Subject: Re: [boost] GGL Review
From: Barend Gehrels (barend_at_[hidden])
Date: 2009-11-17 10:51:42
> and has already suggested a battery of tests. I think this would be an
> excellent place to start.
Hi Brandon, some questions on this:
- did you look here:
http://geometrylibrary.geodan.nl/formal_review/sets.html Probably yes
but I want to be sure that you read it
- did you look in the tests? I don't think we need a "place to start"
because we already have many tests on this, in the distribution, and
more. I mean by this that we're started, on the way. But thanks for
pointing this out.
- clearly we didn't cover *all* cases, because of the bug in the (new)
assemble process, but we know that a lot of tests are necessary and
therefore we have a battery of them
- personally I don't think that batteries of random data will cover *all
*cases. I understand that exceptions or endless loops might become
clear. But with random it is difficult to verify the results. We can ask
Luke of course, how he verifies them, and I've no problem to add them then.
- it is still not clear to me if you refer to "algorithmic robustness"
(handling all cases) or "100% numeric stability" (still being correct in
the ranges where float or double cannot be expressed anymore), see also
our last post about this
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk