Hi!

If you have to write new geostatistical code I think that an object oriented 
and 
generic programming language should be chosen (in primis c++ , it is efficient, 
it is compatible with
c, you can wrap it and work inside R).
The point is that in order to have a great code that is
easy understandable, maintainable and integrable you need
to build your system (for example using UML language) very accurately and 
taking into account the main actions you are performing when doing
geostatistical analysis...that today means to do a lot of different things.
I think that such projecting should be performed by many people working together
trying to get some convergence first on which objects should be built in order
to put together a good system. What I would like to say is that the 
conceptualization
(in objects, relation between objects, and function) of the "geostatistical 
analysis" should
come from a really deep analysis and a general consensus.
Think for example about  the problem of spatial (well, now we need 
spatiotemporal) indexing that becomes
really important for search strategies when you have many data: which kind of 
indexing to use?
could be an idea to use PostGis framework?

One of the most complete geotatistical software is ISatis and the target
should be to build something similar but more flexible.
Then the new Geostat Code should take care about
the temporal dimension of data not only spatial.
Finally I don't think that graphic is secondary above all during
explorative data analysis and spatial continuity study.

I hope that my opinions could be useful.
Bye
Sebastiano Trevisani


+
+ To post a message to the list, send it to [email protected]
+ To unsubscribe, send email to majordomo@ jrc.it with no subject and 
"unsubscribe ai-geostats" in the message body. DO NOT SEND 
Subscribe/Unsubscribe requests to the list
+ As a general service to list users, please remember to post a summary of any 
useful responses to your questions.
+ Support to the forum can be found at http://www.ai-geostats.org/

Reply via email to