Use Advanced Search to search the entire archive.
Re: Fwd: Re: Composable Analytics: A FBP framework for analytical applications
- From: Martin Desruisseaux <
>
- To:
- Subject: Re: Fwd: Re: Composable Analytics: A FBP framework for analytical applications
- Date: Sun, 19 Oct 2014 01:33:14 +0900
- Organization: Geomatys
Le 18/10/14 20:39, Werner Keil a écrit :
>
(...snip...) but it highlights the challenge of "semantic" or
>
mathematic conformity aside from mere language boundaries.
Good.
>
I can't say, if Paul or John will provide feedback to the related
>
question around cast or not to cast, but given they both have been
>
programming longer than pretty much all of us lived (Paul started with
>
IBM in 1953, wrote first banking apps in the 60s, FBP which he
>
invented is among the precursors to modern OOP;-) they might have some
>
opinion on that, too.
>
Yes, if you think that we can. I can send a question proposal to this
mailing list if peoples agree.
>
Paul Morrison scripsit:
>
>
Time to repost the warning I have to give every few years or so,
>
that dimensional conformity does not necessarily mean
>
compatibility.
>
Torque and work have the same units (kg m^2 / s^2), but one is
>
a vector
>
and the other is a scalar; likewise, length, width, and height
>
can be
>
multiplied but normally not added, even though they are all
>
meters.
>
Just before Paul asked if JSR-363 has some solution for that. The answer
is yes, since Unit are parameterized by the Quantity interface, and the
Torque and Energy quantities are two distinct interfaces even if the
Unit have the same dimension. This is in my opinion the greatest
achievement of JSR-275 compared to JSR-108, which was using only the
dimension. This improvement is to put on Jean-Marie's credit.
Note that Paul's "conformity does not necessarily mean compatibility"
comment applies also to dimensionless units: e.g. Angle versus
concentration (e.g. mg/kg) versus percentage. I think we need to explain
that in the specification.
Martin