“John Mashey is known in computing circles for a whole raft of things, among them are his work on the design of the original MIPS architecture, his work at SGI, and a long history of in-depth posts in the newsgroup comp.arch. David Kanter of Realworldtech has taken one of Mashey’s posts and, with the author’s permission, fleshed it out with more data and graphs for posting as a multipart series. Part I of the series is now available, with Part II on the way. One of the things that really struck me in reading the retrospective was just how prominent a role completely non-architectural factors play in the stories of the successes and failures of various processors.”
This is a great article with some interesting history and insight into past business politics that some people newer to the industry (1998+) might not be aware of.
Indeed. Sad to see how computer company divisions sometimes work against their other divisions, as a monstrous multi-headed hydra with mps.
Good article based on a long drawn out thread on comp.arch NG many months ago by the old timers v JM, some of whom have less fondness for the high volume rubbish we are using today v what was once the pillar of computing hardware. Unfortunately c.a has gone down the toilet with all the homework posted.
I thought the article would go over why the Vax would not be practical in full custom today v the Alpha but that seems to be cut. The raw articles are still on the comp.arch NG just search for Mashey.
I remember working on my 1st processor design as a young lad using 1 Vax for 50 engineers knowing full well the chip that would come out would trounce the Vax when it was done and be 1000s cheaper. The writing was on the wall, in those days computing was still far from personal and 1 computer on every desk was still another decade.
just how prominent a role completely non-architectural factors play in the stories of the successes and failures of various processors.
This is the story of the history of computing. If technical factors mattered, we’d all be using micro-kernel OSs on top of RISC processors runnning a Smalltalk-based fully vector-GUI environment. Instead, we’re using 1970’s era monolithic kernels on a CISC architecture running a buggy, unsafe C environment.
Success in computing is far more about being in the right place at the right time than being the best. The people that win are people with products that are just acceptable and timed well.
That is a sad fact
Very true. What we use today is completely crap compared to what we could have.
It’s not so much that success is down to right place, right time, it’s more that people in decision making positions are incredibly conservative. We build on the old stuff over and over again, and are unwilling to throw out the cruft.
This is what disappoints me about Linux, Haiku, and the other open source OS initiatives. Instead of building something that’s genuinely new and innovative, or even building something that would have been high-tech for the 80’s folks are building 70’s era stuff.
More widely this is what makes me very uninspired about modern computing; it’s basically all the same as it was when I was a kid, just with fancier graphics.
It seems to me that the last true era of innovation in software was the late 70s at Xerox PARC. Everything since pales in comparison, and what we use today is a poor immitation, technically speaking, to what the folks at Xerox produced.
It should be no surprise that non-technical factors play a large role in regards to the success or failure of a technology. After all, the individual qualities of a particular technology, is only a small part of the equation.
Any IT-technology is just a part of a larger environment. It is often the battle of platforms, rather than the battle of individual technologies, that shapes the future of IT.
Just think about the Wintel platform. I would be hard pressed to say that Microsoft or Intel delivers individual technologies that are technically superior to everyone else. But they have fitted well together, and they have delivered a fairly cheap “standard” platform.
The platform-paradigm means that the sum is greater than the parts. Unfortunately it also means that we often are stuck with less than ideal technologies. And theres a lot of inertia to replacing platforms.
This is bad from a technology point of view. But from a business point of view, I think it makes a lot of sense. It is simply to expensive to reinvent the wheel every 5 years….