Absolute horseshit. Really. Exactly what "best work in software" is being done by established companies? Practically every innovation in the software technology that really matters has come out of the startup/non-corporate-open-source or academic worlds.
It's fascinating that people are so jealous and insecure, apparently about the success and the moxie of startup folks, that they misread the referenced article. It's like an ink blot. Absolutely amazing..
A lot have actually come out of large corporate research labs, and then been commercialized by startups when the big companies ignored them. Xerox gave us the mouse, the laptop, the GUI, color graphics, Smalltalk, the WYSIWYG word processor, Ethernet, and PostScript. Bell Labs gave us the transistor, sound in movies, six-sigma, the television, photovoltaics, algorithmic information theory, UNIX, C, C++, and plan9 (which is full of innovations that haven't yet been commercialized). IBM gave us the relational database. Microsoft is funding much of the work on functional programming.
You really need both. Startups are an essential part of the economy, but they're not the only part of the economy, and many brilliant inventions have been discovered by researchers working 9-5 at a big company.
No, Doug Engelbart gave us the mouse. You may wish to go through the rest of your list and fact-check it. There's a nice on-line encyclopedia you can use at en.wikipedia.org/wiki/ ; I forget what large corporation invented wikis and later funded the application of them to encyclopedias.
Let me help you with that: Apple. The c2 wiki was inspired by Hypercard.
Speaking of Rorshach Tests. Your example of a CS advance isn't. Wikipedia is an application of a PHP script designed 5 years before the site launched. But you probably wrote that comment using a CPU that is the product of N generations of CPU research at Intel corporation.
Let me help you with that: Apple. The c2 wiki was inspired by Hypercard.
Ward Cunningham invented wikis. Apple didn't. The c2 wiki was inspired by an app he had written in Hypercard, which was itself inspired by Hypercard. But inspiration is not invention.
As fabulously successful and widespread as wikis have become, Ward's name is rarely mentioned in connection with them. That strikes me as a shame, because this is one case where there's a clear and unambiguous inventor.
I started to object to this with a reductionist argument about what wikis actually are, but thought the better of it and concede the point. Ward's wiki is a good example of an indie CS advance.
Yes the wiki was invented by an "uncaged lion", but it was built on over 3 decades of advances by employees of large organizations: Andreeson was an employee of NCSA at UIUC or some other university when he wrote Mosiac. Tim Berners-Lee created the WWW while employed at the big European organization for high-energy physics (CERN). The TCP/IP stack was first added to Unix by grad students and faculty at Cal Berkeley, funded by the US government. Almost all pre-WWW internet software was written by employees of large organizations. E.g., the traceroute utility was written by an employee of LLNL. Sendmail, a grad student at Cal. Earlier mail servers and user agents were written at RAND Corp (which is only medium-sized but is funded almost entirely by the government). The first mailing lists, SF-Lovers and Human-Nets, were started by employees of large organizations and populated almost entirely by them for the first two decades of their existence. The end-to-end argument and the notion of the IP layer were formulated in the early 1980s by employees of large organizations. The father of the internet, J.C.R. Licklider, worked for large organizations his entire life (MIT, Dept of Defense). The first internet hosts were developed by medium-sized defense contractors: BBN was one.
The first entrepreneurs to significantly influence the evolution of the internet were probably the founders of SUN, most of whom came from Cal Berkeley and worked on those government contracts to add a TCP/IP stack to Unix. (The main contribution of SUN to the internet that I know of was to accelerate the number of internet users: every SUN workstation came bundled with internet software -- probably the first time that a marketing department helped drive internet adoption.) After the founding of SUN, not much entrepreneurial influence on the internet that I know of till the internet gold rush starting 1993 or 1994, which was 33 or 34 years after (planning and research) work started on the internet/Arpanet. I know UUNET was an entrepreneurial venture of the 1980s: I am unfamiliar with whether or how UUNET influenced the internet though. In 1992, entrepreneurs/lawyers Canter and Siegal invented spam, but spam is hardly a distinguished contribution.
Also, Engelbart worked for TRW or some other defense contractor (a large organization) and then SRI (medium-sized but funded almost entirely by the government) so pointing out that he was never an employee of Xerox does not exactly score a point for the uncaged-lions side of the ledger.
Funny you mention that, because that list came straight off of Wikipedia. http://en.wikipedia.org/wiki/Xerox_PARC#Accomplishments. The mouse was invented by Engelbart, but Xerox PARC was the first research group to really adopt it and see what it could do.
Hmmmm. Pretty sure unix started as an unfunded side project for a couple of lads. From memory, research on transistors began well before Bell Labs had a hand in their development, IBM didn't give us the relational database; the honor belongs to a university I've long since forgotten. Oh, and Microsoft is funding some of the work on functional programming.
Hmm. Pretty sure Unix is universally credited to Bell Labs, originated in another operating system built by Bell Labs, and received substantial funding from Bell Labs. If your best argument is that Unix is an example of non-corporate research because Richie started it in his spare time, you don't have much of a case.
Remove "academia", which is a feeder for both startups and industry --- disproportionately favoring industry, of course.
Now, defend your argument. Maybe we should start with processor cores? I hear there's some open source HDL you can download for that. Or how about routing algorithms?
It's fascinating that people are so jealous and insecure, apparently about the success and the moxie of startup folks, that they misread the referenced article. It's like an ink blot. Absolutely amazing..