Hi Martin,

Do you think changing it to something more specific like "utilitytype" or 
"utiltype" would make more sense? I suppose that sis/util/type kind of knocks 
that out but still...

Just thinking out loud here.
Adam

On Dec 8, 2012, at 4:39 AM, Martin Desruisseaux wrote:

> Hello all
> 
> The commits in the last few days included:
> 
> Implementations of interfaces derived from ISO 19103: AbstractName, 
> LocalName, ScopedName, MemberName, TypeName. Unfortunately, those objects are 
> not easy to understand, since the ISO 19103 specification is not very 
> explicit. But they appear in XML documents, so we have to support them in 
> some way... The package javadoc does it best for trying to explain them:
> 
> https://builds.apache.org/job/sis-jdk7/site/apidocs/org/apache/sis/util/type/package-summary.html
> 
> I'm uncomfortable with the "type" package name. The current content is more 
> about naming (indeed, the package name in Geotoolkit.org was "naming"), but 
> with the addition of Type, RecordType, RecordSchema and Record classes from 
> ISO 19103, the package starts looking a bit like the java.lang.Class and 
> Field architecture. I have been unable to find something better than "type" 
> for now, but I would still like to find a better none.
> 
> 
> The other commits are in the internal packages. This is a set of JAXB 
> adapters for marshalling/unmarshalling the ISO objects. For now just the 
> basic classes have been committed. The package names try to follow the XML 
> prefixes used in the ISO standards, e.g. "jaxb.gmd" contains adapters related 
> to objects to be written in the "gmd" namespace, etc. It is not easy to get a 
> big picture of those adapters neither. We tried hard to make them as 
> straightforward as possible without too much explosion of the amount of 
> classes, but the ISO 19139 specification still a quite convolved one...
> 
>    Martin
> 

Reply via email to