Posted by jive
on September 11, 2006 at 2:57 PM PDT
Well now that FOSS4G has come and gone I am gathering up notes taken over the couse of the week.
FOSS4G Day One
Wow there is a lot to be done and it is only the first day .... Spent
the morning lost on the "bus" system, apparently M1 indicated a
train, and T1 referred to a bus - always ironic hanging with lost
Oh Catalog where art thou?
Mostly I hung out at the OSGEO booth, adjusting to jet lag, and trying
to assist catalog and metadata refugees wandering from booth to booth
in search of solace. Perhaps with the introduction of
to the OSGEO incubation process.we will be able to handle these issues
a bit better in the future.
The interesting incite is that the OSG catalog specification is viewed
as being complicated, in actual fact is no worse then the WFS
specification (and shares much in common with WFS 1.1 including a
Transaction operation). Let me sum up:
- Open Web Services are usually constructed around a
- OWS WFS is structured around a resource of "features", you
queries against the features, and you are returned a subset
- OWS Catalog is structured around a resource of "records",
issues queries against "metadata" about the records, and are returned a
subset of records that match.
No the difficulty is that when the phrase "metadata" is used, most of
us fall to sleep :-) Currently there are two camps trying to
define a record+metadata profile to store in a OWS Catalog.
Both of these camps are taking so long they are starting not to matter,
my hope is a "BOF" on simple catalog stuff this Thursday will be nice
and focused and offer a third pragmatic approach. I am going to go and
represent "client concerns" - ie give me enough information that I can
make the user experience better.
My fear is that we will get stuck in some SOAP/WSDL/OWS/REST
protocol debate, frankly it does not matter - if useful data is in
there us client programmers will use the protocol to go after it.
GeoTools 10 years on
This FOSS4G is going to be the first large gathering of all the active
GeoTools developers, really looking forward to trying to answer the
many questions Adrian Custer has stored up. Hopefully we can get a
user guide that makes some sense. It is fascinating
It was interesting that nobody wanted to relax and talk about where we
come from, instead everyone was excited about the future and what
needed attention this week!
- OSGEO involvement, we need to get some contribution
document defined and signed so we can be good on legal footing.
- Users Guide! we are not user hostile, honest. We are all
just very very
busy, please next time you engage one of our organizations for a
support contract please ask for a documentation delivery :-)
- Meta Rendering Design - basically we need to back port the
rendering api so we can start to slurp up additional rendering
standards and code (developers working on ISO based portrayal etc...)
- Oracle DataStore - we need to have a good example of
an JDBC datastore, and our oracle support needs help. Half QA and half
rescue - all needed.
- And so much more .. not sure our BOF will ever end.
In the afternoon I attend the Deegree workshop, and was surprised a
couple of times. My big surprise was that their iGeoPortal is
completely defined within the confines of the a WMSContext file.
liked their open - lets say "engagement" rather then "abuse" of this
interoptability format to gather plugins together to define a web page.
Fun stuff. Looking forward to their WFS 1.1 support.
BOF WMS Grid
Always fun about the difference between everyone knowing what needs to
be done, and then actually agreeing to it. At the end of the day it
looks like a seperate capabilities document that provides additional
information for clients so they can make requests in a specic manner so
they can get cached. I am going to look hard at this one as
forms, it was great to see the enthusiasm for getting this solved, as
well as the emphais on backwards compatibility.