Re: [boost] Re:Re new compile-time dimensional analysis and unitlirary

From: Matthias Schabel (boost_at_[hidden])
Date: 2003-12-11 10:38:31

> // quote from test_units.cpp  line 134
>
>     SI<double>::Length    l1(7777.0),
>                           l2(l1);
>     l2 /= 3333;
>     l2 *= 3333;
>
>     std::cout << std::boolalpha
>               << (l1 == l2) << std::endl
>               << (l1 != l2) << std::endl
>               << (l1 <= l2) << std::endl
>               << (l1  < l2) << std::endl
>               << (l1 >= l2) << std::endl
>               << (l1  > l2) << std::endl
>               << std::endl;

I'm guessing that you're making an oblique reference to numeric  
precision issues and their interplay with the comparison operators...   
The units library is templated on value type, which is not restricted  
to POD types, as I demonstrate in test_units.cpp, lines 188-199.  There  
I use a class called Measurement which does simple error propagation  
(just meant as a toy demonstration, not a full-fledged error  
propagating type).  Because the units library merely delegates  
comparisons to the underlying value type, if you implement a FuzzyFloat  
class which performs logical comparisons with fuzzy logic and then  
write SI<FuzzyFloat>::Length l1,l2 etc..., the result of the code above  
will be consistent with those definitions...  I think it's important to  
understand that my units library performs dimensional analysis for  
arbitrary unit systems, with the SI, CGS, etc... systems being provided  
for convenience.  It doesn't do anything with the value type arithmetic  
except delegate, as I believe it should.

Matthias

------------------------------------------------------------------------ 
---------------------------
Matthias Schabel, Ph.D.                           	
Utah Center for Advanced Imaging Research
729 Arapeen Drive
Salt Lake City, UT  84108
801-587-9413 (work)
801-585-3592 (fax)
801-706-5760 (cell)
801-484-0811 (home)
mschabel at ucair med utah edu