Boost logo

Boost :

From: Pavel Vozenilek (pavel_vozenilek_at_[hidden])
Date: 2003-12-21 01:24:28


"Joaquín Mª López Muñoz" <joaquin_at_[hidden]> wrote

> I've just uploaded release 8 of my indexed_set library at
> http://groups.yahoo.com/group/boost/files/indexed_set.zip
>

We would like to see feedback on this library - it may
become first of new-generation 'smarter' containers
so your input may help to shape the future ;-)

Right now its main functionality is to provide multiple views
into data set, each view being sorted differently.

There are few ideas floating around what more features
a 'smarter' container can have:

  - ability to exclude items from a view (e.g. you want to see
    only female employees),
  - ability to group items in data set (e.g. get employees grouped
    by age bracket and each group sorted by name),
  - ability to provide tree-like views for data set
  - optimizations as lazily built indexes

The main advantage is IMHO replacement of traditional
"many simple data structures + complex algorithms" with
"one powerfull data structure + simple algorithms".

Since data structures are usually easily reusable it may
result in higher programmer productivity.

/Pavel

PS: can some good soul try to run the tests under release mode
      with MSVC6.5 and report exit code? All Joaquín machines
      do OK while all my machines crash, having the same
      configuration :-o


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk