Loading ...
Sorry, an error occurred while loading the content.

Re: Schema Rel Candidate 1.0

Expand Messages
  • davewissenbach
    I ve switched to Xerces-C++ parser version 1.6.0, for schema validation. This parser now provides full support for schema. On my web site, the file
    Message 1 of 13 , Jan 1, 2002
    • 0 Attachment
      I've switched to Xerces-C++ parser version 1.6.0, for schema
      validation. This parser now provides full support for schema.

      On my web site, the file

      http://www.cableone.net/cdwissenbach/FileFormat.gpx

      now references two schemas,
      http://www.cableonenet/cdwissenbach/gpsxml_0.3.xsd

      and

      http://www.cableone.net/cdwissenbach/wissenbach.xsd

      for validating public and private namespaces, respectively. This
      example is fully valid using the sample SAXCount application
      provided with the xerces parser. In the gpsxml_0.3.xsd I experiment
      with the derivation of wpt, trkpt, and rtept from a basic point
      type, and the definition of a modelGroup for descriptive data about
      tracks, waypoints, and routes. However, doing this required a slight
      modification of the order, as elements from the base type need to
      always come first! I also added partial annotation, because I thing
      that a usable schema needs to explain itself.

      I believe that this new approach to the Schema provides better
      clarity and maintainability than the Russian Doll design previously
      used--I learned this from the book,

      Professional XML Schemas by Duckett, Griffin et Al. Wrox Press.
      ISBN 1-861005-47-4.

      Comments? If this approach is further acceptable I'll add better
      examples to the annotation.
    • Dan Foster
      Hello, I m going to spend some time this week documenting our work on GPX, and trying to wrap up some unfinished decisions, so we can start deploying GPX 1.0
      Message 2 of 13 , Jan 10, 2002
      • 0 Attachment
        Hello,

        I'm going to spend some time this week documenting our work on GPX,
        and trying to wrap up some unfinished decisions, so we can start
        deploying GPX 1.0 applications. To keep it easier to follow the
        conversations, I'll post each issue as a separate e-mail with a
        relevant subject line.

        I encourage you, if you have remaining questions or concerns about
        GPX, to share your ideas or questions now.

        Best wishes in 2002 to all of you!

        --
        Dan Foster
        TopoGrafix - GPS Software, Waypoints, and Maps
        http://www.topografix.com - mailto:egroups@...
      • Dan Foster
        Hello, Are Schemas required in GPX? Dave proposed a method for specifying a public and a private schema for GPX validation. It looks good, and I haven t heard
        Message 3 of 13 , Jan 10, 2002
        • 0 Attachment
          Hello,

          Are Schemas required in GPX?

          Dave proposed a method for specifying a public and a private schema
          for GPX validation. It looks good, and I haven't heard any
          counter-proposals. But we haven't discussed whether schemas are:
          A. required?
          B. optional?
          C. forbidden?

          B is the default answer. Anyone care to argue for A or C?

          Assuming we go with A or B, it seems like we should all refer to a
          master public schema. I'd be happy to host it (and the rest of the
          GPX documentation) here at http://www.topografix.com/gpx/ Or we can
          use the http://groups.yahoo.com/group/gpsxml/ file area, but that
          isn't accessible without logging in.

          --
          Dan Foster
          TopoGrafix - GPS Software, Waypoints, and Maps
          http://www.topografix.com - mailto:egroups@...
        • Kjeld Jensen
          ... Well let me try arguing for C (for GPX version 1.0) The way I see it we nee to publish a version 1.0 of GPX in order to gain some experience. In my
          Message 4 of 13 , Jan 10, 2002
          • 0 Attachment
            >Dave proposed a method for specifying a public and a private schema
            >for GPX validation. It looks good, and I haven't heard any
            >counter-proposals. But we haven't discussed whether schemas are:
            >A. required?
            >B. optional?
            >C. forbidden?

            Well let me try arguing for C (for GPX version 1.0)

            The way I see it we nee to publish a version "1.0" of GPX in order to gain
            some experience. In my opinion this format should be quite simple to
            implement which will have the following advantages:

            - for people who haven't participated in the discussions it is easier to
            get the idea without having to focus on schemes which requires some degree
            of XML knowledge.

