There Is No Cat

Ten Year Anniversary

This month, or maybe last month (when I was busy getting married), marks ten years since I first came across the web.

I was working for AT&T at the time. I just happened to have a Sun workstation on my desk, a bit of serendipity. My boss had decreed that my Mac would go into a lab where the other tech writers could also use it for their artwork, and knowing of my antipathy to Windows, he arranged to borrow an unused workstation from another group. That group had removed the hard drive and replaced it with two empty 105 MB drives, neither of which was large enough to hold a copy of Sun OS. The tech support group we had at the time had no experience with Sun OS, which used a weird little thing called Yellow Pages to do address lookups; they were used to System V UNIX®, which used DNS, and they didn’t know how to set up the workstation. So I was left to my own devices. It took me a week to figure out how to install Sun OS across two hard drives. I would start the installation procedure and take a guess as to what answers I should give to the various installation prompts, taking notes as I did. When I gave a wrong answer and the installation failed, I would wipe the hard drive and start over from scratch. The first day or two, I could get through maybe seven or eight attempts; by the end of the week, it was taking me half a day to get to the point where I would make a mistake. But after a week, I had a working Sun workstation on my desk, albeit one without the ability to translate machine names into IP addresses. Not only that, but I had root access and the ability to install programs, something I didn’t have on any of my other UNIX accounts.

It was important that I had a Sun workstation, because the initial releases I used of Mosaic in April or May, 1993, were for UNIX machines only. The program wasn’t ported to Windows or Macintosh until later that year, maybe October or so as I recall.

I was a frequent user of the Internet, well-versed in the use of programs like Archie, which helped me find programs to download via command-line ftp, and trn, which allowed me to read and participate in netnews (a.k.a. Usenet), and telnet to get from machine to machine where I had accounts. I was also using Gopher at this point, which was an advance on the command-line tools, but still pretty primitive.

I don’t remember how I heard about Mosaic; possibly on one of the groups on netnews. I fired up my ftp client and downloaded it and installed it on the workstation on my desk. I remember that it was version 0.4. When I fired it up, I realized immediately that this was huge. When I was in college majoring in broadcasting, some of my professors were fond of telling us about the coming convergence, where broadcasting, publishing, and computing would all come together and merge to some extent. They thought that teletext was going to be the medium that brought about this convergence. You’ve never heard of teletext? It was an information service broadcast in the vertical blanking interval on television signals -- the black bar that would roll across your screen when your TV lost synchronization with the signal. Teletext never caught on in the U.S. It had some success in a few countries in western Europe, like the U.K. and Holland, but that was about it. When I saw Mosaic for the first time, I recognized at once that this was the convergence my professors had talked about nine and ten years previous.

I called my boss into my office, all excited to show him the future of our business (both telecommunications, AT&T’s business, and technical writing and publishing, our particular group’s business). He nodded his head, said something non-committal, and went back to his office. Six months later, he was calling me to ask if I’d ever heard of something called Mosaic that one of my colleagues was telling him about. I couldn’t resist pointing out that I’d demoed it to him six months ago. I needled him about that for years. :-)

At the time, there were probably fewer than a hundred web sites in the world. Good thing, too, because as I pointed out above, I didn’t have DNS service on my workstation. In order to jump from site to site, I had to go through an elaborate procedure. I would have a window open with a telnet connection to one of our System V UNIX boxes that had DNS, and another window open in the /etc directory on my workstation. When I came across a link to a new host, I would click on it. My workstation would fail to find the machine, but the URL would be in the window of Mosaic. I would then switch to the telnet connection and do a DNS lookup on the System V machine of that machine name, get the IP address, switch to the window in the /etc directory, open /etc/hosts, and add the host name and its IP address to /etc/hosts. Then I would retry the link. The workstation now knew where to find the errant machine, and I could click on their site until I found a link to somewhere else my workstation hadn’t previously encountered. I think you can see that this approach wouldn’t scale to today’s web.

One thing that I remember about Mosaic was that the early versions had some menu items that hadn’t been fleshed out yet, where the intent was clear but the code hadn’t been written. Maybe it was because I was a tech writer, but one that I noticed clearly was that one of the formats you would be able to save a web page in was FrameMaker MIF. It was a long time before anything came of that. Ultimately, I think there was an external add-on to Mosaic that implemented that functionality, written by a programmer at the Norwegian state telecommunications company. I believe that programmer ultimately went on years later to create the browser Opera, incidentally. There was also the beginnings of an annotation feature that would allow you to attach notes to web pages, in your own browser and in such a way that other browsers could see it. That was another good idea that got lost along the way.

Somewhere along the line, almost certainly in May, 1993, but without question by June 2 of that year (which is the modification date of a file I found a few years ago on my old Mac SE/30 that contained a marked-up version of the FAQ I used to maintain for the rec.radio.shortwave netnews newsgroup), I had figured out enough HTML thanks to the View Source menu command to be able to write my own pages. It seemed very similar to the troff code I used to write books. (I cringe when I look at that code now; I started closing paragraph tags and looking at tags as semantic markup rather than formatting instructions pretty early on, probably in 1994 or 1995.)

