Re: [boost] [Review] Quantitative units review results

From: John Phillips (phillips_at_[hidden])
Date: 2007-04-11 11:12:33

Thorsten Ottosen wrote:

> John Phillips skrev:
> 
>>I’m happy to report that the quantitative units library developed by
>>Matthias Schabel and Steven Watanabe has been accepted for inclusion in
>>boost.
> 
> 
> This is of course good news.
> 
> However, I would like to know what the main difference between this 
> submission and the one previously rejected.
> 
> Has the reasons for rejecting the first library been solved in the new
> submission? Furthermore, what are the main differences between the two 
> submission?
> 
> Thanks
> 
> -Thorsten

   The main difference is probably that Matthais and Steven were not 
trying to be all things to all people in the realm of units. Andy's 
library tries to provide compile time, runtime and I/O facilities, all 
in one package. This makes his library quite complicated, makes zero 
runtime overhead hard to approach and made it so his library 
documentation promised many things as forthcoming features that did not 
exist. Things like adding new units or systems are quite hard in Andy's 
system.

   If you recall the discussion from Andy's submission, there were a 
number of other points involved, but I think these are the most 
important ones. This library solves some of the problems by not trying 
to be all answers at once. By doing so, the issues of I/O format, 
documentation complexity and difficulty of extension are all gone. This 
may not turn out to be a complete enough answer for some potential 
users, but the opinions of the reviewers (many, but not all of whom also 
contributed in the PQS review) are largely that this is a better way to 
go about it than the PQS approach.

   Does that help answer your question?

                        John