            - we will probably get more people to support it as it is easier to
            implement (everybody have to start from scratch)

            - it will be easier for us to implement

            I therefore suggest, that we publish this first version of GPX without
            support for XML schemes. Instead it should be based on a number of simple
            XML compliant directives (authored by Dan and supported by the rest of us)
            published on a dedicated GPX homepage. The homepage should also maintain a
            list of applications which supports the GPX 1.0 format (measured on that
            all the directives are followed).

            This will definitely cause problems, but it is the only way to...

            - move to "working release" as soon as possible (January)
            - gain some valuable experience

            When we discover problems - probably mostly due to the lack of strict
            directives, we discuss them and implement them in 1.01 and so on.

            Along with this we should start focusing on version 1.1 which I believe
            should contain optional use of XML schemes.

            My five (euro) cents

            Kjeld

            ______________________
            Kjeld Jensen
            N 55° 22' E 10° 24'
            Email: gps@...
            http://www.cetus.dk/gps
          • davewissenbach
            Kjeld, I don t think that the schema is necessary, but I think that it can make the format a lot easier to figure out. The sample schema gpsxml_0.3.xsd on my
            Message 5 of 13 , Jan 11, 2002
            • 0 Attachment
              Kjeld,

              I don't think that the schema is necessary, but I think that it can
              make the format a lot easier to figure out. The sample schema
              gpsxml_0.3.xsd on my web site also includes annotations, which
              ultimately could be extended to also provide examples. But there is
              the downside to the schema in that it enforces a particular order on
              elements, which is not actually enforced by the applications. Right
              now my program, which I'm calling Wissenbach Map, interoperates well
              with ExpertGPS with beta gpsxml support, even though ExpertGPS
              doesn't exactly conform to the schema. (It can't, because I changed
              the element order after Dan published!). The problem right now is
              the order of elements, which I arbitrarily changed in order to use
              the principle of inheritance in the schema. (Allowing arbitrary
              element order would make the schema very complicated, which is a
              downside to using a schema.)

              But the schema can provide a central place to insert documentation,
              including a recommendation that applications which conform to the
              GPS standard should be written with a high degree of tolerance for
              such things as order and unexpected tags.

              So I'd say that applications should attempt to conform to the
              schema, but not to validate against the schema. (I validate my own
              output at development time, but I don't validate yours at run time.)
              Then, when you add new tags, with explanation, such as hdop and vdop
              to the schema, you alter the schema to include these new tags, with
              examples and annotation.

              I'll accept Dan's offer to maintain and explain, or at least host,
              the schema. I found that I had to put a copy on my personal web page
              in order to see it and validate. Without the support of an authoring
              program such as ExpertGPS and a reading program such as EasyGPS the
              gpsxml format won't reach critical mass.

              In the long run the presence of a well-documented, professionally
              presented schema, with sample output which corresponds exactly to
              that schema, will be the only way to enable others to interoperate
              with our applications. The 100-odd messages in this discusssion
              group have become way too complex for any of us to follow. Remember
              that standards are what made the World Wide Web possible in the
              first place.

              Dave


              --- In gpsxml@y..., Kjeld Jensen <gps@c...> wrote:
              >
              > >Dave proposed a method for specifying a public and a private
              schema
              > >for GPX validation. It looks good, and I haven't heard any
              > >counter-proposals. But we haven't discussed whether schemas are:
              > >A. required?
              > >B. optional?
              > >C. forbidden?
              >
              > Well let me try arguing for C (for GPX version 1.0)
              >
              > The way I see it we nee to publish a version "1.0" of GPX in order
              to gain
              > some experience. In my opinion this format should be quite simple
              to
              > implement which will have the following advantages:
              >
              > - for people who haven't participated in the discussions it is
              easier to
              > get the idea without having to focus on schemes which requires
              some degree
              > of XML knowledge.
              >
              > - we will probably get more people to support it as it is easier to
              > implement (everybody have to start from scratch)
              >
              > - it will be easier for us to implement
              >
              > I therefore suggest, that we publish this first version of GPX
              without
              > support for XML schemes. Instead it should be based on a number of
              simple
              > XML compliant directives (authored by Dan and supported by the
              rest of us)
              > published on a dedicated GPX homepage. The homepage should also
              maintain a
              > list of applications which supports the GPX 1.0 format (measured
              on that
              > all the directives are followed).
              >
              > This will definitely cause problems, but it is the only way to...
              >
              > - move to "working release" as soon as possible (January)
              > - gain some valuable experience
              >
              > When we discover problems - probably mostly due to the lack of
              strict
              > directives, we discuss them and implement them in 1.01 and so on.
              >
              > Along with this we should start focusing on version 1.1 which I
              believe
              > should contain optional use of XML schemes.
              >
              > My five (euro) cents
              >
              > Kjeld
              >
              > ______________________
              > Kjeld Jensen
              > N 55° 22' E 10° 24'
              > Email: gps@c...
              > http://www.cetus.dk/gps
            • Dan Foster
              Hello, Thursday, January 10, 2002, 5:01:32 PM, I wrote: DF Are Schemas required in GPX? I didn t hear anyone argue that we should *never* support schemas.
              Message 6 of 13 , Jan 16, 2002
              • 0 Attachment
                Hello,

