Re: [patch] First stab at text deltas

Jim Blandy wrote:
> Experimenting with windowing algorithms sounds like a lot of fun.
> Obviously, one wants to have the two windows slide along so as to have
> the greatest degree of resemblance between the source and target.
> Deltas for huge files with small local changes should still be small,
> but if your windowing is not ideal, they could still end up being
> proportional to the overall size of the file.

I tested this by comparing a Makefile with its Makefile.in. When the
window size is about 10% of either file, the delta is about 2.5 times
larger than the on created with a window larger than both files.
(Note that the delta sizes are just an approximation based on the
number and kind of ops in our internal representation; I'm not
creating vcdiffs yet.)

It's interesting to note that the larger delta was about 12% of the file
size, a bit larger than a `diff -n`. Of course, this isn't really a
representative test, anyway.

In any case, we'll be able to play around at least with window sizes
now, at least.

    Brane

-- 
Branko Čibej                 <branko.cibej@hermes.si>
HERMES SoftLab, Litijska 51, 1000 Ljubljana, Slovenia
voice: (+386 1) 586 53 49     fax: (+386 1) 586 52 70