reply !.

[Detlef Morgenstern]

Hi Ravi,

Thanks for your reply. It seems important to me that we distinguish between
the different categories of compression (see my camcorder example).

There is a common understanding, that compression may be loss-less vs.
"lossy" (you will get identically what you had before the compression when
you decompress/inflate/transform/unpack vs. there will be bits which cannot
be restored the original way).

Typically, we will compress (pack, transform, project, ...) a set of data to
make it require less bandwidth or storage capacity. When we want to get
access to the original data, we must decompress (unpack, re-transform,
project, ...) the archive/key/vector set. This is OK in most cases.

For CasC purposes mere data compression will not do. Think of a large
machine code file which is compressed as "binary data". What you get as
compression result is much smaller but it won't run any longer on a
processor (each time you want it to run, you must unpack it first). What we
need instead, is a transformation procedure the result of which will be a
smaller but still functioning "machine code", which makes the machine behave
the same way it behaved with the original code. As far as I can see, there
is still no established name for this type of compression. And I think we
need not coin a special term, because it is in fact the procedure of
abstracting: in the de-compressed state we have lots of coexisting "facts"
and few "rules" whereas in a compressed state we have fewer facts and more
rules. Usually, a rule needs less bandwidth/storage capacity than a set of
facts - hence the compression.

As I understand it, the driving force for Gerry Wolff's CasC initiative is
not the goal to save bandwidth. He concludes the other way round. "If I can
compress it, it will be more abstract after the compression". Make a
technology of it, and you can solve a lot of problems computing has today.

Do you see chances applying the compression methods you work on to a
procedure like this? (I meant the three steps a), b) and c) being part of
any kind of compression.)

Regards,

Detlef