Jody Garnett ha scritto: > Andrea Aime wrote: >> (18.32.12) aaime: Hum, it seem the feature model review triggered a >> good chain reaction >> (18.32.17) aaime: but Jody is missing from the mix, sigh >> (18.32.34) aaime: I guess he's the only one willing to face the >> data/metadata mix issue... >> (18.32.59) aaime: btw, you're here. jgarnett, when we'll see your >> opinion on the matter? :) >> > I am still "missing" until the weekend; FOSS4G prep and a project > deadline have my kind of hammered. I have started an open office > document to answer your questions, (using code examples when I can). I > tried to do a cost/benifit of the various features to help us out in the > functinality/complexity compromise. And have a refactoring idea that can > take away some of the "extra" methods (like getAttributes() and > getAssociations()) from SimpleFeature - nothing earth shattering but it > will help people starting out in an IDE for the first time. > > I think the think the hold up comes down to the following design choice > between "Multiplicity" (what we did) and "Nested Children" (what deegree > did): > - http://docs.codehaus.org/display/GEOTOOLS/Feature+Model+Design+Discussion
Being an OO type, I find what deegree did much more natural, saner. > From a users standpoint (someone hitting up the data structure for > answers) you can tell the difference when you do xpath queries. Keeping > attribute name/value paris in order allows us > to model exactly what the XML is doing. Sorting the keeping attributes > as name/collection pairs does not. > 1- allows us to match xml > 2- allows us to match xpath > 3- allows us to take on observation and measurement things? (Verify with > Gabriel) > 4- allows us to handle data "as is" without regard for validity > > Point number #4 was the killer for me, I want us to be able to handle > data even when it is wrong or we do not know the schema. > Jody > > PS. Now that you understand the trade off you can see why we were not > finding List to be that useful a binding, it is only useful when > attribute order is significant. > PPS. As I tired to find the link I realize that we have spent more blood > sweat and tears on this FM thing then anything else, ever. I understand. Yet I feel we lack a middle ground allowing to represent associations in an efficient way, the same way deegree did. Cheers Andrea ------------------------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Still grepping through log files to find problems? Stop. Now Search log events and configuration files using AJAX and a browser. Download your FREE copy of Splunk now >> http://get.splunk.com/ _______________________________________________ Geotools-devel mailing list Geotools-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geotools-devel
