Here is my feedback:

* getInfo()

I guess this is a little off topic of this proposal and related to the 
last. But I was surprised when i saw it go through today. I remember 
having voiced feedback about those interfaces not being fine tuned 
enough. All those attributes: schema, publisher, title, etc... dont 
apply in the common case, they really only apply to WFS as far as i can see.

What I want is to have a bare minimum. Like just uri. Everything else 
like the above can be thrown in a subclass. That way implementors can 
chose to implement additional metadata interfaces if they choose, and 
only a minimum in the normal case.

Also some of the attributes (like icon), need to be rethought. An icon 
is something that a ui will want to set, not something provided by a 
datastore implementor.

I also think ServiceInfo and GeoResourceINfo are bad names. They imply 
catalog, and as it says in the proposal "they are not catalog".

I dont know why my vote says +0... i remember having serious 
reservations on this one. Perhaps it is because i disapeared over this 
last week.

* FeatureAccess

I dont know how I feel about a new interface. I think we may want to 
consider a different solution with generics. Not that I love generics 
but in this case they may better apply. I assume gabriel has thought 
about this route?

I think the name of this class is pretty bad to.

* FeatureStore

Having FeatureStore extend ComplexFeatureStore seems odd to me. I know 
that this mirrors the feature model but... In the least we *need* to 
change the name of this to something without the word "Complex" in it.

But again... the new interface bugs me, and is invasive. From what i can 
tell the main motivation is to be able to override methods which take a 
feature reader with methods which take a ComplexFeatureReader. I am not 
sure this is necessary, see next point.

* ComplexFeatureReader/ComplexFeatureWriter

Again, not to beat a dead horse... but i think these super interfaces 
are unnecessary and generics might be cleaner. I would even rather just 
have next() return an object and have client code cast, rather then 
break out a new interface.

So to sum up, this proposal needs to consider the alternative way of 
doing this which is cleaner. Adding a bunch of super classes which are 
poorly named gets a -1 in my book. Perhaps its impossible to do with 
generics... but i think its worth consideration.

-Justin

Gabriel Roldán wrote:
> Hi all,
> 
> the proposal to extend the geotools data access interfaces to allow working 
> with GeoAPI Feature and FeatureType as a superset of the current API based on 
> SimpleFeature and SimpleFeatureType is ready for your evaluation.
> 
> The URL is 
> <http://docs.codehaus.org/display/GEOTOOLS/FeatureAccess+super+class+for+DataStore>
> 
> And feel free to raise any concern as comments in the jira tracker for it at 
> <http://jira.codehaus.org/browse/GEOT-1701>
> 
> Best regards,
> 
> Gabriel
> 
> -------------------------------------------------------------------------
> 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
> 
> !DSPAM:4007,47a9db0026015332866982!
> 


-- 
Justin Deoliveira
The Open Planning Project
[EMAIL PROTECTED]

-------------------------------------------------------------------------
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

Reply via email to