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 16:12:18 +0900
- Organization: Geomatys
Le 17/10/14 09:09, Werner Keil a écrit :
>
Aside from all EG members who have an opinion, just like others e.g.
>
354 the last word is with Spec Leads.
Only on topics where there is a choice. For every topic fully determined
by the laws of logic, nobody, including the Spec Leads, can make a choice.
Now my question to Werner: why since the beginning of this tread I have
been unable to get an argument on the ground of logic with you? Do you
understand the following sentence?
If /A/ implies /B/ and /B/ implies /C/, then /A/ implies /C/.
If you accept the above sentence, can we try to restart the analysis
step-by-step? I start with one sentence, we make sure that we agree on
it, then I follow on a next sentence until we reach together the clear,
indisputable evidence that UNITSOFMEASUREMENT-62 breaks current Java
parameterized type safety. No democracy here, no possible vote, no Spec
Leads choice. The only way you can dispute my claim is to find a flaw in
the chain of logical arguments.
>
Otavio argues the wildcard version in need of a cast was "old style"
>
and frankly some new JEPs like JEP101 go in a similar direction.
This is not true. "<?>" is not a substitute for raw parameterized type.
It means that the type is unknown or can not be expressed by the current
syntax of Java parameterized type. For example I challenge any of you to
find a *safe* non-wildcard version for the following methods from
java.lang.Class:
* Class<?> getDeclaringClass()
* Class<?> getEnclosingClass()
* Constructor<?> getEnclosingConstructor()
I will answer about JEP-101 in a separated email.
Martin