A look back at our industry in 1986

(This essay was posted on my Web site in September 1996 and reprinted here for reference.)

This issue marks the first anniversary of Web Informant. I thought that it would be a good time to take a look back even further in the past to the fall of 1986. Back then, I began my writing career working for a special supplement to PC Week called Connectivity. At the time, I had left working for Transamerica Occidental Life Insurance Company in downtown Los Angeles, in a 30 person information center. This group was part of their internal IS department to support end-user computing, and I was getting more interested in local area networks, having installed the first one at Transamerica that summer of 1986. I remember working in an Information Center (as we liked to capitalize it). It was a proud profession: there were trade mags, shows and even IBM product lines geared towards us.

First, let’s talk about the Internet. Back then, there were still a small number of different networks that had grown up from various US government-funded projects. A prophetic paper by Dennis Jennings, Larry Landweber and David Farber for Science magazine mentioned how “NSFnet will probably have the most impact on science of all networking activities in the US at this time (February 1986).” NSFnet went on to evolve into the Internet backbone that we have today.

For those of you new to the Internet, as recently as five years ago a private corporation would have had lots of difficulty finding an Internet Service Provider, let alone getting their own .com domain name established. When I began Network Computing magazine in the summer of 1990, we had to piggyback on a university’s email system to get Internet access for our editors! Even as recently as three years ago, there was a single ISP with numbers in my local area code that I could call: now there are over a dozen that I can choose from for Internet access.

Ten years ago, most of the computers on the “Internet” were running proprietary operating systems or Unix. The idea of having a desktop PC running IP was ludicrous. Now it is taken for granted, and comes included with all desktop operating systems.

Mobile PC products certainly have gone through quite an evolution: back ten years ago, I remember my first portable PC was the Radio Shack model 100, a unit I am proud to say I still have somewhere in my office. It had a terrific keyboard, a 1200 (internal) bps modem, and an eight-line by 40 character screen that I used to file my stories for PC Week. Steve Roberts sent me a photo showing him typing on one next to the massive bicycle-cum-office that he rode around America back in 1984. Since then, I think I have used about 20 different laptops to write my articles. My favorite was the NEC Ultralight notebook that I used in 1989 (one photo I have is of me typing away at the hospital shortly after my daughter was born). Unlike my present portable, it was small, light (less than 4 lbs.) and had great battery life.

Roberts’ article mentions the “Plus [automated teller] System promises to make all this [mobility] easy someday, but at the moment [May 1984], its nodes are far more sparsely scattered than are those of CompuServe.” Interesting how both networks have grown over the years, but I think the number of ATMs far exceeds CompuServe nodes at this point. Roberts is now on to outfitting a sailing ship with various computers, by the way.

What computers were we using ten years ago? Well, at Transamerica the most popular device (in terms of numbers, not necessarily in terms of emotions) was still the 3270 terminal — by the time I left in 1986 we almost had 2,000 apiece of PCs and mainframe terminals. The best PC at the time was the 386 with 640 Kilobytes of RAM, introduced by Compaq that fall. However, most of the machines we had at Transamerica were 8088s, with some ATs.

What was the predominant software back then? Why, Lotus 1-2-3 of course. I still have my copy of version 1A, and not too long ago I installed it on a machine and was gratified to see its familiar grid pattern. And also I was gratified to see that I could still remember how to use it.

Back then, a good portion of my end-user support effort was getting the right video drivers (remember Hercules graphics on IBM monochrome monitors?) to work properly. MicroChannel and EISA bus machines had not yet been invented, and Apple had begun selling Macintoshes a few years earlier. Most monitors were woefully small by today’s standards.

Networking was a very different picture back ten years ago. Of course, the biggest networks were still those connecting terminals and PCs with 3270 cards to IBM mainframes. In the fall of 1986 the Manufacturing Automation Protocol was picking up interest (and actually on version 2.1!). A company called Industrial Networking Inc. was formed to sell products from Ungermann-Bass and General Electric, only to fold a year later. UB, by the way, was the first company besides IBM to sell token ring gear that summer. Now the automotive companies (the core group of MAP’s original sponsors) are fully behind IP and the Internet.

Novell and IBM were the predominant LAN software vendors back then — NetWare was one of the first products to take advantage of the protected mode of the 286 processor, something that IBM finally delivered on with OS/2 several years later. Token rings were just 4 megabits, and used passive “MAUs” for hubs that still are around today (mainly because they are one of the few hubs that don’t require any power to operate). 3Com was selling Ethernet cards by the truckload but having a hard time with the original “network computer,” a diskless workstation called the 3Station. Some things never change.

Looking back at PC Week Connectivity (yes, I still have the back issues), I am amused how many of the stories written then still cover many of the same themes we have today. One of my first reviews published for PC Week ran in Jan 1987 about Attachmate’s 3270 emulation products. At the time Attachmate was a brand-new company, and I stated “No matter how good the Attachmate product is, it will be tough to gain market share over DCA and IBM.” Since then the company bought DCA and IBM has played a lesser role with 3270 products. Oh well, can’t always call ’em.

