And from
January 1987
's editorial are the rest of Phil Lemmons' comments on the eventual move from 16 bits to 32 bits:
Editorial
Two Brief Conversations with Ben Rosen
by Phil Lemmons, Editor in Chief
With the arrival of the 68020 and the 80386, we've heard a lot of people saying most users don't need the power of a 32-bit processor. We've heard this sort of thing before. Back in the mid 1970s, some technically astute people said things like, "I wouldn't know what to do with 16K bytes if I had it," and "You ought to be able to do anything in 8K bytes." Programs like VisiCalc and WordStar made everyone realize that 64K bytes of RAM has its uses.
When 16-bit processors arrived in the early 1980s, more than a few experts said, "A 780 and 64K bytes can do anything anybody needs." But Lotus 1-2-3, Framework, Javelin, Reflex, Paradox, Q&A, and dozens of other products have shown that an 8088 and 640K bytes can do things an 8-bit processor can't.
Since the introduction of the Macintosh, and especially since the Macintosh Plus gave that 68000-based system adequate memory and mass storage for its graphics environment, software developers have given us glimpses of what 16-bit processors with a large li
near address space can do. Can you imagine products like Excel, STELLA, Balance of Bower, PageMaker, and More running on a Z80 with 64K bytes?
But how can we be sure that the move from 16 bits to 32 will be as important as the move from 8 bits to 16? As it turns out, Ben Rosen is an interesting gauge of this.
Ben is not easily impressed. His venture capital firm, Sevin-Rosen Management, has been involved in some of the most successful start-up companies in personal computing, including Lotus, Compaq, and Ansa. He has seen a million proposals from start-up companies.
At the Personal Computer Forum in Phoenix in early 1986, I ran into Ben during a coffee break. He was outside on the patio, thinking aloud about the talk he was to give the next day. Many people from the software community were questioning whether there was room in the market for any new software products, no matter how good. I asked Ben what he would do if Mitch Kapor walked into Sevin-Rosen seeking funding then, in early 1986,
rather than years earlier. "That's an interesting question," Ben said with a smile. He seemed intrigued with the prospect of a product like 1-2-3 being turned down. This was a depressing commentary on the opportunities in software in the heyday of the IBM PC AT.
I bumped into Ben Rosen again last November at COMDEX in Las Vegas. We were in the Quarterdeck booth watching software run under Quarterdeck's DESQview environment on a Compaq 386. The machine was simultaneously running a desktop publishing program under GEM, a CAD program under Windows, and another application written for TopView.
Ben said, "We're beginning to see proposals for new application software products for the 80386. There are going to be some very nice things written for that chip."
"AI?" I asked.
"That, and other things," he said. He refused to be drawn out further.
With the 32-bit processors, as with the 16-bit processors before them, we will all find we need them as soon as software developers have had time
to exploit the new chips. How will developers break new ground? There are many possibilities: in the graphical user interface, in natural language, in communications, and so on.
But developers
will
break important new ground. When a seasoned venture capitalist like Ben Rosen finds proposals interesting again and attributes this to the 80386, we can all be sure that the 32-bit processors will make possible some improvements in software much more dramatic than the great increases in speed that have already been observed. You better begin to budget now for a machine that will run the software that's coming.
photo_link (68 Kbytes)
10 Years Ago in BYTE
