Use Advanced Search to search the entire archive.
[jsr363-experts] Re: Should JSR-363 API provide the SI class?
- From: Werner Keil <
>
- To: "
" <
>
- Subject: [jsr363-experts] Re: Should JSR-363 API provide the SI class?
- Date: Tue, 16 Dec 2014 12:20:15 +0100
Let's hope I am, but we are unlikely to have the Public Draft out before
say JavaOne (that would be a good time to present stuff also to a wider
audience;-) so either in January or June I will certainly be at a F2F.
January is all down to PMO/Oracle travel ("Santa Larry") still getting me a
ticket "through the chimney";-D
Regards,
Werner
On Tue, Dec 16, 2014 at 12:10 PM, Leonardo Lima
<
>
wrote:
>
>
We could ask both the ME folks at Oracle and PMO e.g. at our next
>
> meetings, but I feel we need to implement both.
>
>
>
We should do this, yes, when we're presenting at the F2F we ask for their
>
input.
>
>
Regards.
>
Leonardo.
>
>
>
On Tue, Dec 16, 2014 at 8:50 AM, Werner Keil
>
<
>
>
wrote:
>
>
>
> No, you either use ALL or none.
>
>
>
> Since Area, Angle,... all extend Quantity I don't think an implementation
>
> would be forced to use all of them (we would have to check that in the TCK,
>
> the important element is Quantity, the others are only Generic "markers")
>
> but e.g. for e.g. the spi package, an implementation would IMHO have to
>
> implement QuantityFactory, SystemOfUnits and each of the Service interfaces
>
> to be compliant;-)
>
>
>
> We could ask both the ME folks at Oracle and PMO e.g. at our next
>
> meetings, but I feel we need to implement both.
>
> The Parser is extended by UnitFormat (but as emphasized is crucial on its
>
> own for creating a UCUMParser or similar parsers) so anything that
>
> implements UnitFormat automatically implements Parser, too, hence you get
>
> two for the price of one there;-D
>
>
>
> Werner
>
>
>
>
>
>
>
> On Tue, Dec 16, 2014 at 11:42 AM, Leonardo Lima
>
> <
>
>
> wrote:
>
>>
>
>> To make it short: I think if we define a package as optional, everything
>
>> inside is optional, so you can make a package with say 10 interfaces
>
>> optional and implement 2 of them.
>
>>
>
>> On Tue, Dec 16, 2014 at 8:34 AM, Martin Desruisseaux <
>
>>
>
>
>> wrote:
>
>>>
>
>>> Le 16/12/14 19:27, Leonardo Lima a écrit :
>
>>> > I guess if we have each as a class under an optional package, it would
>
>>> > be ok.
>
>>>
>
>>> Maybe phrasing my question in an other way: does optionality have to be
>
>>> at a package level? Can it be at a class/interface granularity level? Or
>
>>> at a member level?
>
>>>
>
>>> Martin
>
>>>
>
>>>