On Sun, May 16, 2010 at 5:53 AM, Paul Gilmartin <[email protected]> wrote:

> I mostly agree.  PACK/UNPACK were designed for a reliable repository
> such as local DASD.  I (and many of the followups) are targeting an
> anecdote.  But lacking other tools, one of the questions tech support
> asks early is, "How many bytes in the file?"

I think the purpose of PACK (when implemented in COPYFILE) was to
reduce the disk space occupied by the data. Simple run length encoding
was probably best you could with slow CPU and limited memory (all
before LZW, I presume). Transfer between systems was not an issue
because there was only one system to most folks. Since then we
discovered that it's helpful that PACK can protect record boundaries
on hostile transfer media.

CMS Pipelines does the pack/unpack to simplify reading and processing
packed files.That's extremely helpful when you process files that are
archived on disk in packed form.

Rob

Reply via email to