Re: [boost] Preview 3 of the Geometry Library

Subject: Re: [boost] Preview 3 of the Geometry Library
From: Brandon Kohn (blkohn_at_[hidden])
Date: 2008-10-27 12:38:34

--------------------------------------------------
From: "Bruno Lalande" <bruno.lalande_at_[hidden]>
Sent: Monday, October 27, 2008 11:11 AM
To: <boost_at_[hidden]>
Subject: Re: [boost] Preview 3 of the Geometry Library

>> In looking more at Brandon's code, I have to repeat my observation that 
>> he has done an impressive job of defining a generic geometry type system 
>> and using boost libraries to do it.  I think that if we make his accessor 
>> dimension agnostic (either runtime or compile time indexing) we have a 
>> winner.
>
> Indeed, it's a very interesting part of his library. Unfortunately we
> really have to find a consensus about the compile-time / runtime
> access before we can do anything together... Maybe I will open a new
> thread about this precise subject.
>
> Brandon if you read this, there's a typo at 3 places of your code:
> "dimenstion" instead of "dimension".
>

Sorry for the double reply, but I thought I should point out that my latest 
code-base also tackles the run-time/compile time access issue via 
specialization of the access types. I have a type indexed_access_traits 
which is specialized for coordinate_sequences (of which point and vector 
types specialize via coordinate_sequence_traits). The specialization of 
indexed_access_traits incorporates a flag which says whether the code should 
use run time access or compile time access for each particular 
coordinate_sequence type. If the user's point/vector types have either 
get<Index> or operator[] for the corresponding compile-time/run-time 
accessor.. then they can simply use a supplied macro to declare their point 
as a valid coordinate_sequence type... and all the indexed_access_traits 
work for free. I've even been playing around with operators for dot product 
and scalar arithmetic on vector types defined using this mechanism... so all 
in all pretty good stuff in my code base so far. I've also tacked the 
reference frame issue a bit.. providing mechanisms for users to use either 
implicit frame transforms or to proceed as if they are in complete control 
of the state at any given time.

Please do have a look. I would love to hear feedback from any who are 
interested.

Cheers,

Brandon

P.S.
(This is a preliminary site that a friend put together for me.... the text 
is a bit brash :D.... please ignore and go to the download link to find SVN 
access.)
http://gengeomalg.sourceforge.net/

> Regards
> Bruno
> _______________________________________________
> Unsubscribe & other changes: 
> http://lists.boost.org/mailman/listinfo.cgi/boost
>