                Thursday, January 10, 2002, 5:01:32 PM, I wrote:

                DF> Are Schemas required in GPX?

                I didn't hear anyone argue that we should *never* support schemas.
                (Kjeld argued that we should support them at a later date). So it
                sounds like there's consensus that schemas are a good thing, and GPX
                should eventually support them.

                Unfortunately, if we're to have backwards compatibility, we need to
                make some decisions now (like the order of elements) so that we'll be
                able to apply a reasonable schema to GPX in the future.

                As I sat down to start writing documentation, it became clear that the
                detail I was putting into each element description was really just a
                schema in English. Also, I started thinking about how we were going
                to "certify" new GPX developers (to make sure their GPX output was
                valid). Having a schema is a very fast way to write a validation
                checker - just see if the new files validate against the schema!

                So, here are my recommendations:
                1. We should create an official schema before releasing GPX 1.0.
                2. We should agree on an official validation tool and method for
                certifying that GPX files are valid.
                3. We should adopt a namespace versioning method so that new versions
                of GPX each have a new schema.
                4. I should write some documentation to accompany the GPX schema,
                since reading schemas is a bit tricky. (But the schema is still
                the official definition of GPX.)

                I've read through the schemas proposed by Kevin and Dave, and created
                a new version 0.4 schema which draws on both ideas. Strangely, I
                found that the great documentation in Dave's latest schema actually
                made it harder to read, so I've stuck with inline comments. I stuck
                with Dave's element ordering, but used the flat format of Kevin's
                schema to make it easier (to my mind) to determine what tags are
                allowed for each object.

                To support versioning, I've created sub-folders at
                http://www.topografix.com/GPX/ for the major and minor versions of the
                standard, so version 0.4 lives at
                http://www.topografix.com/GPX/0/4/gpx.xsd

                As far as validation goes, I've started using the Xerces 1.6.0 parser
                as suggested by Dave, and suggest that we adopt it as our validation
                tool. I'll put together some instructions for validating GPX files if
                we choose to go this route. (Dave: I'm using the following options:
                "saxcount -n -s -f -v=always test_gpx_file.gpx")

                I've updated my software to reference this schema, and have created a
                private schema for testing purposes.
                http://www.topografix.com/GPX/Private/TopoGrafix/0/1/topografix.xsd

                I'll post later tonight once new beta versions of my software are
                available for use.

                Please give some thought to the recommendations I've proposed, and
                share your comments.
                --
                Dan Foster
                TopoGrafix - GPS Software, Waypoints, and Maps
                http://www.topografix.com - mailto:egroups@...
              • davewissenbach
                ... created ... stuck ... Yes, as I was writing the 0.3 schema I came to the same realization, and my intention was to patch it up with examples in the
                Message 7 of 13 , Jan 16, 2002
                • 0 Attachment
                  --- In gpsxml@y..., Dan Foster <egroups@t...> wrote:
                  > Hello,
                  >
                  ...
                  >
                  > I've read through the schemas proposed by Kevin and Dave, and
                  created
                  > a new version 0.4 schema which draws on both ideas. Strangely, I
                  > found that the great documentation in Dave's latest schema actually
                  > made it harder to read, so I've stuck with inline comments. I
                  stuck
                  > with Dave's element ordering, but used the flat format of Kevin's
                  > schema to make it easier (to my mind) to determine what tags are
                  > allowed for each object.
                  >

                  Yes, as I was writing the 0.3 schema I came to the same realization,
                  and my intention was to patch it up with examples in the annotation.
                  And of course you recognize that some of the documentation in that
                  schema was lifted straight out of your own posts. What I was doing
                  was experimenting with the idea of inheritance in the schema.

