David Boggs has died
nytimes.com
nytimes.com
David and Ron Crane hired me for my first job in California right around 2000. I had no clue who they both were we just met at a trade show and I said I was looking for a job, we had this crazy interview where we just talked about anything and everything, and voila I was working in silicon valley. I think we spent about 2 hours talking about how we could float a balloon beside an AM radio station and how light up a light bulb...
Anyways, who fresh out of school doesn't work for the inventor of ethernet/fast ethernet/a core founder of 3com, etc? I only spent 2 years working with them, and I've always been looking to work with similar talent/kindness/etc. How they put up with a fresh grad I will never know.
I remember David and I working on an SDSL project, and we were just having the worst time ever. We couldn't get it train and finish setting up the link, it did almost everything but then just failed at the end. It was meant to be easy....but we just couldn't figure it out. We spent a crazy amount of time on it, maybe a month. Finally Ron got fed up, and asked "have you tried reversing the pairs?" and it worked! Turns out we had plugged the cable in backwards, and trying to streamline/debug the code had removed the final bits of cleanup code that checked if the pair was reversed. Ah well.
They were amazing mentors and friends. David invited me over to his home for wine tastings, to meet his cats (Palo and Alto), etc. I unfortunately lost touch with him over the years as I moved, he moved, etc.
Thank you David, you welcomed me to California and you'll be missed.
2000 would've been around the time of LAN media, no?
Ron and I kept in touch. (since he replied to emails) I lived overseas for a while, and when I was in the bay area we'd meet up, I remember visits with him to get breakfast at Black Bear Dinner, or touring the computer history museum together, a talk at parc. I so loved his excitement, he was so excited to meet up with me...made me feel special. I loved how he'd just chat/teach and that curiosity and patience.
Ok a story. So on this SDSL project we need to simulate 10,000 feet of cable, so Dave and I got some big boxes of cable and put them under our desk. But as the project progressed we need fancier setups. Again Ron pitched in, shook his head at out huge spool of cable and disappeared into the lab. About 30 minutes later he came back with this little mass of resistors, caps all soldered together in a ball and said something like "here, this is a good enough model of 10kfeet of cable with a 1k foot tap..." I just remember this little ball of parts...and having no clue how he just came up with it out of thin air. I think he explained it, but woosh. Analog, complicated, not 1s and 0s.
Or when I left....I gave notice and Ron was really nice. But he kept suggesting I should give 3 weeks notice. Just stay on for another week, etc. I was young, and I never listened well anyways so was firm with my date. Too bad...if I'd listened stayed 2 more days I would have collected my rather sizable retention bonus....ah well.
I ran into him once at the Computer History Museum in Mountain View a while back and he was willing to grab lunch with me and some friends and shoot the shit for a few hours. I can tell why he was a great mentor - approachable and knowledgeable.
Do you have an email address or a contact method?
His one conceit to vanity? The personalized plate on his older Mercedes SL coupe was something like ETHERNT, which is why I initially approached him.
You could then do “simple” vampire tap into such Ethernet RF cable to expand your network.
Because the then-10Base-5 signal is a guided RF-based wire, it was originally made to allow communication in one-direction only … at a time (half-duplex). Each end of the wire had to wait their turn (just like telegraph operators did by beaconing their intent to send info by pre-sending some well-known code).
Both IBM and Xerox Parc diverged into Token Bus and Star network configuration to deal with this “half-duplex” issues using the same collision backoff algorithm.
I think David unknowingly cemented the IEEE 802.3 standard as a winner when he prototyped the N-backoff algorithm for upcoming 802.3 for the 802 IEEE Working Group.
This half-duplex constraint of 802.3 (and 10Base-5) technology was losing out to the IBM token bus (IEEE 802.5) methodology in maximum bandwidth utilization.
Of course, without this N-Backoff algorithm, 802.3 wouldn’t have made the next step of leveraging a twisted pair (precursor to 802.3 10Base-T) possible to achieve this full-duplex and modern Ethernet we know to this day. Token bus and token ring both had effectively lost out in the Ethernet “Duplex War”.
While Bob Metcalf may have made it possible the 10Base-5 communications, he also got a lot of hardware help so David Bogg actually did the thinking of backoff algorithm … at hardware level. And the PARC performance team refined the backoff algorithm to near perfection.
Any error are mine and mine alone. I was close enough to that circle, others may be closer.
Respect!
I think the idea was that anything more than this was likely to lead to long enough round-trip times between devices at the far ends of the Ethernet that the CSMA/CD algorithm couldn't be guaranteed to work.
And it really did not scale. If you got too many computers on one Ethernet network together, things would start to really bog down.
We would address this by splitting an ethernet network into two. When it got too slow, we'd take our network of (say) 1.2.3.0 - 1.2.3.255 and split it into two networks, 1.2.3.0 - 1.2.3.127 and 1.2.3.128 - 1.2.3.255. Each of these ethernet networks would get its own separate port in the IP router. (Which could require adding cards to the router. And put extra processing load on it.)
