home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.next.advocacy
- Path: sparky!uunet!spool.mu.edu!uwm.edu!rpi!usenet
- From: gad@eclipse.its.rpi.edu (Garance A. Drosehn)
- Subject: Re: A CPU on Every Desk?
- Message-ID: <b+p38fa@rpi.edu>
- Nntp-Posting-Host: eclipse.its.rpi.edu
- References: <1993Jan25.192117.2503@trilithon.mpk.ca.us>
- Date: Tue, 26 Jan 1993 00:46:20 GMT
- Lines: 87
-
- henry@trilithon.mpk.ca.us (Henry McGilton) writes:
- > As Steve Abell said, ``Oh, how short people's memories are!''.
- >
- > Anybody who (like me) suffered through that computer industry mass
- > psychosis of the 1970's called ``time sharing'' while trying to
- > get software out the door would never want to go back to centralised
- > anything. The biggest liberating factor I had in my career was
- > when I quit using a Control Data 7600 timeshared machine and got
- > myself a Cromemco Z2 box with a Z80 in it -- mine, all mine!
-
- The question you should ask is, "Were my problems the result of time-sharing
- per se, or were they particularly bad because of the OS that I happened to
- be stuck with to do that timesharing?"
-
- > Unfortunately, I see the disquieting trend back to time sharing,
- > except it's now called ``client server computing'', as if giving
- > it a new name changes the result. In these new bad old days, the
- > resource being competed for (competition, you understand, NOT
- > sharing) isn't so much CPU cycles as network bandwidth. The end
- > results are the same, though:
-
- One of my biggest complaints with "the brave new world of computing" is the
- lies that are involved. People tout the benefits of standalone systems, but
- what is really sold is "client-server" computing. I agree with you that
- "client-server" computing isn't all *that* much different than mainframe
- computing, once you peel away the fancy marketting layers.
-
- > o the decisions as to what's on your desk and how well (or
- > mostly badly) it'll be ``served'' are made by somebody
- > other than you. The person making the decisions usually
- > has no accountability and usually doesn't have to suffer
- > the same lousy service you're getting.
-
- Again, this shows a problem in your particular situation. System developers
- need to be on the *exact* same system as the users, or the problem you note
- is unavoidable. This is true whether we're talking mainframes or
- stand-alone PC's.
-
- > o the administration of the central resource ends up being
- > done by what rapidly becomes an elite priesthood of ``gurus''
- > and their acolytes. This elite priesthood rapidly develops
- > into a defensive faction with their own agenda which has
- > little or nothing to do with providing adequate service to
- > the users of the resource.
-
- That defensive mechanism comes from having too many people asking for too
- many things. This is unavoidable with any centralized system. Of course,
- with any completely distributed system, you end up with a similar problem.
- My stand-alone Mac has all kinds of gee-whiz, high brow, fancy productivity
- enhancements. I'll be willing to bet that most the Macs on campus don't,
- because *I* know where to find these things and the average Mac user doesn't
- have a clue about them.
-
- While I do believe there's a difference there, it's hard for me to put into
- words just what (or where) I think the overall advantage is.
-
- I'm also all too painfully aware of users who *think* they know what they
- are doing, so they install some new piece of software --- and then call us
- up (us = a central computing service) to figure out why nothing on their
- machine works anymore. I've lost many a day to such "productivity
- enhancements" as we have moved towards the distributed computing approach.
-
- > o the decisions as to the level of utilisation of the central
- > resource are made by bean counters, who usually don't have
- > to suffer from their decisions. The bean counter mentality
- > indicates that the resource must be utilised to and beyond
- > the limit, to the point where the resource becomes useless.
-
- The influence of bean counters on all this can be a problem. But again, it
- can also be a problem for standalone systems. Why buy a Mac Quadra when you
- only use that Mac Plus 2 hours per day? The real problem here is that you
- are using other people's money to get resources for your own work (work that
- they do not do). You have to convince them to spend the bucks.
-
- And since I work at a university, I am very familiar with this problem...
-
- > One would think that the people in the industry had learned their
- > lessons in the so called ``software crises'' of the 60's and 70's
- > and would not want to repeat the nonsense.
-
- It is not clear that everyone can really get what they need with any of the
- current models of computing, if that one model is followed slavishly.
-
- --
- Garance Alistair Drosehn = gad@eclipse.its.rpi.edu
- ITS Systems Programmer (handles NeXT-type mail)
- Rensselaer Polytechnic Institute; Troy NY USA
-