historical note (was: New Block Closures)

Jarvis, Robert P. (Contingent) Jarvisb at timken.com
Fri Jun 22 18:28:16 UTC 2001


> From: Michael Rueger [mailto:m.rueger at ACM.ORG]
> We seem to have become worse at solving soft problems (like morphic),
> maybe because we are not really solving them, just adding more and mre
> patches to our incomplete understanding of how things should work.

Well, where were we 20 years ago:

	Systems:   Most programming was done on large, monolithic systems.
			Transaction processing (ala CICS) was the hot
technology.
			16 megabytes of real memory was a *lot*.  Disk
drives
			were the size of washing machines.
	Language:  COBOL was the language of choice.
	Database:  Flat files and simple non-transactional indexed files
were
			common.  Hierarchical and CODASYL databases were
new;
			relational database were still "too new".
	Microsoft: Was a little compiler company in Arizona IIRC.
	Me:	     My "computer" was an HP-41C programmable calculator.
			Did some really neat stuff with it, though!

Where are we today:

	Systems:   Most new development takes place on small, networked
			computers.  (I say "small" here in terms of size and
			cost.  The PC I'm typing this on has more CPU and
memory
			horsepower than the mainframes which 20 years back I
was
			drooling over.  (Which was appropriate seeing as how
those
			old mainframes *were* water-cooled :-)).  16 megs of
			memory is nothing.  16 *gigabytes* of real memory,
now,
			that's a *lot*.  Disk drives fit in your hand.
	Language:  *My* language of choice is Smalltalk in all its variants.
			The "industry" language of choice is probably C,
with
			C++ and Java in the hunt.  Dinosaurs still program
in
			COBOL.
	Database:  Relational databases are common.  Object databases are
new.  I
			don't know what is "too new" these days.
	Microsoft: Is a multi-gazillion dollar cyber-Godzilla in Washington.
	Me:	     1.2 GHz Pentium 4 (good) running Windows Me (bad, but
			needed to run games for the kids).  I hardly get to
			look at it, let alone do any programming.

I think that the problems we're trying to solve today are much more complex
than the problems we dealt with 20 years ago.  (I can't remember the last
time I had to cook up some master-file update logic).  

Where are we going (2021):

	Systems:   Don't ask me.  Maybe multiprocessing systems (one
process,
			one processor)?  World-wide wireless networking?  16
			gigs of memory won't be enough to do anything
useful.
			16 *terabytes* of memory will be a lot.  Rotating
			disks will have vanished, replaced by multi-terabyte
			RAMsticks or something.
	Language:  Beats me.  I hope things will change/improve.  I fear
			they won't.  Dinosaurs will still use COBOL.
	Database:  Ditto.
	Microsoft: Will go bankrupt in 2013 when everyone "Just Says No" to
			rebooting their PC's every five minutes, at $100 per
			reboot.  Bill Gates becomes hunted fugitive; when
			surrounded by disgruntled investors at their
corporate
			HQ in Antarctica he reboots himself all the way to
Mars,
			thus escaping prosecution *and* opening up an entire
			new market for Windows ET (Extra-Terrestrial
edition).
			This causes the financial markets to melt down, but
the
			government saves the day by using the Social
Security
			trust fund to bail out Wall Street.  ("Which is more
			important", then-president Newt Ging-grinch will
ask,
			"giving money to a bunch of whiney old people who'll
			soon be dead anyways, or giving money to a bunch of
			whiney rich people who will give a bunch of it BACK
			in political contributions!?!?").
	Me:	     I'll be within a year of retirement, which I probably
			won't be able to afford (even without having counted
			on Social Insecurity), the last of my kids should
			be in college, and I might finally have time to do
			some programming for fun.  Does Morphic still need
to
			be refactored?  :-)

Bob Jarvis
Compuware @ Timken





More information about the Squeak-dev mailing list