From: Phil Richards (news_at_[hidden])
Date: 2006-06-22 14:51:58
On 2006-06-21, Robert Ramey <ramey_at_[hidden]> wrote:
> Phil Richards wrote:
> > It's nice to see that there appears to be a move in the direction
> > of separating the dimensional analysis and units into orthogonal
> > bits, though. It's the way I've gone... :-)
> More than that - I believe its the only way something like this (which is
> really needed) is going to get into boost at all.
I agree with that too.
[the stages to a solution...]
> layer 1 dimensional analysis core
> layer 2 specific dimension systems - usages of the above library - probably
> more than one
Probably need to do 1 and 2 together - otherwise the library is
practically useless. Demonstrate that layer 1 works with a layer 2 SI
implementation (the one that the majority of people would want, I
suspect), but with a clear example of how other variants can work.
> layer 3 units core - extension of dimensional analysis library
This is one where I've been having a hard think. I think it is possible
to have units done independently of DA. When units are applied to a
dimensioned value, then various smarts can kick in to do with
multiplication and division, but there's a lot that can be done with
just addition, subtraction and scaling (I wish my Maths degree wasn't
20 years ago or I'd remember the proper name for this type of space).
In particular, we could make sure that absolute+relative (point+vector,
or whatever else you want to call it) stuff works cleanly. (i.e., the
difference between conversion of a temperature difference of 1C being
1.8F, and 0C being 32F.)
> layer 4 specifice units systems.
Similar comment to layer 2.
> Now I know from experience that lots of people will find say - "if it
> doesn't include xyz... its just useless". Well, those people are wrong.
> Also it would make the library(ies) much easier to understand and use.
And on this too.
-- change name before "@" to "phil" for email
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk