Re: Modernization and multi-member files

On 18/07/2008, at 12:32 AM, Terrence Enger wrote:

Indeed, I remember hearing a presentation by an IBMer where she said
that multi-member files were designed explicitly for source code. It
was an unpleasant surprise that customers used them for data. Now, the
system as a whole is stuck with the situation.

Don't think so. If it were the case that the DESIGN was for source code only then part of that design would have included ensuring ADDPFM checked the type of the target file and rejected non-source files.

Further, CRTPF would also have been constrained to disallow MAXMBRS > 1 when FILETYPE(*DATA) was specified.

Since neither of those things are true (nor to my memory were ever true) I doubt the veracity of the statement that multi-member files were never designed for data.

There is nothing wrong with multi-member files as a solution to a problem. Not very relational but so what? That's not the only way to design a database. Multiple members can be handled quite well using OVRDBF and USROPN. Even SQL copes with them via OVRDBF or CREATE ALIAS, and multiple members are how Rochester chose to implement partitioned data files so they can't be all bad.

Regards,
Simon Coulter.
--------------------------------------------------------------------
FlyByNight Software OS/400, i5/OS Technical Specialists

http://www.flybynight.com.au/
Phone: +61 2 6657 8251 Mobile: +61 0411 091 400 /"\
Fax: +61 2 6657 8251 \ /
X
ASCII Ribbon campaign against HTML E-Mail / \
--------------------------------------------------------------------