In the mid-to-late 70's I took part in a brainstorming meeting in the computing department of a British University. They were trying to decide which directions the departmental research should take. It was an interesting time for a computer science department then, because British Universities around that time were having to deal with threatened cuts on the one hand, but computer science was seen as a future growth area. The head of this particular department was in the interesting position of being required to write two reports for the University management – one on how to reduce the size of the department, which every department had to produce, and the other on how to enlarge the department, which only he had to produce.
You have to mentally take yourself back to remember what the computer culture was like then: no personal computers, Unix just emerging, little-to-no penetration of computers into daily life (lots of people I knew then were saying they would never use a computer), no CDs, digital watches just appearing, pocket calculators still a relatively new concept. The department was very proud of its new computer with 1/4 Mbytes of memory, and 10 Mbytes of disk, running Unix 6.
I had done my homework for the brainstorming meeting, and I knew about Moore's laws, and what that meant about likely available technology in the near future. When asked for suggested research topics, I suggested digital sound. An engineer present blinked and did a back of the envelope about how many bits per second that would entail. "Digital Photography" I suggested (amongst several other things). But the problem was, the people present didn't really believe that memory would ever get that affordable, nor that computers would ever get that fast. They knew that computers had got faster, and that memory had got cheaper, but, as they explained to me patiently, technology has its limits, and it looked like those limits were close to being reached.
Now, in the era of computing power too cheap to meter, we know better. In 1983 a 10 Mbyte hard disk for an IBM PC cost something like $3000; now, a disk 100 times more capacious (and many times smaller, and many times faster) costs 10 times less. Digital sound in fact happened earlier than I had expected then, principally because CDs allowed larger amounts of storage than otherwise would have been possible, and digital photography on the other hand has happened later than I had expected then.
At lunch recently I was talking to some colleagues about how the cost-per-bit of hard disks and RAM, though both going down, was dropping faster for RAM than for hard disks, and the lines would meet in about the year 2015, and so presumably there would be no reason to use hard disks any more after that date, since it would be just as cheap to use RAM. I then remarked on how this might affect how we use computers and how the interfaces to programs might change. Their response was "I wish I had a crystal ball like yours", and they then started to explain to me how technology has its limits, and that it looks like they've nearly been reached now...
I hadn't even considered this to be a particularly shocking prediction (I hadn't actually considered it a prediction): we've already reached the age where hard disks sometimes get left out: the PDA in my jacket pocket, which by the way is twice as powerful and has twice as much memory as a machine I used for serious computing less than 10 years ago, stores all its data in 1Mbyte of (battery protected) RAM. The portable computer I use has twice as much RAM as the hard disk I had on my portable-before-last.
Even short-term predictions seem to surprise people. In the age of the Commodore 64 (with 64K RAM), I once gave a talk to a group of hobbyists, and managed to cause great hilarity by telling them that the system I was describing (intended for personal computers) would never run on a machine with less than 128 Kbytes. Only a couple of short years later (in the age of the Atari ST) I was able to produce almost equal hilarity by relating this self-same story, but then caused outrage by telling them that the system that I was describing at that meeting ran acceptably fast on a Sun 3, and had to calm them down by explaining that they too would have machines that fast in the very near future.
Let me ask you a question: in the late 80's people joked that one day we'd all have Crays on our desks. When will we actually reach that point?
In fact, that time has long gone. The portable I am writing this on is about 4 Crays-worth of computational power, and my desktop is about twice that.
We have the good fortune to work in a relatively predictable field. If we don't take advantage of this, we run the risk of making the same old mistakes. Even now we're still living with the legacy of the person who thought that we'd never need more than 640 Kbytes of RAM.
© Copyright Steven Pemberton, Amsterdam, 1997. All rights reserved.
First published in the SIGCHI Bulletin, January 1997