If instead it is not happy of this it should repair what has break .
>
> I guess an public administration should not fund this reparation,
> otherwise every commite could think to have no proble to do what it like to
> do and after ask to the PA to fund for the repair from its break commit.
>
> Andrea.
>
>
>
> Il 08/06/2014 07:08, Alex Mandel ha scritto:
>
> You are assuming a new deployment and one where QGIS is not used on the
> desktop, where some groups may decide to use it for both in order to cut
> down on IT staffing requirements. You are also assuming the choice is as
> simple as picking a new one, there is plenty of people time involved in
> deploying a different strategy and redo-ing all the styling to fit the
> new way. Sure QGIS can push to geoserver now but running geosever means
> running a JAVA stack. Switching to Mapserver would be more similar with
> an apache setup but the QGIS export to Mapserver has long since fallen
> into disrepair so now you'd need to make separate style rules for 2
> programs.
>
> I completely agree, QGIS should maintain compliance. My argument is that
> organizations that are shifting money from proprietary licenses should
> invest in ensuring an open source product meets their needs. To not do
> so means at any time there be no options that fit their requirements.
>
> I think I've also provided the links that show QGIS can be WMS compliant
> and still have extra options, the spec provides for how to do it, we
> just need to make the modifications and test it.
>
> Thanks,
> Alex
>
> On 06/07/2014 02:40 PM, Maurizio Trevisani wrote:
>
>  Hello,
> you say
>
> "As to why fund it? If QGIS provides other value to your organization in
> some other way, total cost of operation may be lower to simply ensure
> it's compliant rather than to switch software or have to use multiple
> software."
>
> but as long you, Public Administration, can choose among several
> GFLOSS to implement your OGC services (that must be completely
> interoperable with all the potential clients and other PA), why should
> you choose a product that has choosen not to be interoperable?
>
> The problem is not the product, but the people who doesn't mind to
> write interoperable code: they are not reliable - so why a PA should
> fund them and their products?
>
> Qgis is an important product in the GFLOSS reality, but now more than
> ever should take care to complethely adhere to all the international
> standards, especially those which are at the basis for
> interoperability.
>
> Bye.
>
> 2014-06-07 22:17 GMT+02:00, Alex Mandel <tech_dev at wildintellect.com> <tech_dev at wildintellect.com>:
>
>  On 06/07/2014 01:06 PM, Andrea Peri wrote:
>
>  Yes also this is possible,
> but pay attention to use it correctly.
> I guess it is no really simple to use (ie to define the extension).
>
>  It looks really simple to use according to the docs. If it works and
> cascading WMS works with other WMS servers, and it passes the schema
> check I see no issue.
>
>
>
>  In the SLD world this was allowed and a unfortunately and worst
> understanding of it will born a lot of incompatible dialects.
> Also in the metadata world (iso19115) the possibility to extend the specs
> will produce incompatibility monster.
> :)
>
>  This exists in the html world, over time there are winners. If you don't
> care to use the extra features you are always welcome to use the base
> which is 100% compliant. The winners or some compromise variant end up
> in the next version of the spec.
>
>
>  I guess surely better and easy is put the new functions in in a distinct
> and new kind of request.
>
>
>  After reading the WMS doc I believe using the tags I mention is the
> correct way to do it. Technically the result is WMS 1.3.0 compliant.
> Clients are free to ignore the extra functions as not using them does
> not remove any required features.
>
> As to why fund it? If QGIS provides other value to your organization in
> some other way, total cost of operation may be lower to simply ensure
> it's compliant rather than to switch software or have to use multiple
> software.
>
> Thanks,
> Alex
>
>
>
>  Andrea.
