Larry explains...

> Using the POSIX namespace allows the names to be "database##.###"
> where the "##" is the dataset number (just like of old) and the ".###"
> is
> the jumbo file number.  The MPE namespace file "database##" is a
> control
> file which includes information such as how many ".###" files you
> should
> have.  The first file that contains data would be in the POSIX
> namespace as
> "database##.001", the second as "database##.002" and so on.
>
> This was debated my many :), because of the issue you bring it.  It
> forces
> jumbo set users to *have* to use the POSIX namespace.  However, in the
> end,
> I believe the choice was made because it was the best of the known
> options
> at the time (I'm not implying that there is another known option
> today,
> just that there could be one in the future).
>
> [Neil Harvey]  And I'm sure (hoping?) that in future, MPE will be able
> to handle files that are terabytes long, all within MPE space :)
>
>
> Neil
>