Re: [Geotools-devel] day 1, core and defaultcore experimental

[Ian S. <Ian...@ar...>]

Hi,

>You have indeed been productive, a new email from the cvs server seems
to 
>arrive every few minutues!

Yeah, sorry about that. We thought about working off of a repository on
a different box, so as not to cause excess traffic.


>Very much needed work, there was indeed a lot of dead junk.  There
should 
>be no need for any method to be marked as deprecated before we hit our 
>first beta, we should be culling them all prior to release.

Yeah, its kinda wierd to go beta w/ deprecations...

>We may want to have some form of name or label meta data in the 
>FeatureCollection that the Layer can use, but I see no reason any more
for 
>Layer to have any connection to a DataSource.

This brings me again to the idea of a FeatureDocument. The
FeatureDocument would be a FeatureCollection, but also have more, like
offering containment policy, locking, more meta-info, etc. whereas a
collection is really just an aggregation of various Features, whether
temporary or not...

>I'll take a look at what you are doing - the code in sun.misc is
actually 
>quite simple so we can just replicate the parts we need if necessary.

This part is now done and works.


>They won't be the first classes to end up in core, other factories live

>there.

Okay. I also moved DataSourceFinder in there.

>Note, that when I say alternate implementations I'm not necessarily
talking 
>about alternates that would work with the rest of gt2.  For example
someone 
>may want the SLD reader to create instances of their own Style objects
and 
>not use anything else from gt2.

Absolutely, which is why datasources must be very careful about their
usage of FeatureType and AttributeType - hence thowes repast problems.


>Sounds like things are going very well, keep up the fantastic work!

Thanks for the feedback.

Ian