Geek Trivia: Net losses and Ethernet gains

How did Jerry Saltzer's denunciation of Ethernet "help" 3Com capitalize on the technology?

Twenty-five years ago this week—Sept. 30, 1980—Xerox, Intel, and Digital Equipment Corporation (DEC) jointly published the formal specifications for Ethernet technology, sparking a technology showdown that led to the near-ubiquitous Internet, without which your favorite trivia column would likely never reach your browsers and inboxes.

Ethernet is one of the many ahead-of-its-time technologies developed at the Xerox Palo Alto Research Center (PARC) in the 1970s. The PARC is perhaps most famous for the Xerox Alto, the first personal computer with a graphic user interface and a mouse—developed way back in 1973.

Xerox executives infamously refused to approve the Alto for production, feeling there was no viable market for such devices. Steve Jobs was among the many future tech luminaries who saw the Alto in action, and word has it that he used the ideas at work as the inspiration for his own WYSIWYG-based computer revolution.

Often overlooked in this well-worn tale of opportunities lost and found is Ethernet, also utilized by the Alto. This technology also suffered a long and winding road to success beyond the walls of the Xerox PARC.

The man most often credited with inventing Ethernet is Robert Metcalfe, a researcher at PARC in the 1970s. In 1976, Metcalfe, along with assistant David Boggs, published the seminal academic paper, "Ethernet: Distributed Packet-Switching For Local Computer Networks." When Xerox failed to immediately realize the value of Ethernet, Metcalfe left PARC and started his own company in 1979. He named it 3Com.

Alas, neither Ethernet's nor 3Com's stories were ones of instant success, as a pair of established, proprietary technologies already had a stranglehold on the fledgling computer networks of the day. IBM preferred Token Ring networks, and ARCNET systems also enjoyed market success in the early 1980s. Both blocked Ethernet's entry onto the early LAN landscape.

Indeed, no less an authority than Jerry Saltzer—a computer science pioneer whose work laid the foundations for both the UNIX operating system and Internet protocols—openly criticized Ethernet in academic circles. Ironically, Metcalfe often cites Saltzer's critiques of Ethernet as the basis for 3Com's eventual Ethernet-fueled fiscal successes.


How did computer science pioneer Jerry Saltzer's open criticism of Ethernet technology help 3Com capitalize on Ethernet's eventual market success, at least according to 3Com founder and Ethernet inventor Robert Metcalfe?

Saltzer, an influential voice in computer science circles—both then and now—co-wrote an academic paper professing that Token Ring networking systems were inherently superior to Ethernet systems. From a certain perspective, this was true, as Token Ring systems were far more complex than Ethernet's more simplified protocols. And despite the fact that Xerox, Intel, and DEC had all agreed to promote Ethernet as a networking standard in 1980, Saltzer's words had a significant early effect.

By Metcalfe's measure, Saltzer's endorsement of Token Ring LANs caused major hardware developers to shy away from Ethernet, which forced 3Com to develop simplified peripheral Ethernet cards that the company could sell as aftermarket add-ons—rather than integrated components of computer systems. As the utility of Ethernet's simpler-than-Token-Ring technology began to catch on in the early 1990s, 3Com began selling large numbers of these network cards, quickly becoming one of the high-profile darlings of the Internet boom.

As a result of this reversal of fortune, early Ethernet proponents jokingly remarked that "Ethernet works better in practice than in theory"—thumbing their noses at the "theoretically superior" Token Ring systems that were quickly falling out of favor. Ironically, Ethernet has become so ubiquitous that many motherboards now feature integrated Ethernet ports.

We can summarize the aforementioned simplicity of Ethernet with a six-step algorithm (quoted below from Wikipedia), originally developed in the 1960s to avoid radio interference on ALOHAnet channels in Hawaii.

  1. Start: If the wire is idle, start transmitting; else go to Step 4.
  2. Transmitting: If detecting a collision, continue transmitting until reaching the minimum packet time (to ensure all other transmitters and receivers detect the collision); then go to Step 4.
  3. End successful transmission: Report success to higher network layers; exit transmit mode.
  4. Wire is busy: Wait until wire becomes idle.
  5. Wire just became idle: Wait a random time, then go to Step 1, unless exceeding maximum number of transmission attempts.
  6. Maximum number of transmission attempt exceeded: Report failure to higher network layers; exit transmit mode.

This elegant, scalable set of instructions is the basis for all Ethernet systems even today—to say nothing of fueling some always fascinating Geek Trivia.

Check out the Trivia Geek's blog!

Keep in touch with Trivial Pursuits, the Trivia Geek's online journal of rants, opinions, crazy ideas, half-baked notions, bizarre concepts, wild schemes, and trivial observations unfit even for Geek Trivia.

The Quibble of the Week

If you uncover a questionable fact or debatable aspect of this week's Geek Trivia, just post it in the discussion area of the article. Every week, yours truly will choose the best post from the assembled masses and discuss it in the next edition of Geek Trivia.

This week's quibble comes from the September 7 edition of Geek Trivia, "Time and time again." A TechRepublic member, the always astute AlanGeek, called me out for an intriguing omission when it comes to poking holes in integer-based UNIX time calculations.

"I thought for sure you'd mention the end of UNIX time, at least as it stands currently for 32-bit-signed integer time: Jan. 19, 2038, at 03:14:07. Just for fun, I set my Solaris box time ahead to watch it roll over to Dec. 13, 1901. Wikipedia mentions this issue and also the impact of the leap second on UNIX time."

Well, I promised my editor that I'd keep these little articles under 800 words—OK, I said 700, so 800 is my grace limit—so this tidbit didn't make the cut. Thanks for giving me an excuse to slip it in later, and keep those quibbles coming.

Falling behind on your weekly Geek fix?

Check out the Geek Trivia Archive, and catch up on the most recent editions of Geek Trivia.

Test your command of useless knowledge by subscribing to TechRepublic's Geek Trivia e-newsletter. Automatically sign up today!

The Trivia Geek, also known as Jay Garmon, is a former advertising copywriter and Web developer who's duped TechRepublic into underwriting his affinity for movies, sci-fi, comic books, technology, and all things geekish or subcultural.


Jay Garmon has a vast and terrifying knowledge of all things obscure, obtuse, and irrelevant. One day, he hopes to write science fiction, but for now he'll settle for something stranger — amusing and abusing IT pros. Read his full profile. You can a...

Editor's Picks