>
>
>
> 2014-06-07 21:56 GMT+02:00 Alex Mandel <tech_dev at wildintellect.com> <tech_dev at wildintellect.com>:
>
>
>  I just checked the WMS 1.3.0 specification documenthttp://portal.opengeospatial.org/files/?artifact_id=14416
>
> Extended optional features are allowed. There is a specific way to
> include them. See section 6.9.5 "Extended capabilities and operations"
> <_ExtendedCapabilities> or <_ExtendedOperations>
>
> So perhaps we just need to wrap those extra options in a specific tag
> for them to pass schema testing.
>
> Thanks,
> Alex
>
> On 06/07/2014 12:35 PM, Alex Mandel wrote:
>
>  I understand the issue now. In order to be WMS 1.3 complaint you can
> only use what's in the spec.
>
> Looking at an analogy with html specs I find this limitation appalling
> short-sighted. It means there can be no innovation testing new features
> with the spec unless you manage to get it into the future spec. I find
> it hard to comprehend that clients don't just skip tags that fail to
> match a known tag. In html land its very common for some browsers to
> know some non-standard tags, which are new features in testing to be
> proposed or reworked into future standards. IE's policy of only
> adhering
> to the spec and including no experimental tag support has been seen be
> web designers as discouraging to any change. Why, because their is no
> way to publicly test new ideas.
>
> So from the QGIS side, in order to comply we would need to reply with
> only allowed tags if a user requests WMS=1.3.0, we can reply with more
> stuff like GetPrint if they don't specify that version. Or perhaps we
> have to invent a 1.3.0+ variant specifically for when a user knows it's
> QGIS server.
>
> Anyone more familiar with WMS that can shed more light on the best way
> to work around this issue and have both compliance and the ability to
> add extra features that have no standard equivalent yet.
>
> My point still stands, that EU agencies with this concern should be
> funding compliance efforts, not removing funding for lack of
> compliance.
>
> Thanks,
> Alex
>
> On 06/07/2014 12:23 PM, Andrea Peri wrote:
>
>  Hi,
> I need to be more clear.
> My english is tremendous.
> :)
>
> The Interoperability mean to have a small set of operation euals on
>
>  EVERY
>
>  Server WMS.
>
> Equals mena same reqeust , same response.
>
> So when a Cleit WMS send a Request of GetCapabilities, The response
>
>  should
>
>  be the same from QGIS-server or from GeoServer or From Mapserver.
>
> The same response mean that every product use the same dialect the
> same
> tags and so on.
>
>
> The XSD OGC is the dictionary that every wms client and server should
>
>  use
>
>  to know the right language and tags.
>
> When the QGIS_Server response to a request GetCapbility with an XML
> that
> contains the GetPrint tags.
> The client wms say "hey what is this ? It is not in the XSD OGC. This
>
>  mean
>
>  your response is wrong."
>
> Of course there are some client wms that don0t do a validation of
>
>  response,
>
>  they HOPE that the response will be exactly as they exected.
> If this is not true. They go in crash or other bad situation.
>
> Again the resence of a Tag not compliant with XSD OGC will create
> incompatibility.
>
> Think to a client that will parse the xml response and say:
>
> ok the GetLegendGraphics tag is passed now there is "this well know
>
>  tag".
>
>  Instead arrive a GetPrint tags.
>
> The client wms become crazy.
>
> Of course QGIS will understand it.
> But this is because you (qgis group) manage it to work.
>
> But other clients don't know that tag and so they are not able to
>
>  extract
>
>  all the information from Capabilities response.
> This is a bad practice also because create artiiciosally an
>
>  incopatibility
>
>  with other products.
> Instead Inspire ask for INteroperability from every product.
>
> Interoperability don't mean use all the same unique product. (This is
>
>  the
>
>  microsoft philosophy)
> Interoperability mean All the product must use the same little set of
> command and the response at these command should be compatible
> (interoperable) between all of them
>
> Actulally this is not true for the response xml of qgis-server at a
> getcapability request.
>
> Hope to be better explain, now.
>
> Andrea.
>
>
> 2014-06-07 20:49 GMT+02:00 Andrea Peri <aperi2007 at gmail.com> <aperi2007 at gmail.com>:
>
>
>  Hi Alex,
>
> The question is not the print capability.
