The Web: a generation old, a link unexpected

Web browsing, two decades ago

Web browsing, two decades ago

On this twentieth anniversary of the first popular web browser, consider how its potential was viewed just one year after its birth. If only to realize that sometimes in tech the Law of Unintended Consequences can pay off in positive ways, too.

In September 1993, the National Center for Supercomputing Applications released the beta of NCSA Mosaic, appearing at the same time as Windows and Macintosh versions ported from Unix. Mosaic had then-revolutionary features for navigating the still-new World Wide Web: icons, bookmarks and the actual ability to display pictures “inline” with text. Which meant images showed up in the same, not a separate, window, something NCSA claimed “made the software easy to use and appealing to ‘non-geeks.’ ”

But this “graphical” browser, and even the web, didn’t face a straight, non-critical path to the dot-com boom (and later bomb) of the same decade. As a matter of fact, what we now take to be a natural marriage — hypertext and multimedia/web browsing — was viewed by some practitioners at the time as a shotgun wedding.

In September 1994, the Association for Computing Machinery’s European Conference on Hypermedia Technology was held in Edinburgh, Scotland. I was there, one of about 350 delegates split almost evenly between U.S. and European attendees. I took notes using the dominant technology of the day for that task, paper (and a Fisher Space Pen). Still-recognized big names spoke: Tim Berners-Lee (then CERN’s inventor of the World Wide Web, now a “Sir” and director of the World Wide Web Consortium); Jakob Nielsen (then SunSoft, now an influential web usability consultant); and Doug Engelbart (then Bootstrap Institute, and this year, sadly, late).

CompuServe email delivered the '94 agenda

CompuServe email delivered the ’94 agenda

My handwritten scrawl for six sessions I attended, as well as the detailed agenda distributed by email on the definitely non-graphical CompuServe Information Service, reveal a web of conflicts (and not just the session titled, “HTML – Poison or Panacea”):

“Technically better” would win, not “comfort.” An ongoing debate was whether a technically superior solution was more important than one that was “easy to use.” Even Nielsen, responding to one of my tremulous questions about whether a good interface should provide both access to information and a degree of user ease in getting to that info, responded that it was debatable — Apple’s overly cute (and later doomed) eWorld online service had provided a metaphor he thought had been taken too far.

While Nielsen may have ultimately been right, the conflict turned out to be moot. William Nisen, a conference speaker who at the time was president of hypertext pioneer OWL International (its U.S. headquarters was located in Bellevue), reflects now that then, “technically superior products were also the best customer products. Speed and flexibility that were the cornerstones of a great customer product could only be achieved through great engineering … Moore’s Law (about how rapidly computer hardware advances would allow computing power to double) allows us to luxuriate in a customer-centric design world.”

Too cute to live

Too cute to live

We don’t need no stinking “multimedia.” While today it’s impossible to imagine the web without heavily linked video, audio, and high-resolution photos, the dislike of “multimedia” was almost palpable among many at ECHT ‘94 representing academic and commercial interests. (One panelist for “Does Multimedia Make a Difference?” went so far as to call multimedia a fad and “useless” for hypertext.)

Keep in mind, at this juncture, linking was almost all in local network, internally connected hypertext documents. Hypermedia and/or linking everything-everywhere was a pie-in-the-sky concept, suitable for visionaries like Ted Nelson and his longtime Xanadu all-world-knowledge-is-available-to-anyone project, documented in my dog-eared copy of Nelson’s book Computer Lib/Dream Machines.

“Ted was envisioning a hypermedia-connected universe while we were concentrating on an efficient hypertext/hypermedia engine,” notes Nisen. “The major conundrum … was the discovery of the killer application that was going to propel hypertext/hypermedia out of the lab and into the market.”

Navigation was nasty. Every product on the market today for hypermedia has little or no support for justification and typographical control … few fonts look good on screen … no hyphenation/justification can be done in real time when re-sizing windows.” My notes back then would be a horror story for web designers now. And icon or navigation conventions? One speaker discussed using “highway signs” for icons. Another who had developed what he thought were great, intuitive icons, left on vacation for a week, and was told a superior had the icons made more “Windows-like.”

A major development noted on one panel was that Mosaic “ages” colors on a clicked link, something we today take for granted. Engelbart stressed, in his session prophetically titled “Long Distance Perspectives on Hypermedia,” the now-obvious observation that people don’t “scroll” in real life, they “jump.”

Overall, the ECHT in 1994 illustrated an inherent mutual distrust of academic by commercial interests, and vice versa. As one speaker from the company InfoAccess noted, “Universities don’t like actually developing things because it causes all sorts of problems.”

ACMlogoNow consider what happened next. NCSA Mosaic achieved skyrocketing popularity, going from 5,000 downloads per month at its 1993 launch to 50,000 per month by mid-1994, then more than 50 percent browser market share by 1995, and became the foundation or inspiration for subsequent killer apps of the 1990s, the browsers Netscape Navigator and Microsoft Internet Explorer.

In doing so, the web browser overcame the largely backward-looking concerns of some ECHT speakers, who could only imagine the future based on what had come before. Some saviors of multimedia-linked web browsing were out of their control (Moore’s Law, Internet uptake) and other issues turned out to not require anyone be saved (fear of getting lost on the web, inconsistent navigation metaphors).

Browsers were a textbook case of where the choice wasn’t peanut butter OR chocolate. It was AND.

The result? Two different developments, hypertext and digital multimedia, joined at the hip to create something completely new. In many ways, it was not conceptually unlike how WiFi/mobile broadband and smartphones combined, more than a decade later, to propel social media.

The tech future is not additive. It’s not even exponential. It’s quantum. Both in uncertainty and in its potential.

frankthumbnailFrank Catalano (@FrankCatalano) is a strategist, author and veteran analyst of digital education and consumer technologies whose regular GeekWire columns take a practical nerd’s approach to tech. See the column archive. He thanks Bill Nisen for making it possible for him to both attend ECHT as his guest and tour some killer Scotch distilleries afterwards in a Fiat Ferret, er, 500.

  • Rick LeFaivre

    Great article, Frank. Those were heady days. I was running Apple’s Advanced Technology Group at the time, and we weren’t great fans of eWorld. One of the many university research groups we funded was Larry Smarr’s NCSA group at Illinois. I remember them attending a university research open house we held each year, in (I think) the Spring or Summer of 1993, and demoing the early beta of Mosaic. We were perhaps a little blase (hey, we’d had HyperCard for years!), but Jim Clark saw it some months later and Netscape was born. BTW, at that time we ran http://www.apple.com off a Mac server in ATG, posting research papers and the like. Shortly thereafter Apple corporate said, “uh, we think we need to take it over …”

    • http://www.intrinsicstrategy.com/ FrankCatalano

      Rick, I appreciate your additional perspective from inside Apple at the time; it’s fascinating. I remember being a beta tester for eWorld. There were great hopes for its success (as a closed, AOL-like online service, yet more graphical) based on the activity when it was a free beta. Then it tried to charge for access, and the connect fees were breathtakingly higher than many beta testers had expected. eWorld rapidly depopulated.