There were maybe 100 web servers in the world at the time. (The official history of the web at W3C mentions that there were about 200 in October, 1993.) I don’t remember if I set one up on the Sun. I do know that I talked to one of my friends in the support group and showed him the web, and pointed out where he could download the server software from CERN, and shortly thereafter, we had a running server on our System V UNIX box. It was behind the corporate firewall, so the rest of the world couldn’t see it, so I suppose it was one of the first intranet sites in the world, or at least one of the first outside of CERN. I set up a site for my department. Since we were tech writers, I called it The Inkwell. It had stuff like job numbers and a list of the members of our group. By this time, there were other people within the company who were catching on. No more than a few dozen out of the 200,000+ employees, but enough that other servers were starting to show up inside the firewall. One was run by the company library, and they had done some interesting stuff that would be useful for writers, like putting a dictionary online. I linked to it from The Inkwell.

After a few months, the group that had loaned me the Sun wanted it back (now that I had gotten it working and it wasn’t useless any more, after all), and my boss saddled me with a PC for a few months before deciding that since I was the only person using the Mac in the lab, I could put it back on my desk. I don’t think the Windows and Mac versions of Mosaic had come out yet, because I remember being a regular user of the line-mode client that you could telnet to on info.cern.ch. Someone did a quick-and-dirty port of the line-mode client to the Mac, and I used that for a while, too. Eventually Mosaic came along for the Mac.

Throughout 1993, the web was more of a hobby for me. My boss wasn’t paying much attention to what I was doing with it at first, but after a while, he started showing it to our clients, and by 1994, they started asking us to do web stuff for them. My boss was always very good at ferreting out business for the group. He managed to get the stockroom to fund us to put their catalog up as a web site. At first, I didn’t want to take on the work, since I enjoyed having this as my hobby, and I enjoyed tech writing so much. So he got my officemate to work on it. When I saw what he was having to do, I talked to my boss about how it would be a good idea to place all this stuff in a database rather than manually creating pages for all of the items the stockroom had. Actually, the stockroom had some of it in a database. I eventualy got sucked into the process. We wound up getting a programmer to write a program that would nightly access the database and replace placeholders we had placed in our static files with the current prices. It didn’t make things any easier. My boss didn’t think it was important to use a database for the pages, and I didn’t know how to set one up, so we continued a mad scramble to keep the stockroom catalog updated, pretty much on a daily basis. It was nuts.

I remember when O’Reilly & Associates brought out their first guide to the Internet, written by Ed Krol of the University of Illinois. It had one tentative chapter about the web. There was about the same about Gopher, and lots of stuff about the command-line tools of the traditional Internet. I couldn’t find a copy anywhere near me. I wound up asking my brother, who was in grad school in Philadelphia at the time and had a branch of the great tech bookstore Quantum Books in his neighborhood, to pick me up a copy. If I’d waited a couple of months, I could have gotten it locally, because it caught on, got wide distribution, and sold a ton of copies.

At some point in here, before Netscape came out, so it must have been in 1994, the lead writer on my tech writing project and I discussed the possibility of licensing Mosaic to incorporate into some of the network management programs I was documenting to give us a way to provide online help. We talked with the local representatives of Spyglass, who had been named the University of Illinois’ reps for commercial applications, but eventually decided not to use Mosaic and HTML, because it didn’t support tables yet. I wound up creating a web page describing the process we went through and the decision we came to.

I was still doing tech writing at this point as well, but my boss eventually prevailed upon me to switch to web work full time, since I was the person in the organization with the most experience with it. I was reluctant, and more than once told him no, but ultimately I didn’t really have a choice. And as it turned out, that was okay. I came to feel that working on the web was something of a calling. I didn’t really want to do anything else.

It was an interesting time to be involved with the net. Things were changing so quickly. Gopher and Archie hung on for a while, but it was clear before long that they didn’t have much of a future. (The University of Minnesota, home of Gopher, made some mistakes, but ultimately I think the web was just better.) I was well acquainted with the history of the introduction of radio in the early part of the 20th century, and there seemed to be a lot of parallels between that and what was happening to the net at the time. The introduction of Mosaic was kind of like when the superheterodyne radio was introduced, and direct tuning, making it possible for people other than the geeks and gearheads to use radios. I felt privileged to be able to be a participant.

When I first thought about writing this, I was wondering if perhaps after ten years it was time to move on to something else and return my web work to the status of a hobby. Prolonged unemployment will do that to a person. The age of the webmaster, the person who could administer the server, code the pages, write the text, and add in the graphics, pretty much ended some years ago. The time of the pioneer is past. We’re still feeling our way, but the pathways are a little better marked than they were before. Of course, as soon as I started thinking about that, I was bound to find a job. I think this essay would have made more sense as an elegy to a time past, a farewell to the profession, but it looks like I’ll be sticking around for at least a little while longer.

Posted at 9:08 PM