On 19 déc. 2011, at 16:17, Hardy Ferentschik wrote:
I am sitting in a similar boat than Sanne. I don't know how much
useful feedback I can provide w/o first getting to
know the actual problem space better.
Looking at the Jira issue and comments, I like the idea of having the @Spatial
annotation. One think which comes
to my mind here is how this all relates to supporting geospatial data types in Core -
https://hibernate.onjira.com/browse/HHH-6509
Will our approach work with these types? Is this completely orthogonal to the mentioned
Core data types?
It's a very good question.
Both problem are mainly isolated and these Core types are not used by Hibernate Search but
the Java object they convert into are JTS types and
another type that Karel is planning to use. So both features will be consistent.
What exactly offers this JTS library?
It's the defacto standard in Java to represent spatial objects like Points etc.
Regarding the DSL, I would definitely go for the approach where we have a unit
(.within(2, Unit.km)) instead of .within(2).km().
Interesting, care to articulate your reasoning? I think I like you approach but I
can't grasp why.