Re: [boost] [Review] ITL

Subject: Re: [boost] [Review] ITL
From: Wuttke Peter (wuttke.peter_at_[hidden])
Date: 2010-02-26 15:02:10

To Begin with the most important point:
        YES, boost::ITL should be accepted as a boost library.
        
(Please be don't be to disappointed of my english, for i'm a developer
from Germany, and this would be my first review on Boost).
        
The ITL (interval template library) is  a very valueable part of our
(the Cortex Software GmbH, incidently the company i work for) daily
development routine.

In our industry (which is health care) sampling of Status-Informations
for  collections  of evaluations (in time)  are an essential part of
our work, so we are using the ITL-Library for some years with great
success. (We have to aggregate Events based on equivalent properties,
eg. same admission - which diagnosis, same discharge  - which
diagnosis, and so on. ) .

As far as I know, there is really no alternative for using  the ITL do
work with Intervals - I've tried is with plain SQL, and, belive me:
It's possible - but really no fun - no,no,no. (I've also seen
Experiments trying to implement the ITL-Solution for specific
problem-domains - they aren't usefull or elegant, either. Forget
them.)

Until now, we only used the library for intervals of datetimes (coded
as integers), but i Im   sure, that its usefullness goes far beyond.

Concerning the overall architecture, I think, the actual version is
more than complete, I'm especially impressed withn the idea of big,
sparse bitsets implmenented by the ITL - it's cool.

The documentation (using samples) is clear and completly comprehensive
- couldn't hope for more.

So, being not a specialist in template/generics programming (or in
algebraic specification), I LIKE this Library - and use it.