[OSGeo-Discuss] How open source and OGC/ISO can work together - GeoAPI example

[Jody Garnett]

I mentioned GeoAPI a couple of times as an example of a "bridge" between 
open source software and the OGC / ISO standards process. I passed on an 
informal report we got in our IRC meeting on Monday, here is an update 
we got today.

Have a read of this email; and think if this level of involvement will 
be enough for us. It may be easier then setting up a formal arrangement 
- it has the strength of OSGeo making contributions in a format we do 
well (ie code) and OGC making contributions in a format they do well (ie 
large pdf documents and xml schema). Also note that the GeoAPI project 
is not limited to Java projects.

Martin Desruisseaux wrote:
> Hello all
>
> Sorry for being late, I have been in a rush for the past few days.
>
> There is a quite report about what happened at OGC meeting, and what could be 
> our road map if peoples agree (at the very end of this email). First, a quick 
> reminder of GeoAPI history:
>
> * At some point in the past, the OGC 01-009 implementation specification
>    ("Coordinate Transformation Service Implementation Specification") was
>    accompanied by a set of Java interfaces created by the OGC 01-009
>    authors. Those interfaces were defined in the "org.opengis.cs",
>    "org.opengis.ct" and "org.opengis.pt" packages.
>
> * An independent contributor offered in 2001 additional Java interfaces
>    in the "org.opengis.cv", "org.opengis.gc" and "org.opengis.gp" packages
>    derived from OGC 01-004 ("Grid Coverage Service Implementation
>    Specification").
>
> * GeoAPI has been created in 2002 by James McGill in an effort to bring
>    some open source projects (especially Geotools and Jump) to a common
>    set of interfaces. While we used OGC interfaces (GeoAPI 1.0 was just
>    a packaging of the above-cited org.opengis cs, ct, pt, cv, gc and gp
>    packages), the GeoAPI project was formally independent from OGC.
>
> * About at the same time, a OGC working group, namely GO-1, had similar
>    goals. The OGC GO-1 working group (leads by Polexis) was in touch with
>    the GeoAPI community. In June 2004, we formally merged the projects. We
>    wrote a charter (link below) and got it approved by a OGC vote, as OGC
>    document 04-059.
>
>    http://geoapi.sourceforge.net/stable/site/charter.html
>
> * The join effort between OGC GO-1 working group and GeoAPI community
>    resulted in a rewrite of the referencing interfaces (for alignment
>    with ISO 19111) and addition of geometry interfaces (ISO 19107).
>
> * GO-1 final (OGC 03-064r10) has been released almost one year later
>    (May 4th, 2005) and approved as a OGC implementation specification.
>    GeoAPI 2.0 has been released slightly after GO-1 (June 7th, 2005)
>    in order to reflect the GO-1 state. Many GeoAPI 2.0 interfaces are
>    explained in this OGC 03-064r10 document.
>
> * Polexis (now SYS Technologies) seems to have disengaged from GO-1/GeoAPI.
>    I'm not aware of any progress in the GO-1 project since the OGC 03-064r10
>    release in 2005. Consequently, GeoAPI lost its representative at OGC.
>
> * GeoAPI work continued, thanks to the GeoAPI community, but we were
>    failing to report our progress at OGC. I did some attempts to bring
>    "GeoAPI request for changes" at OGC in 2006, but failed either because
>    the document was not posted on the OGC web site (none of us were OGC
>    members), or because I was not able to attend to the OGC meetings.
>
> * At the OGC technical meeting at Edinburgh (June 30, 2006), OGC members
>    voted the dissolution of GeoAPI working group on the basis that it was
>    inactive. It was not a move against GeoAPI itself, just the normal
>    procedure for a group that didn't reported any progress for 2 years.
>    Maybe the apparent absence of activity was caused by the OGC's GeoAPI
>    mailing list being totally silent; not all OGC members were aware that
>    all our discussion happen on the SoureForge mailing list.
>
> * None of us in the GeoAPI community were aware of this dissolution. I
>    learned this fact only in May 2007 from an other OGC member (Spot Image).
>    I wrote to the OGC director and we decided (on his recommendation) to
>    submit the GeoAPI request for changes at OGC anyway. We got a two hours
>    slot on Monday, and Adrian explained the GeoAPI context to the members
>    present. I presented the Request For Changes after that.
>
> * It was not possible to get a vote on the Request For Changes because
>    the GeoAPI working group was not existent anymore. However we got a
>    vote on the following motion:
>
>      "The GeoAPI ad-hoc meeting proposes to the Open GIS Consortium
>       Technical Committee to establish a GeoAPI Standards Working Group
>       to revisit and revise the GeoAPI interfaces which expand the GO-1
>       Implementation Specification (which is the OGC approved GeoAPI
>       Implementation specification)."
>
>        Moved by Martin Desruisseaux
>        Seconded by Ron Lake
>
> * In order to form such a working group, we need at least 5 OGC members.
>    Adrian and myself will be on this group. We need 3 additional members.
>    Jody, is Refraction Research an OGC member? If yes would you like to
>    be on this group? I don't think that it would involve much additional
>    work compared to your current GeoAPI involvement, but it would make
>    your involvement visible to OGC. If Stephane Nicoll accepts to be on
>    this group too, it would be very nice.
>
> * After some informal discussion between Adrian and Carl Reed, we are
>    going toward the following proposal:
>
>      - Forget about GO-1.
>
>      - As a working group, our task would be to produce a brand
>        new specification. I suggest to call this new specification
>        "GeoAPI 3.0".
>
>      - Since we would start on new grounds, we have some freedom
>        for cleaning some pending issues in GeoAPI. We are not too
>        tied by the past. However I suggest to continue what we
>        already do ("deprecate, then delete" cycle, document changes
>        in RFC document when we can) as a respect for our users.
>
>      - We could make GeoAPI 2.1, 2.2, etc. releases on our own as a
>        transition toward GeoAPI 3.0, but we should make it very clear
>        that they are not official OGC specifications. Only GeoAPI 2.0
>        was and GeoAPI 3.0 would be, if approved by OGC members.
>
>
> Any comments, discussion, objections? Any volunteer for being member of OGC 
> working group? I suggest that OGC members continue the discussion on the usual 
> SourceForge mailing list (no secret mailing list).
>
> 	Martin