DHCP did not exist yet, so we had to manually go to each computer and change its broadcast address, subnet mask, and default route.
This was disruptive, so we tried to keep all addresses in the lower portion of the address range so that they would continue to function well enough during such a transition. For example, a host at 1.2.3.10 with subnet mask 255.255.255.0, router address 1.2.3.1, and broadcast address 1.2.3.255 would still function when we changed its ethernet to be 1.2.3.0 - 1.2.3.127 instead of 1.2.3.0 - 1.2.3.255. Its broadcast address and netmask were wrong, but it wasn't completely offline. (By contrast, a host with address 1.2.3.200 would be out of range for the new network parameters and would have problems immediately.)
Then Grand Junction came up with the idea of ethernet switches to isolate traffic, and we didn't get one because they were a bajillion dollars. I don't know the exact prices, but I would guesstimate something like $5000 to $10000 for a switch with around 10 ports.
But eventually they came down in price. With an ethernet switch, everything was amazingly easy in comparison. It basically did the same thing I'm talking about (splitting an ethernet into pieces so the devices don't try to talk over each other), except instead of it being a weeks-long manual process, it happened automatically in real time.
Once switches became standard, ethernet was pretty scalable. Broadcast traffic still isn't scalable if it's used too much, but otherwise it basically works fine.
(Also, ethernet bridges were a thing at the time, and they are basically 2-port switches. But arguably a switch with a bunch of ports was a different way of thinking about how to do an Ethernet network.)
https://www.engadget.com/ethernet-co-inventor-david-boggs-di...
1) Sees someone struggling over some problem and asks if he can help.
2) Together, they get it working and it goes on to be used globally
3) ...
4) "And the Nobel Prize for xxx goes to ..."
Seems a bit understated.
I did not know him personally, but I was always grateful for his humility and grace while wearing the mantle great innovation.
Respect. Rest in peace.
It's true but kind of funny since ethernet has changed 180 degrees from the original design. I remember tapping thick cables with vampire taps when it was a shared medium with collision detect and now cables are plug and play with a star topology.
Someone told me about this crazy non-IBM protocol where you just shoved data into the wires. If it collided with someone else's data - just try again!
Thanks to a good helping of fud, there was no doubt in my mind that this was a rubbish approach that wouldn't scale (though obviously it was pretty handy not having to wire your network into a ring configuration).
Requiring shielded twisted pair cabling? Processor costs (when it mattered) and Ethernet's benefiting massively from cheap switches (when hardware costs had dropped)?
Token ring cost 5-6 times as much, and required a special MAU device (the thing that sat on the ring). Then the actual networked computers had a point to point connection to the MAU. Special cabling, not just twisted pair. IBM charged a lot for anyone else to licence the tech.
Token ring was clever and worked well though - a heavily used 10Mbps token ring network was far faster than the equivalent ethernet network due to no collisions.
Whereas ethernet degraded more gracefully.
(ii) If Token is lost: To deal with such situations, it is the responsibility of Monitor to generate new Token. Monitor will wait for Max Token Return Time and if the token is not received then it will generate a new Token.
This is a very common pattern in these types of things.
Usually, the "concept artist" gets all the credit, but the one in the back room was every bit as essential as the "idea person."
In my experience, "idea people" seldom understand how incredibly valuable good "back room implementation" people are.
I'm biased, though, as that's my forté, and I have had to fend off a lot of highly insulting "idea people," over the years.
> His response was unequivocal. “Seems Ethernet does not work in theory,” he said, “only in practice.”
Ooohhh... burn
None, that I know of.
Unsuccessful ones, though...pretty much every single one that I've ever met.
Story time:
I have a couple of friends that are quite rich. They run an apparel company, but a "white-label" one. You've probably worn clothes by them, but never knew it.
The "creativity" behind the company, is the wife. She's pretty awesome. But her husband is a very sharp businessman. Really humble, and low-key, but woe be unto anyone that tries to pull a fast one on him.
Without her, he would be nothing, but without him, she would probably still be in a small, windowless, room, in some Manhattan building, helping someone else get rich.
He treated people like crap. Not sure if that extended to his implementation people. I’m pretty sure he gave them a lot of agency and money, though.
I don’t really recall him surrendering the spotlight to his backroom people. Cook seems to be much better at that.
Pronouncing it forté is actually a confusion with the Italian word forte which comes from music and means loud.
Thanks!
You may like this: http://queenofwands.net/d/20031003.html
The only thing I see here to quibble with would be the acute accent on the final "e", which is rarely if ever a feature of modern English orthography in any case. Beyond that, I think GP's prescriptivism has, like that of the Académie Française, proceeded to a fault - specifically in GP's case, the fault of recommending you mispronounce "forte".
I'm simply describing how to sound like an educated person.
From what dialect are you prescribing here?
That has to be one of the worst description of ethernet that I’ve seen!
His response was unequivocal. “Seems Ethernet does not work in theory,” he said, “only in practice.” "
So, for the record, the rival technologist was right, and that is why we have IP (internet protocol).