9 Sep
2017
9 Sep
'17
5:55 p.m.
below...
On Sat, Sep 9, 2017 at 12:34 AM, Steve Johnson <scj(a)yaccman.com> wrote:
...
At the time, the IBM 360 required that you run a special job step to
create a file (we're talking punched cards here). And then you had to pull
that job step out of the deck because trying to create a file that already
existed was an error.
Point taken and sort of...TSS/360 and MTS was not (quite) as bad (although
it could be if you use the batch processing system), but Steve is right...
the idea of persistence was really not something people considered as
'easy' because it usually cost them money (real or allocated from the
computer center).
So I suspect part of it was the economics of storage at time. On line
(magnetic) storage was way more expensive than cards. Its has been
pointed at the the original PDP-7 Ken used did not have a disk on it, it
was custom special DEC's CSS group had splicing a PDP-15 disk to it. The
disk unit itself was manufactured by someone else (as were most/many of
DEC's disks for years).
Clay Christensen in "*The Innovator Dilemma" *has curves that actually
start a few years later when he studies the disk drive industry. But its
the just part of the same effect he is talking about.
The key is the what UNIX was doing was not considered practical by the
mainframe folks, so people did not consider it. It was a resources to be
protected (and to an extent, hoarded maybe). Moore's Law, et al, was the
engine that allowed the UNIX innovation to really see light. The way the
mainframes were doing things just did not make functional sense, but until
it was economical to do it otherwise - we were stuck.
As Steve says... it really was mind blowing for making things like this
easy.
Clem