Re: [boost] Sparse Bit Vectors

From: joel de guzman (isis-tech_at_[hidden])
Date: 2001-06-20 22:26:45

----- Original Message -----
From: "joel de guzman" <isis-tech_at_[hidden]>

> ----- Original Message -----
> From: "David Abrahams" <david.abrahams_at_[hidden]>
> >
> > ----- Original Message -----
> > From: "joel de guzman" <isis-tech_at_[hidden]>
> >
> > > -> Macintosh style graphical regions. A vector of sparse bit
> > >     vectors can represent something like Regions in the Mac.
> > >     in fact the Mac-Region inspired this implementation. [  I
> > >     am aware that this is patented. I don't know if my
> > >      implementation infringes on this patent. But I am also
> > >      aware that this patent will expire soon (now)? ]
> >
> > I think the Mac implementation uses a some kind of difference encoding
> from
> > row-to-row, so that, e.g., a gray halftone (checkerboard) is actually
very
> > cheap to encode. If you're not doing that, you probably don't infringe.
>
> Ok. I'm a bit aware that they do some row to row stuff. My implementation
> is just a 1D Region (In Mac parlance). Extending that to 2D is straight
> forward.
> Mac's optimisations seem interesting though. I'd assume that was patented
> pre 1983. So that makes the patent open just about now (?).
>
> >
> > > -> Compresses bitmaps. Ok, same as above :-)
> >
> > Or at least bitmasks...
>
> Compressed, drawable and hit-testable as is.
>
> >
> > > -> This is interesting... A set of floating point numbers
> > >     and ranges. Say we want a set of floating point numbers
> > >     from -1 to 1 except 0. You'd write:
> > >
> > >        fs = range<float>(-1.0,+1.0) - 0.0;
> >
> > Hmm, we should talk. There is a nearly-finished implementation of
> half-open
> > ranges in CVS. This work is at least related.
>
> Cool, I'm all ears. My implementation uses closed-ranges. But I guess
> it is trivial to use half-open ranges instead.

Oops, I'm not sure if that's what you mean by 'half-open' and 'closed'
ranges where 'closed' means the 'from' points to the start and the 'to'
points to the end (rather than 1 past the end). I was just guessing
from the context. But I guess you know what I mean. :-)

Cheers again,
Joel de Guzman