If we had known everything we now know about the SPARC project, we’d never have started it.” Incredible words spoken by Vinod Khosla, one of the engineers who architected Sun Microsystem’s revolutionary chip architecture a quarter of a century ago.
Khosla, now a venture capitalist whose investments help the dreams of other tech visionaries become a reality, has a point. And what it comes down to is simply this: where there’s a will, there’s a way.
Speaking at the 25th anniversary of SPARC’s launch, Khosla was emphatic that no matter how improbable something looked, it was worth attempting, even if others mocked the early attempts and doubted the idea’s ability to succeed.
After all, isn’t there the old myth that according to the laws of aerodynamics, bumblebees should be incapable of flight? Of course, bumblebees live in blissful ignorance of the laws of aerodynamics, and therefore, are not crippled by self-doubt every time they take to the air. Engineers facing skeptical investors and peers are not always so lucky.
“There are lots of risks we took that we shouldn’t prudently have taken,” Khosla said. “I say, take this as a model to go try it, go do it!”
All human progress depends on the unreasonable man, said Khosla, lamenting the lack of innovation in the modern chip space and imploring engineers to have the guts to go that one step further, beyond the logical, beyond the assumed possible.
“It can’t be that in 25 years there’s no need for an architecture change,” said Khosla, “I do urge people to go and innovate again and take larger, bolder risks.”
“We had engineering problems, the company was low on cash, we had no tools, few people, and we didn’t know it couldn’t be done,” added fellow founding engineer Bernard Lacroute. “Sometimes, when you don’t know something cannot be done, you can get it done.”
Indeed, as the founders reminisced, the theme came up over and over again. They had all been told multiple times that the idea made no business sense, and yet, they did it anyway.
“We knew the risks were enormous, but if you don’t take big bets, you’re never going to be great.”
Rather heroic in my opinion. A bright "sparc" to inspire the engineers of today to create the architectures of tomorrow. But that’s just my two cents. What do you think?
We've been flying around in tubes with wings that are powered by underslung turbofans for 50+ years now. The control surfaces and control interfaces on modern aircraft would be totally recognizable to aviators from 100 years ago.
Of course modern aerodynamics and materials science have advanced considerably since that time so innovative designs like Boeing/NASA's X-48c Blended-Wing-Body are now technically possible.
The question (as EREBUS implies) is are airports, airlines and passengers willing to take the plunge? It could be that the X-48c is one of those technologies that's always "15 - 20 year in the future."
SPARC was never that great of a design and was subsequently crippled by hacks that made it fast for the 80's but hurt its performance long term.
I am not sure what he means by innovation. The SPARC was not a great innovation, it was a design choice for increasing the speed of computing. As other chips got faster, SPARC could no longer offer a competitive product and they became overtaken by normal speed progressions of hardware.
Having a technological advantage is wasted if the user cannot differentiate the value. Being better and faster is no longer a major selection requirement, especially when your networks cannot keep pace with even modest speed devices.
Just my opinion.
LTE smartphones will ultimately offer comparable to WiFi and wired connection speed.
This will surely accelerate our entry into an entirely different world -- the "mobile cloud" era - with many new services and - massive technological innovations at component (high frequency, packaging) and system levels
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.