Re: [boost] Re: Yet another physical quantities unit library (0.1) uploadedtoyahoo groups files section

From: Phil Richards (news_at_[hidden])
Date: 2004-01-17 10:58:28

On Fri, 16 Jan 2004 21:46:27 +0000, Andy Little wrote:
> Have tested yapqul on VC7.1 . (aside "pqs-script.cpp" Needs VC7.1 language
> setting set to "force conformance in for loops" BTW...which is correct but
> not always default)
 
Well, that's a (fairly pleasant) surprise.  What does the "force
conformance in for loops" refer to - just out of interest?
 
> Overall comparing the three pqs, Yanl and yapqul... and looking at scope
> for convergence/conversions. I think there is good reason for optimism.
 
My feelings too.  I have made a start (and some significant progress) on
porting over to a boost::mpl implementation.  Although I got something
that basically works as the current form (side note: factor of 2-3 fold
slow down on compilation times), I'm having a play with the MPL to see if
I can remove even more of the "hard-coded" sizing.  (All that is left are
the conversion and printing routines.)  (It's taken a few hours to get my
brain back into functional programming style thinking - now, if we could
add Haskell to C++... :-)
 
> So at the higher level the three libs should start to look relatively
> similar.(Your params order is probably better than the one in my pqs
> Concept Check post.) On that issue... could take a look at pqs-1-00-01
> dimensional analysis error handling, which uses (basic) concept checking
> to help distinguish these type of errors.
 
Will do.  I have considered using concept checking rather than letting the
compiler generate rather cryptic error messages - pre-mpl version, it was
ok - with the mpl, the error messages are both terrifying and unreadable -
though using gfilt does produce a wonderful zig-zag pattern for about 5
minutes.
 
> There are still a lot of parts that I havent looked at yet, but
> (possibly because it is meant for a similar area of work to mine)...
> basically I like it :-) There are some parts that I think I wont agree
> with( boost archives for more:-) )... but the basic test is if I had
> found the lib on boost, I would probably have quietly downloaded and
> been on my merry way.
 
Thanks for the feedback.
 
phil

-- 
change name before "@" to "phil" for email