Boost logo

Boost :

Subject: Re: [boost] Fwd: Re: [Boost-users] [boost-users][QVM] Boost.QVM formal review is ongoing
From: Emil Dotchevski (emildotchevski_at_[hidden])
Date: 2015-12-14 21:59:24


On Mon, Dec 14, 2015 at 6:23 PM, Adam Wulkiewicz <adam.wulkiewicz_at_[hidden]>
wrote:

> It seems that you've send a response to the boost-users list and I'm
> guessing Emil is not a subscriber because he haven't responded. So I'm
> forwarding it to the developers mailing list.

I am a subscriber to boost-users but I've missed Paul's message, thanks for
the reminder!

> 2 Boost prefers clarity to curtness. The naming is much too terse making
> the resulting code rather
> 'write-only'. I'd at least go for col_mat for col_m and transpose for
> transp etc. (I know any
> changes will be a big PITA but shows that asking sooner than later is a
> good idea for would-be Boost
> library authors ;-)
>

I'm not concerned about the PITA part :) and I hear this message loud and
clear, transp, trans and perhaps a few other things will be renamed. That
said, in a library called QVM it should be clear what the _q, _v and _m
suffixes mean. I've found this naming convention very useful, there are
many QVM names that are consistent with it. I'd rather not change these.

> 4 Does this mesh at all with the quaternions in Boost.Math library?
>

This issue falls under the category of "it would be useful if QVM had
available bindings for this or that type", and I agree 100% with that. I
plan to add a directory qvm/bindings, where I'd be adding API-specific
headers defining QVM bindings for various types, e.g. OpenGL, DirectX,
Boost Math, Eigen, Maya, whatever.

Thanks!
Emil


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