                  I think that going back to the anonymous type method proposed by
                  Kevin Read is OK. The schema will probably be slightly more
                  difficult to maintain and extend, but much easier to read.

                  > To support versioning, I've created sub-folders at
                  > http://www.topografix.com/GPX/ for the major and minor versions of
                  the
                  > standard, so version 0.4 lives at
                  > http://www.topografix.com/GPX/0/4/gpx.xsd
                  >

                  I think that versioning will work well. I assume that a version 0.5
                  schema will validate a 0.4 document--or that we'll try to keep
                  backward compatibility as the schema is extended to include more
                  concepts.

                  > As far as validation goes, I've started using the Xerces 1.6.0
                  parser
                  > as suggested by Dave, and suggest that we adopt it as our
                  validation
                  > tool. I'll put together some instructions for validating GPX
                  files if
                  > we choose to go this route. (Dave: I'm using the following
                  options:
                  > "saxcount -n -s -f -v=always test_gpx_file.gpx")
                  >
                  I used the same command line but left off the -v=always.

                  > I've updated my software to reference this schema, and have
                  created a
                  > private schema for testing purposes.
                  > http://www.topografix.com/GPX/Private/TopoGrafix/0/1/topografix.xsd
                  >
                  > I'll post later tonight once new beta versions of my software are
                  > available for use.
                  >
                  > Please give some thought to the recommendations I've proposed, and
                  > share your comments.
                  > --
                  > Dan Foster
                  > TopoGrafix - GPS Software, Waypoints, and Maps
                  > http://www.topografix.com - mailto:egroups@t...
                • Dan Foster
                  Hello, I ve updated my beta software to read/write GPX 0.4. The new software can be downloaded from: EasyGPS: http://www.easygps.com/beta.asp PanTerra:
                  Message 8 of 13 , Jan 18, 2002
                  • 0 Attachment
                    Hello,

                    I've updated my beta software to read/write GPX 0.4.

                    The new software can be downloaded from:
                    EasyGPS: http://www.easygps.com/beta.asp
                    PanTerra: http://www.topografix.com/beta.asp
                    ExpertGPS: http://www.expertgps.com/beta.asp

                    --
                    Dan Foster
                    TopoGrafix - GPS Software, Waypoints, and Maps
                    http://www.topografix.com - mailto:egroups@...
                  • Dan Foster
                    Hello, I ve written several GPX documents to provide an overview of GPX (for developers and for GPS users). I still need to write a developer s document
                    Message 9 of 13 , Jan 18, 2002
                    • 0 Attachment
                      Hello,

                      I've written several GPX documents to provide an overview of GPX (for
                      developers and for GPS users). I still need to write a developer's
                      document explaining the specifics of writing a GPX.

                      Please feel free to comment on the documentation. Offline comments
                      can be sent to gpx@...

                      Overview: http://www.topografix.com/gpx.asp
                      Links to:
                      GPX for GPS Users
                      GPX for Developers
                      Validating your GPX output
                      GPX Resources
                      GPX Schema 0.4

                      If you'd like your GPX application or webpage listed on the GPX
                      Resources page, send me your URL. I won't list your app unless you
                      request it, since we're all still in development.

                      --
                      Dan Foster
                      TopoGrafix - GPS Software, Waypoints, and Maps
                      http://www.topografix.com - mailto:egroups@...
                    • Kjeld Jensen
                      Dan, I must say that your GPX promotion homepage looks really great! Naturally I would like Cetus GPS http://www.CetusGPS.dk to be added. I have spent an hour
                      Message 10 of 13 , Jan 19, 2002
                      • 0 Attachment
                        Dan, I must say that your GPX promotion homepage looks really great!

                        Naturally I would like Cetus GPS http://www.CetusGPS.dk to be added.

                        I have spent an hour looking into the schema, and after that it really
                        doesn't look that complicated. I suppose my opinion will be modified a bit,
                        but I will get back to that.

                        For now I'll just ask why the elements

                        <sat></sat>
                        <hdop></hdop>
                        <vdop></vdop>
                        <pdop></pdop>

                        aren't a part of the schema? I think we agreed to this earlier?

                        Best regards
                        Kjeld Jensen

                        ______________________
                        Kjeld Jensen
                        N 55° 22' E 10° 24'
                        Email: kjeld@...
                        http://www.CetusGPS.dk
                      Your message has been successfully submitted and would be delivered to recipients shortly.