From: Maxim Shemanarev (mcseem_at_[hidden])
Date: 2002-05-05 22:50:10
> I ran into the clipping problem you describe some years ago. Because the
> graphics package didn't handle the integer overflow issues, we hand to
> pre-clip all inputs, which was a major pain in the whatever.
I had similar problems many years ago working on a CAD system on PDP-11.
> For a lot of digital geography applications, there is a serious issue of
> who, when, where, and what has the responsibility for conversion from
> to device coordinates.
Precisely. It's amazing that monster-like graphic libraries don't
allow to control it.
> Can coordinate conversions likewise be done at the
> point in the pipeline that makes the most sense for the application?
Yes. And you even don't have to write any special converters for that.
The proper use of the affine transformer included in a *proper* place of
the pipeline will give you all you need. You can even have more than one
transformer (say, before and after some converter) if you need.
The only reason it could be preferable to use a specialized converter
is to avoid 2 extra multiplication operatons per vertex.
I'm trying to propose an idea similar to the one used in BOOST
tokenizer library. But, I encounter some difficulties with using the
exact STL-like iterators paradigm (see my previous post).
Boost list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk