On Wed, Nov 24, 2021 at 1:43 PM Larry McVoy <lm@mcvoy.com> wrote:

> SMIT - just say no.

There were a number of things IBM did well with AIX so I'm not quite so
glib knocking everything from them.

But I agree that SMIT was the not so well thought out piece and never fully
understood why it ended up being such a bad example of systems SW. ... but
 .... I always suspected that SMIT was an example of what IT managers
running mainframe thought of UNIX.  The folks at IBM set out (and did) a
thorough and well thought out requirements study IBM style ... and then ...
they only talked to Mainframe IT folks, not people that actually had
experience in running UNIX in a production setting from their (like on a
BSD or Ultrix based Vax or SunOS - i.e. instead of talking to the folks
that came to a USENIX LISA, they talked to customers that came to a SHARE
meeting).  So they solved the wrong set of problems.  SMIT was a force fit
of UNIX to mainframe shop and never was quite right for either group.  I'm
not sure the IT folks really liked it much better than the UNIX folks, but
at least for them it used their terminology and their concepts (*e.g.* DASD
*vs.* DISK).

BTW:  HP, I thought had a similar issue and they did not really understand
the UNIX user.   DEC parts of so got it/parts did not.  Many DECies wanted
Ultrix == VMS (and really wanted Unix to go away since VMS and RXS were
really better in their hearts), but at least there were a ton of folks
inside of DEC doing the Unix work that 'got it.'

As I have said before, it was always interesting having all of them as
customers at LCC.  You got to see the good and bad of all the systems
vendors.