Hey Jody, all,
A brief update on my repository work, since you "asked".
On Thu, 2008-01-17 at 16:04 -0800, Jody Garnett wrote:
> Bet you had forgotten about this thread acuster :-)
"forgotten" is one way to think of it. "overwhelmed" is another. I am
however working actively on this, albeit a minor but recently dominant
activity looking at ISO geometries.
In brief, the problem domain as I understand it, includes all the
aspects of "Content Storage". This is a superset of uDig needing to
maintain references to resources, of geo(or other)server needing to
track preferences and type information, of a web metadata registry
server needing to store schemata, vocabularies and instance documents,
and of everyone to track user info and passwords. So far I have finally
gotten a good handle on what I need to understand completely and how I
can realistically get started. Here's an outline of what I need to grok
(understand completely, from grok-'to drink fully'):
Content
* Define the atoms that need to be stored:
e.g. ServiceInfoBlob, DataBlob
* Understand the lifecycle of the different elements
this includes the changes in status we need to display
* Understand the workflows in which users interact with the
elements
Model
* Learn the ebRIM (Registry Information Model) enough to
understand how to map the former consistently into a registry
model. Compare this to how one would do it in the strictly
hierarchical view used by the JCR approach.
At this level, I want to tackle how we guarantee the integrity
of our storage mechanism and how we recover after an integrity
violation. This should also assess how versioning might one day
be added into the model.
View/Controler
* Map the workflows into the control needed by the user
* see what 'views' of the content will be required
Interfaces
* Complete the initial exploration of JAXR and JCR
* Get a handle on the interfaces used for security (login/password
management)
Implementations
* security both for JVM based app, for eclipse rcp, and for JEE
* storage: compare OMAR (a JAXR impl) and Jackrabbit (JCR)
GUI
* Interface to the 'registry service' which is the controller that
allows one to define, add, remove, or alter 'stored content'
* Interface to the 'repository admin service' which allows one to
load, dump, and monitor the different blobs of stored content.
* Interface to the 'catalog service' which allows one to query,
Each of these, will also need to have 'user admin' components.
After understanding a lot more of this, I will be able to tackle the
code aspects of this work. My personal focus is on the uDig registry and
hopefully I can do this at the Geotools level. Our MDWeb collaborators
have a JEE level mandate, a narrower set of needs, and two students so
hopefully they can tackle the parallel aspect of doing this in a JEE
environment (glassfish).
So the aim is:
'Design for the full case, build piecewise the narrow cases'
Jody, it's great that you are ripping out the 'catalog' stuff out of
Geotools for a clean start. I gather there were pieces in gt api, in
main, and all the way out to spike. Please let me know when you are done
so I will know where I can play once I allow myself to look at code
again.
--adrian
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel
