That brought back memories for me, I worked there from 1978 - 1984. It was special and I agree with a lot of what the author said, but I would also highlight the people that ran the place. My boss was amazing, too long to list all the reasons but never forgot him. It takes a special type of person to manage highly skilled technical people effectively. I think there are some places like that today (maybe Google, I don't know first hand only from friends that work there). But in my view, that is the key, creating an environment that those types special people are able to thrive in. Bell Labs did that, maybe it was because they were a monopoly that they had that luxury, but I think there was more to it than that.
I called the guy who gave me my start in telecom up a good 4-5 years ago, and spent two hours on the phone with him thanking him for giving me a chance at it, without him I never would have made it as far.
> It takes a special type of person to manage highly skilled technical people effectively.
Can you please elaborate on this one? On the article it says at Bell Lab the manager had strong track record who appreciated scientific work - but is there something more?
I would very much like to know that too. I think Alan Kay said something about the middle management at Xerox being excellent people, too, so I would be tempted to think this is the one important factor to get the kind of tech oasis that were PARC and Bell Labs.
Some of us end up in the few dream-team environments that already exist today, but most will have to build those ourselves; any hint would be very welcome.
"At Bell, the financial motive was not grants but contributing to the company’s product lines. This seems reasonable to me, both because telephone
service is a public good and because, as Gertner notes, the challenges of improving phone service motivated technical advances that benefited other areas as well."
The thing is AT&T was the monopoly of telecommunication service in the US back then. When Paul Baran went to present his "packet switching" (which is also independently studied and proposed by UK's Donlad Davies, and both combined with Leonard Kleinrock's theoretical analysis became packet switching in the development of the "Internet") to AT&T, AT&T laughed it off and said no, mainly because AT&T did not want to adopt the digital communication and wanted to maintain its monopoly. The engineers at AT&T (not Bell Labs) pretty much said to Baran he lacked the fundamental understanding of how communication system works. Just to be fair, at work, in the beginning RAND Corp engineers were also not convinced, but nonetheless Baran was able to get enough support and appreciation from his fellow coworkers and the company's management that he demoed to AT&T.
This is a story of a classic corporate engineer who has the "know-it-all" mindset vs the classic researcher/engineer who appreciate challenges and differences.
I recommend reading "When Wizards Stay Up Late".
I think companies which allow some degree of hacker/research culture really do benefit from the openness and innovation. Google's 20% (although I read most engineers don't get any free time at all and 20% was not a formal policy, and 20% was just a spontaneous thing in the early stage of Google), FB's internal Hackathon crated some of the most-used features/products at the respective company. I think another modern example is Stripe, I consistently heard something great about its culture.
AT&T was responsible for much of the base technology that made packet switching even possible. They did however stuck (largely for good technological reasons) with circuit switching all the way thru, even if it was largely digital by 1984.
Even today there are considerations whether packet switching is really better than circuit switching in particular if reliability is a concern: For circuit switching one can (often) guarantee, say, the bandwidth of the service offered over wire, for packet switching this is between hard and impossible.
So while I can accept that for many applications packet switching is the better choice, I can understand AT&T at least with respect to why they still wanted to offer a circuit-switched service.
I subscribe to the school of thought that mindstate while In Flow is real and significant, so it is easy for me to believe this anecdote/quote. Mindstate scaled up to multi-year or even multi-decade research efforts likely also exists, and would matter across the entire team and not just individuals. If this is true, then breaking off activity mid-effort and restarting must exact a steep price in re-connecting, no matter how strenuously we try to mothball the effort.
One follow-up question I'd like to see asked: how did the leadership at say, Xerox PARC and Bell Labs decide relatively consistently how to pick and choose which efforts to consistently fund? What markers were they looking for, what process did they use, or what were their thought processes? How did they decide to pull the plug on specific lines of inquiry (surely they didn't always pick all winning hands, so to speak)?
I think having a fountain of money is certainly a good thing, but other companies today enjoy consistent funding and they're nothing like Bell Labs.
For example, Verizon and Comcast. Both rake in vast sums, but where is the fundamental research and contributions?? It appears that mostly gets folded into boring vertical business initiatives.
I suspect a key difference is that the old Bell Labs wasn't beholden to excessive Professional Project Management running the show and making sure everyone has their nose to the grindstone at all times.
I suspect a key difference is that the old Bell Labs wasn't beholden to excessive Professional Project Management running the show and making sure everyone has their nose to the grindstone at all times.
I think that's true, and was also true to a greater or lesser extent in many of the other hotbeds of 20th century research -- perhaps including those embedded in more conventional companies (IBM Research and Xerox PARC both come to mind).
Short of being wealth enough to self-fund, are there any truly low-managerialism research places left today?
AT&T was a legal monopoly, so they were bound by law to have research and I think a certain percentage of profits had to go to a development wing (Bell Labs) focused on improving technology. Neither Verizon nor Comcast have any such obligation.
This is a little off topic but maybe it highlights the sort of people that worked there.
When I was an undergrad I had some Unix question so I emailed Dennis Ritchie (...!research!dmr) and asked him about it, we went back and forth a bit. Finally I asked him for his phone number and he sent it, I called him and we quickly sorted out my question. Over the years, I've called him a number of times. I suspect that if he were still with us he'd remember me, he seemed to enjoy chatting and passing on his knowledge.
I had to think about to remember if it was just email or if we talked live. My phone database put that straight:
$ call ritchie
Dennis Ritchie 908-582-3770 (W/bell labs)
I was just a kid in college, nobody important, I hadn't contributed anything of significance to the field. Yet he was happy to take time out of his day and share his knowledge. I've had the same experience with bwk, I had some ideas I wanted to hack into awk, emailed him and told him what I was thinking, not too long later I open up my mailbox and there is a tarball of ~bwk/awk, all the source, the docs, and the book in english and french.
Bell Labs was something special, pretty much everyone that I've met/talked to who worked there were really nice people and more than happy to share their knowledge. They only cared that you wanted to learn (and that you had a clue). Smart (enough) guy who wants to learn more? Sure, here's my phone number.
BTW, there is mailing list for old Unix hackers that I enjoy, if you like that sort of thing you might check out http://www.tuhs.org/ - it's got former Bell Labs people who comment on stuff pretty regularly. It's pretty nerdy, there is a long thread going right now about old Unix vulnerabilities (eg, passing the right negative number to kill(2) would get you root) but I find it fascinating. Another thread that's going on is why C choose to pass arrays by reference but everything else by value; another is a discussion of C's type syntax (not a lot of love for the pointer binding to the variable, people wish that "char* p1, p2" got you two pointers, not a pointer and a char.)
The breakup of AT&T likely made the world we live in today possible, but it also killed the innovation engine that came up with much of the underlying technology that the world we live in today was built on.
No, AT&T did patent much of their inventions, they were prohibited from doing something other than phone service and related equipment, so they licensed or gave away their IP. They also had broad cross-licensing agreements with their competitors in the industry (ITT/Kellog and Automatic Electric). AT&T however was prohibited from selling computers until the 1956 consent decree was vacated in 1984.
Our local chapter of Papers We Love has been looking for articles sort of like this one but a little longer (maybe 3-10 pages), on either Bell Labs or Xerox PARC. Any suggestions?
I think that everyone should read The Idea Factory, for a reason similar to the reason that everyone should learn Lisp. Bell Labs no longer exists, and you will never get to work there; but you can learn how research and supervision could be done, and that will help you route around whatever damage keeps the handbrake on at the place where you do work.
Many years ago I recall seeing an image of the office layout where it showed Dennis Ritchie's office being located near Ken Thompson's office (I believe). I'm unable to locate it now, though.
That must've been an interesting work environment.
I had the pleasure of taking over Al Aho/Brian Kerhighan's office for a summer as an intern. Although, this was while Bell Labs was finishing dying -- both of the long term residents had moved on long ago.
For those interested in learning more about Bell Labs, I highly recommend "The Idea Factory". It's a history of Bell Labs, focusing both on the biographies of the engineers and also a meta story about how you create an organization that consistently produces impactful innovation.
Post divestiture i thought I was hot stuff with my cps and math degree. Only to find out I was the dumbest in my BL group.
I remember people's email was their first name. I had to use first initial and last name.
Eventually I moved to sales/marketing for consumer side. Consumer was fighting MCI and share loss, looking for help from the Labs.
They created something that virtually eliminated all background noise (Feldspar??) but it was not used/promoted afaik.
Then someone suggested that all AT&T long distance customers be given an emai account: their number @ att.com (2015551212@att.com eg) aol was still mailing 3.5 floppies
There was really no vehicle to speed up the product side....