The December 1986 PC Week had some interesting prices: 2400 bps internal modems from Hayes were selling at close to $800, 80386 PC 18 Mhz clones with 512 k bytes of RAM were going for $4500, Microsoft was selling version 1.03 of Windows and version 3.1 of Word, and 3Com’s servers cost $6000 and had whopping 70 megabyte disks!

What was the computer trade publication landscape back then? Well, PC Week, Infoworld and Computerworld were the predominant news weeklies. I re-read the parody issue called ConfuserWorld which was printed in 1983 and contained headlines such as “IBM calls it quits”. The best part of the parody publication were its ads, though: one for a new computer-related TV show called Happy Daze where “the Fonz fixes an HP-4000 by kicking it in the drive unit.” Another ad for Kodex maxi- modems promised that your could “communicate with confidence” on those “special times of the month when communications is at its peak.”

There weren’t any Internet-related publications, and just LAN Times and LAN magazine were devoted to networking topics. Networld and Interop were separate trade shows, and Comdex was still too crazy even then.

We have come a long way in a decade. Thanks for all of you that have sent me documents and anecdotes from that era — I appreciated reading all the responses. It has made for an interesting trip down memory lane.

Computerworld (1996): The rise of web-based user interfaces

The jury is still out over whether having the web as a universal graphical interface for applications is a step forward or backwards. I started keeping track of the more notable examples on this page of my website (and haven’t touched it in years, sorry.) Clearly it isn’t just another pretty interface for all your applications, as I originally wrote this story for Computerworld back in June of 1996.

Some analysts aren’t thrilled about Web UIs: Alan Cooper, a user interface consultant in Menlo Park, Calif., thinks that “HTML has set computer programming back 30 years and is about the worst technology I’ve laid eyes on.”

Others are more positive: Fred George, who used to work on IBM’s user development for OS/2 and is now an independent consultant based in Boulder, Colo., says he is “glad to see that we have moved beyond the 1960s-era teletype-style user interface,” and has hopes that web technology will eventually catch up with the operating system graphical interfaces. And software developer Bruce Fram who runs his own Silicon Valley company called Relations Software, says “Web technology is the information mechanism of the 90’s. Application developers who are not using it as their primary end-user interface will be in the same position as scribes after the invention of the printing press.”

Just about every software category now has at least one instance of where a vendor has taken to the web: you can access calendars, groupware, email, and even 3270 terminal emulators via an ordinary web browser. And plenty of hardware products, including printers and routers, now come with their own built-in web server for management purposes.

There is an important cross-platform advantage for having web interfaces: software vendors can cancel their Mac and Unix development efforts and concentrate on figuring out new ways to shoehorn functions into HTML tags. And there is always the hope that webification will simplify training users on how to do things, since they just point at a link and click away rather than have to learn a new command syntax.

“Writing HTML interfaces means that software developers tend to develop to the lowest common denominator,” says Elizabeth Rosenzweig, a usability manager with Kodak’s Boston Development Center in Lowell Mass. That could be trouble, although certainly any graphical interface is better than the cryptic telnet command-line interfaces used to manage routers and other hardware devices.

The software side of things is harder to judge. Some products make a great deal of sense to webify: take network-based calendars, for example. B.W. (Before the web), you had to go through a messy synchronization step just moments before catching your flight out of town. Now one can leave the data on their servers and view it with their browsers, and more importantly, actually have a chance at keeping their calendars updated.

So the web has shifted the debate from whether Win16 or Win32s is the best programming interface — now we can argue whether HTML extensions from Netscape or Microsoft are better. Much more understandable by the common user, and much more fun to watch. But having all software go to the web user interface might hasten having an all-Windows world: since multi-platform apps can be supported by back-door HTML, developers have moved away from Everything Else and concentrated on Everything Windows.

 

 

 

Breaking the Internet Speed Barrier

Want to go into the wayback machine and see what early 1996 Internet access tech looked like? Here you go. This is one of my favorite feature stories, written for the Ziff print pub Windows Sources under the able direction of Jackie Gavron. Jackie is one of the best editors in the business, and sadly, we haven’t had the opportunity to work together for many years.

Hanging Up Your Shingle — what questions to ask your Web hosting provider

Please note: this article is 11 years old.

"Hanging Up Your Shingle — what questions to ask your Web hosting provider."

One of the great things about the Web is that your stuff can live forever, if someone has taken the time to archive it properly. Back at the dawn of Web time, I did some work for an O'Reilly publication called Web Review. The guys and gals behind this effort did a fantastic job of creating one of the more thoughtful and interesting pubs back in the middle 1990s when we all didn't really know what we were doing. 

Lucky for me, Jennifer Robbins, the graphic designer of the publication, keeps one of my stories on her site. What I find interesting is that apart from the names of some of the companies (remember when Prodigy offered Web hosting? Digex?) and protocols (gopher anyone?), the piece still holds up well and has a lot of great advice.