27 Sep
2000
27 Sep
'00
11:20 a.m.
On Wed, 27 Sep 2000, Clint Wolff (VAX collector) wrote:
On Wed, 27 Sep 2000, Jerome Fine wrote:
Hmmm! Seems like I remember something about punched card input to tube
type computers in that era. Wouldn't that qualify?
- don
I think that it would be helpful to have a discussion on the different
phases of the "computer" market with a view to identifying the
characteristics to be expected for the next phase. Based on
40 years of seeing the market charge, I have a bit of perspective,
but I am probably so out of touch with the current market that
I don't have the needed insight to identify the next phase. So, how
about some comments?
Phase I - Characterized by very high hardware costs for tube type
computers and a total lack of operating systems - up to about 1960?
Dominated by IBM
Not really computers IMHO (no flames please). These systems were
wired up with patch cables to perform specific calculations, then
rewired for the next. Although the patch cables can be considered
a form of 'stored program', it's just not the same. Phase II -
Characterized by the first introduction of transistors,
many more small companies along with the emergence of Snow
White (IBM) and the Seven Dwarfs (Burroughs, DEC, UNIVAC,
CDC, Wang, ?, ?). Some operating systems started to emerge -
up to about 1975 - Still dominated by IBM
Operating systems weren't available until memory became small
and cheap enough to give up large portion to a resident I/O
library which included functions that weren't needed by the
currently operating program. Anyone have an example?
Phase III - Characterized by the start of LSI.
Hardware costs were
still high and operating systems were mostly bundled. Volume was
still low in most cases - up to about 1985 - IBM was still the major
player, but no longer so dominant
MSDOS wasn't available separate from the hardware until much later.
I remember running a copied version of DOS on a brand new '286 clone
I built from parts purchased from Jameco. I searched high and low,
and couldn't find MSDOS unbundled anywhere. This was about a year or
so after the '286 was released (early '90s? I don't remember)
Phase IV - Here I start to become less able to
see the market. I know
that M$ started to become important along with Intel. In addition,
the internet was just starting to become useful and wide spread although
nowhere near the 2000 situation. The marketing model was about to
shift from low volume/high mark-up to high volume/low mark-up as
MS-DOS started to penetrate and the PC stated to become a
commodity item. IBM become less important. Can someone else
help to define this Phase IV? - up to about 1993???????
Hmmm... I think you've blurred the lines a bit too much here. The PC
market with MS software became mainstream in the mid to late 80's.
IIRC, that is when PC's (and to a much lesser extent Macs) became
standard equipment on people's desktops. The internet was being used
by universities since the early 80's, and the basic tools were in place
ftp,mail,archie,gopher,etc. (anybody remember veronica?)
The WWW wasn't 'invented' until the mid-90s. My first exposure
was NCSA-mosaic in '94, prior to the founding of Netscape by
the students who wrote mosaic.
Phase V - The internet starts to become very
important. Linux
becomes available. How do these two aspect interrelate? What
are the dominant features of Phase V? Will Phase V continue
for a few more years?
Linux 'is' because of the internet (but not vice-versa). Very few
people could afford the costs to copy an operating systems onto
floppies and nobody had a 9-track tape attached to their PC
(well almost nobody, I hung around some odd characters).
The wide availablility of the internet made it cheap and easy
to download.
Also, most people forget Linus T. only wrote the Linux kernel
(copied portions from Minix IIRC). The remainder is the Free
Software Foundation clones of the AT&T Unix tools (gcc, etc).
Linux wouldn't exist without these tools also.
Phase VI - Here is the key aspect of why I am
sending this email!!
Is it possible by looking at the past 50 years to identify the key
characteristics of Phase VI (which are probably already present)
which will take us up to 2010? Notice that I think that the number
of years in a given Phase seems to decrease. Is that correct?
If possible, the goal of this discussion would be to achieve some
sort of overall agreement on the different past phases (not really
too important) and some sort of forecast on what the next phase(s)
will consist of. If even the probable characteristics of the next
phase can be identified, that would seem to be a major achievement.
My stake in the ground:
1) Very few On-Demand services (music, video, etc). The current
bandwidth has trouble supporting the limited number of users running
'conventional' music and video feeds (everyone gets the same feed).
It will not be possible to supply everyone with their own selection
of audio and video feeds. You might have noticed there aren't
commercials (AT&T?) advertising movies on demand anymore.
2) Continued net congestion, with the creation of more private
high speed networks. As network bandwidth increases, more users will
jump on, reducing the available bandwidth to the current level of
pain. This is self regulating. As the network gets slower, fewer
people use it, speeding it up. Companies that can afford it will
use private high-speed networks for internal communication. This
is going to be a huge growth area, but will plateau relativly
quickly since the number of companies that can afford such a service
is small.
3) A continued move away from standalone PCs, toward the net-pliance
system. Based on the failure of DivX, I don't think people will
rent software, so the appliance needs to be programmable and upgradable,
so a secure facility for selling software across the internet needs
to be developed. Perhaps software manufactures should give away
CD-ROMs with their latest wares in demo mode, and sell the keys to
unlock varying levels of functionality.
4) A move back to the glass walled computer room. IT departments
are beginning to realize the cost of upgrading everyone's desktop
machines every few years. The current processor speeds are adequate
to run MS Word, and normal users don't need to upgrade (although
they want to). With the growth of high speed LANs, most companies
have moved back to a central RAID disk storage system, and have
started moving toward the 'compute farm'. Now each user has a local
(low power) system that is used to read email and write memos.
Computational problems are submitted to the racks of high power
machines for crunching. It is a lot easier and cheaper to upgrade
20 machines in a rack than 200 machines in peoples offices.
I suspect that while this topic is probably OT,
it is also a rather
interesting question and for most list members could be considered
vital for the future.
Somehow, when I read the comment that Allison made, I started to
think about these aspects and I began to wonder just what might
be in store. Being stuck (by choice) in RT-11 and the PDP-11
does not provide a reasonable viewpoint, but I suspect that the
younger members of this list might benefit from a bit of hindsight
and the older members could be helped by the younger members
in being able to identify future trends. Does this make any sense?
I am fascinated with older machines are OSs because I can understand
them. I doubt anyone understands EVERY function in MS Windows or
the Intel Pentium processor. The older systems are much simpler
in implementation, and more elegant in design.
Sincerely yours,
Jerome Fine
I'm all for identifying future trends. It's the only way I'm going
to get rich :) unless I win the lotto...
clint