Morning Frank:  

On Monday, 10 September 2012 at 4:20 AM, Frank Gasdorf wrote:  
> Hello List,
>  
> After the pull request of Wayne Beaton I read the proposal and like to
> discuss or suggest some points I had in mind:
>  
>  

I was going to go through the feedback this evening; and fill in a few more 
blanks. It looks like we need to hunt down the current list
of comitters for example.  
> Scope
>  
> * I guess the scope is to bring data from web mapping services, local
> as well as remote vector and raster data sources to a rich client
> applications, where the user can create, modify or delete spatial
> objects or their attributes AND Analyze/Query different sources with
> the same with via common interface
>  
>  

Other then "I guess" that is pretty good.  I would stick with "spatial data" 
rather then specifically raster and vector (as other data sources exist such as 
point clouds and so on).

> * gives bus business applications an easy to use extension to
> display their data on a map to give users an other point of view
> on their data sets
> * analyze data sets from a spatial perspective
> * create reports with spatial contents
>  
>  

Not too sure about this one, we may be able to roll "reports" into business 
applications item. "Provide business application easy to use components to 
display their data on a map suitable for interactive use and reporting".
> and whatis uDig NOT?
> - A GIS Application for the web to integrate into backend infrastructure.
>  
>  

Strangely enough we have an example of just that being performed a few years 
ago on the mailing list. The uDig "catalog feature" was intended to just focus 
on data access and would be suitable for this use-case.

> For the description I guess it would be helpful if we mention EMF for
> a model driven development projects. And probably this is also a kind
> of expectation to come closer to the top-level Eclipse projects like
> EMF/RCP, which means
> - review of the current code
> - with their knowledge a smoother integration into existing projects
> - a better marketing over all Eclipse projects AND in the Eclipse
> universe (User perspective instead Developers point of view)

Yeah when they hit their "Architecture" topics on LocationTech I would like to 
see EMF models for things like SLD taken up. I kind of want to hit them 
ourselves in any case.

> I guess a great benefit for the Eclipse Community would be the
> possibility for the Java-World to have an easy access to a wide range
> of spatial datasets through the connectors provided by geotools, gdal,
> jgrass and others.

Yep; the uDig Catalog IResolve system actually can let you pick and choose. We 
had an example integration with OSSIM. And I also experimented with using 
deegree (rather then GeoTools). The API changed too often due to their rewrite 
policy.  
> I personally would prefer a common lightweight API to add spatial
> content (via AdaptorFactories) to a map that acts like a Facade for
> business apps which hides the complexity for spatial access. For
> example you have several objects in your "host" application that have
> spatial attributes and the user would like to see these on map (per
> selection or whatever) the implementor should write AdapterFactories
> to convert this IWhatever Objects to an lets say IMapObject .. and so
> on
>  
Agreed; central to that story is an adapter that can convert "object" to 
"feature" (i.e. with a location). And a geocoded is normally
used for the majority of business data … which is another point for the 
"architecture" group.
  
>  
> I know the background about "It is our intension to transition as
> quickly as possible and promptly issue a release." but I don't know
> whether its a good idea to stay where we are. From my point of view
> the release cycles of other eclipse project are to long for a
> community driven not really sponsored project like ours.

I expect we will be in a position to release every six months (now that 
GeoTools has gone to that timeframe). We will probably
lag a month behind the GeoTools release based on our history doing QA.
> From the other side we're very close to the users in the geospatial OS
> Community and would like to keep them happy abput short term response
> and issue fixing. Nevertheless, probably we should try to mix this
> "half-a year release" and kind of adhoc-release a bit and should turn
> into a better planed release management. It doesn't exclude patch
> releases for the community ;)

  
> The question is, do we have time to
> - upgrade for Eclipse release train to a new Eclipse version and what
> about long-term support and support for already released versions
> - test the application itself and the integration into the current or
> older eclipse release(s)

We have traditionally lagged behind one point release for uDig; I think this 
discussion can take place once we have
migrated to LocationTech. As far as I can tell there have no goals been set in 
this regard.  

(I understood this section of the proposal to be about how quickly they could 
expect an initial release - rather than ongoing release policy)
> Technical I've JAI /Image-IO on my list because of lack of knowledge
> how to handle the "redistribution" license. IMHO we already
> redistribute it with JRE but I'm not sure, whether we are right.

I am confident we are releasing the JRE + JAI correctly (within the license 
restrictions
of the JRE and plugin system). It may actually raise more questions when 
release the JAI as
a jar rather then embedded in a JRE.  
> The other discussed option is bundling theses dependencies into bundles
> and fragments with native support. We should clarify this before
> moving on (thats the reason why I closed the pull request a few weeks
> ago).

Okay; so we can update the proposal to match.  

Aside: can you think of anyway we can ship with the non-native jars bundled; 
and override them if
JAI is available on the booth class path?

> What do you think about a skype/google hangout to discuss the points
> from the proposal. What kind of expectations, issues do you have.

Sounds good; my usual time for open source endeavours is:
- 
http://www.timeanddate.com/worldclock/fixedtime.html?msg=uDig+Hangout&iso=20120910T23&p1=47&ah=1
> I'm looking forward and interested in your opinions.
>  
> Cheers, Frank
> _______________________________________________
> User-friendly Desktop Internet GIS (uDig)
> http://udig.refractions.net
> http://lists.refractions.net/mailman/listinfo/udig-devel


_______________________________________________
User-friendly Desktop Internet GIS (uDig)
http://udig.refractions.net
http://lists.refractions.net/mailman/listinfo/udig-devel

Reply via email to