>
> The question is to LOST THE INTEROPERABILITY
>
> If qgis response an xml that is not OGC complaint it is not
>
>   interoperable
>
>   with other product.
>
> As example:
>
> if an public Administration will eed to do a cascading wms with the
>
>   server
>
>   wms of another public administration.
> The server before of all call for a GetCapability.
>
> If the response has a tag proprietary. If fail.
> This need Not Interoperable.
>
> I dont say do not do a getprint.
>
> I say remove tha tag GetPrint from the GetCapabilities response.
> It is not a OGC tag and so that response is not interoperable as
>
>   requested
>
>   from Inspire specification.
>
> Regards,
>
>
>
> 2014-06-07 20:36 GMT+02:00 Alex Mandel <tech_dev at wildintellect.com> <tech_dev at wildintellect.com>:
>
> On 06/07/2014 11:19 AM, Andrea Peri wrote:
>
>  Hi,
>
> AFAIK the qgis server is not complaint with Inspire.
>
> This beacausethe Response to GetCapabilities is not responding to
> the
> requisite that the OGC will require for it.
>
> Originally the qgis was simply generate an incompatible response
> for
>
>    the
>
>    XSD of OGC.
>
> The response is ncompatible for thre thinks:
>
> 1) the GetCapabilities is in the wrong namespace.
> This is a silly question anc could be easily resolved.
>
> 2)
> The presence of the GetStyle that is dismissed from OGC wms 1.3.0.
> Please notice that the Inspire require the WMS 1.3.0 .
> To resolve this the QGIS groups has copied the XSD of OGC and
>
>    modifica
>
>   it
>
>  to redirect to a different XSD not in the OGC site.
>
> 3) The presence of a Proprietary tag inserted without any reference
>
>    to
>
>   any
>
>  standard.
> The GetPrint.
> This is not present in any other product.
>
> My question is for any person of a Public Administration that plan
> or
>
>  are
>
>  funding QGIS.
>
> In Europe the Inspire directive will ask to promove the
>
>  Interoperability.
>
>  The interoperability strategy ask that every produc that allow the
>
>  inspire
>
>  directive will speak the same language using the same tags and
> functionality.
>
> The QGIS solution to add a proprietary tag and to write a own
>
>    different
>
>   xsd
>
>  that overlap the standard OGC xsd will create the presuppost
> (AFAIK)
>
>    to
>
>    vilate the Inspire directive.
>
> If this is true A Public Administration should not use the QGIS.
>
> This is a realproblem for us that invest many fund on qgis.
>
> So I like toknow the opinion of other public administration.
>
> Before still fund a product that seem to violate the Inspire
>
>    directive
>
>    principles.
>
> Thx,
>
>
>
>  To me the question is flipped. What needs to be funded, probably by
> EU
> agencies to ensure INSPIRE compliance of QGIS Server?
> It looks like you've put together the list of what needs to be
> fixed,
>
>   so
>
>   the target should be easier. I am little puzzled about not allowing
>
>   for
>
>   extra functions that are not in the standard. Unless the WMS has a
>
>   print
>
>   standard an extra print add-on doesn't break any expectations. Who
> knows, maybe that should be submitted as an extension to WMS.
>
>
> Note, this should have no effect on funding and usage of QGIS
> desktop.
> Maybe Paolo has good numbers on if EU agencies are funding Server vs
> Desktop features.
>
> Thanks,
> Alex
>
>
>    _______________________________________________
> Qgis-user mailing listQgis-user at lists.osgeo.orghttp://lists.osgeo.org/mailman/listinfo/qgis-user
>
>  _______________________________________________
> Qgis-user mailing listQgis-user at lists.osgeo.orghttp://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
>  _______________________________________________
> Qgis-user mailing list
> Qgis-user at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20140608/15d56eff/attachment-0001.html>


More information about the Qgis-developer mailing list