Boost logo

Boost :

Subject: Re: [boost] [xint] Boost.XInt formal review
From: Vicente Botet (vicente.botet_at_[hidden])
Date: 2011-03-03 15:58:48


Hi,

This is not a review, just some thoughts.

The Big Int library I would like to have in Boost should:

* Separate between data representation and algorithms. Algorithms should
work on concepts not on representation.

* Provide specific data representation that is a model of fixed and
unlimited Big Int Concepts.

* Optional features as COW should not add anything when the user doesn’t
requires them.

* Make use of move semantics when available by using Boost.Move.

* The implementation should avoid virtual functions as no needed in this
domain. Instead the library should use the CRTP which is well adapted to
this kind of abstractions.

* The library should provide expression templates so the user can get the
best performances she can expect. Or at least show that is open to be
enhanced with expression templates with some basic examples.

* A clear documentation including a good tutorial, examples and a reference
document with something more than the C++ syntax.

* Proving that the user can build other features on top of it in a non
intrusive and efficient way.

* That doesn’t reinvent the wheel.
 

Best,
Vicente

--
View this message in context: http://boost.2283326.n4.nabble.com/xint-Boost-XInt-formal-review-tp3331338p3334186.html
Sent from the Boost - Dev mailing list archive at Nabble.com.

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