Subject: Re: [boost] generative geometry algorithms library idea
From: Simonson, Lucanus J (lucanus.j.simonson_at_[hidden])
Date: 2008-10-05 00:14:21
>The only people who have
>responded to my request for feedback seem to be other geometry library
>writers, and from those responses, I didn't really get the sense that
>had really investigated what my library tries to do (I.e. all talk
>merging with their works with little feedback about the substance of my
I'm confused that you seem to be disappointed by the response you have
gotten. I skimmed the code and noticed several things I like about your
code. It was neat, had good comments, followed Hungarian notation
(actually I don't like Hungarian notation itself, but I like that you
have a disciplined and professional style), implemented valuable
algorithms that I am interested in, and used template techniques so
similar to my own that they even have some of the same names:
point_type, coordinate_type, coordinate_traits, etc. I emailed you off
list immediately, asking questions to help me investigate further into
your code, to inform you of the lay of the land in boost wrt. Barend's
library and to suggest we work together. After you didn't reply, I went
back to my own work and didn't bother to read your code further. If you
can't be bothered to talk to me, why should I review your code or offer
feedback? I guess we are both disappointed. Whatever expectation you
had when you submitted your library to the vault, I'm sorry that it
wasn't met, but you could have set that expectation more realistically
by looking at the list archive from the last several rounds of geometry
One tip, post code snippets to the list. Don't expect people to dig
through your vault submission to find out what you are doing. Very few
will. Package your design for the list in a concise email with examples
and explain it. You will get the feedback you wanted that way.
If you really want to talk to people on the boost list who are
interested in geometry other than Barend and myself, I can send you a
list of their email addresses (or you can look them up in the list
archive) and you can ask them directly to take a look at your library.
It is possible they weren't subscribed at the time or overlooked your
submission in the volume of unread boost email that can easily pile up
in an inbox.
>The notion of a concept driven traits based interface is a bit
>different than simply having users define their points in terms of a
>boost::point type. Same for the segments and polygons/lines. With my
>I'm really trying to define an interface where any user defined
>type can be used and the traits specialized on their type provide the
>necessary mechanisms for interaction with the algorithms in the
Please read the list archive, this is not a new idea around here.
Depending on when you had the idea you may have thought of it first, but
this is what we've already been discussing wrt. geometry type systems.
I'm optimistic that Barend's new interfaces will look like what you
describe and we can start moving toward a consensus. I also welcome the
fact that you are advocating for the same type of interface as I am,
because it makes it more likely that an eventual boost geometry library
will work the way you and I appear to agree that it should.
>I'm still trying to finalize some design issues and would like to
>that before writing documentation. Soon hopefully!
Please share these design issues with us. Perhaps the boost community
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk