Use Advanced Search to search the entire archive.
Re: Remove "generic" multiply/divide operations from Quantity
- From: Martin Desruisseaux <
>
- To:
- Subject: Re: Remove "generic" multiply/divide operations from Quantity
- Date: Sat, 18 Oct 2014 15:18:48 +0900
- Organization: Geomatys
Hello Leonardo
Thanks a lot for your replay, it is an excellent summary :-)
Le 17/10/14 04:46, Leonardo Lima a écrit :
>
In the end, I understood that:
>
What I (really) wanted can't be achieved without "exploding" the JSR
>
to have bajilion methods. That's not feasible.
Right, I think we all agree on that.
>
The convenient way to achieve somewhat what I wanted (the signature
>
that sparked this thread) breaks Java rules of generics expectations.
>
That's Martin's main point.
Exactly. Millions of thanks for understanding that!!!
>
The best way to continue with Generics is using the <?> and then force
>
the API user to cast.
>
>
1. But, then, why not drop it from the signature?
>
2. And add signature in JSR 363.1 when better support is
>
available in the language?
>
Right. Question 1 is an open question which could be the subject of a
separated thread, once the UNITSOFMEASUREMENT-62 issue has been
understood by everyone.
Werner, the question #1 raised by Leonardo above is a question that
could be submitted to a vote, because it is a _*choice*_ that the group
would need to do. But the fact that the current UNITSOFMEASUREMENT-62
breaks Java parameterized type safety *can not* be submitted to vote -
it is a logical _*fact*_. The only choice we could do regarding
UNITSOFMEASUREMENT-62 would be "/do we accept to break Java
parameterized type safety?/" I argue that this choice is not under
JSR-363 power.
Martin