>>I think the big change will be a move to RISC architecture (which ARM has). Power and temperature are the drivers. Per watt and per therm, RISC has a large advantage. 10 years (Craig's estimate) seems right to me.
>
>That was a big prediction in the early 90s. Didn't pan out.
>
>It might this time.
>
>>
>>>Well, one of us will. I am convinced that it won't be me, unless a 128-bit or 256-bit commercially available Intel '86-style CPU is released.
>>>
>>>>You're going to be surprised. I'm giving it 10 years.
>>>>
>>>>>My guess is VFP9 will work with every new Intel version of Windows for the next 25-30 years.

Another big thing in the late 80s and early 90s was CASE tools. Remember that big wave? When I joined Kraft (Kraft General Foods at the time) we made a huge bet on CASE. They were good design tools but they couldn't quite get it to the code level. The grand redevelopment project turned into a steaming pile of poo. Some execs literally lost their jobs, which was not the usual way at Kraft. Usually failed project managers were busted down to some offensive title that told them it was time to seek greener pastures. That time heads rolled.

I was among the first to bail out. I returned to my big fish in a small pond, highly paid, VP role at a small software company based in downtown Chicago. And what do you know? They were on a CASE mission. The company was closely involved with IBM, which eventually bought them out, and IBM decreed that you are going to choose a CASE tool. One good thing about it was I went to visit an Atlanta vendor called KnowledgeWare. During the sales meeting the CEO, Fran Tarkenton, walked in to say hello. He wasn't a figurehead, he really ran the company. So I got to shake hands with a hall of famer.