Re: [boost] [review] PQS

From: Andy Little (andy_at_[hidden])
Date: 2006-06-10 05:50:47

Hi Oleg,

Sorry I missed your review before...
BTW. I will look back and see if there are anyones questions that I have missed,
but  if anyone feels that they asked a question that I havent attempted to
answer ( in direct reply or to someone else ), especially if they think I have
tried to evade a point, then do bring it to my attention. Please understand that
at the start of the review my review was looking like a ghost town. Now the
problem is opposite!

"Oleg Abrosimov" <beholder_at_[hidden]> wrote in message
news:4488574F.7050202_at_gorodok.net...
>I believe that in current state PQS can not be accepted to boost.
> Reasons:
> 1) It is almost useless for scientists. (of cause, it is very useful for
> engineers, though). It lacks or makes it really hard to plug-in
> different systems, such as C.G.S.E. (centimeter, gramme, second, and
> intensity of electric field from one electron is given by formula: E =
> e/r^2. Note, no k coefficient, as in SI), standard model in theoretical
> physics (velocity of light is 1, plank constant is 1, etc.),
> relativistic model, where only velocity of light is 1 (it means that
> dimensions of time and length are equal. length is time and time is
> length). This list can be expanded dramatically (almost every physical
> problem can benefit from some sort of special dimensions usage).

I said previously in reply to John Phillips, that I had decided to concentrate
on the SI system. However I have been having a rethink on that. Perhaps I need
to spend some time to see how feasible it is to make the library more generic.

 I have no experience outside the SI system and my use of units is limited to
engineering. I could try to acquire knowledge of other systems but if experience
of the SI system is anything to go by, then each system will have its own quirks
etc, etc, IOW to acquire knowledge of a particular units system is probably not
a trivial pursuit.

OTOH rather than try to make PQS more generic, what I can do is to sit down and
try to redo the documentation of the current library.  It would then be used as
a discussion point  for your ( as in all those making the requests for other
unit systems) input as to whether the concepts as documented are a good fit for
other systems or not.

In Walter Browns SIunits library, he mentions that his library can convert
between systems.

http://www.oonumerics.org/tmpw01/brown.pdf

If that is also a requirement, then I will say now that my math and physics
understanding is simply not up to the  task and I would suggest that someone
else will need to take over the design of the library.

In fact in both areas I am merely being realistic in saying that I simply do not
have the facility in maths and physics to do it justice and ideally it needs
someone else with those skills to take over the design. I would be happy then to
maintain and develop the SI 'facet' of this library though.

> 2) In PQS concept of dimensional analysis is tightly coupled with
> concept of units. I believe that they _must_ be uncoupled and
> implemented independently.

In fact they are in PQS but this is masked because I am trying to make the
library easy to use. A lot of the detail in the documentation is describing a
'view'  or interface on the underlying components, because if dealt with in the
raw they will be very unwieldy to use, and futher I dont think anyone will use
the library unless it has a nice interface.

> Dimensional analysis has its own value even without any units at all.

What is its value In practical computing without its numerical data payload?
OTOH maybe you are saying that there should only be one unit (IOW only using SI
base units with no prefixes etc)?
Or again are you saying that ( as Leland Brown was describing IIRC) that
dimensional checking should be switched on for dimensional analysis checking and
then switched off?

> I hope that tomorrow Ill find time to provide more ideas of what
> Boost.Dimensions and Boost.Units should be to meet my expectations from
> such a library.

AFAICS. Your wishlist is going to require a considerable amount of work to
implement. I will try to help, but I dont think I am in the best position to
fulfil it. Bear in mind that my Physics and Maths stops at pre-university level
and I am sure that is not adequate for the task.

I have been thinking again on what David Abrahams was saying regarding lack of
clarity in the documentation for PQS too. It may be most productive for a future
units library, rather than extend the PQS library or fix its faults to simply
try to get the documention, Concepts etc to the best possible standard. PQS may
then be more useful as a starting point for a more generic units library. In
many ways perhaps that has been its job, as has been brought out informally
during this review period.

Thanks for your review.

regards
Andy Little