This issue of TidBITS marks our 12th anniversary of continuous publication on the Internet. On some of our previous anniversaries I’ve written about the early history of TidBITS, lessons we’ve learned over the years, and how things have changed from the early days. Those articles remain accurate and relevant, so rather than regurgitate them here, I’d merely encourage you to go back and check out the originals in our article database, particularly if you weren’t a subscriber back then.
For artifacts from our earliest history, though, we’re indebted to Google, which recently finished bringing 20 years of Usenet archives online – a total of over 700 million messages. My memory certainly isn’t good enough to tell just how complete the archive is, but I was able to find two items of interest: the first announcement of TidBITS to the comp.sys.mac newsgroup and the first actual issue of TidBITS as sent to the comp.sys.mac.digest newsgroup as a binhexed HyperCard stack (bonus points to those who can remember and perform all the steps necessary to read the stack in HyperCard in Mac OS X).
This year I don’t want to look back or dwell on past successes. Instead, I’m going to pretend briefly that TidBITS is a public company that has to reveal the challenges facing our business going forward. Some of these, such as the possibility of Apple going bankrupt, are sufficiently severe (and unlikely) that there’s little point in planning for them. So I’ll focus on two very real things that would keep me up at night if I didn’t have a small child in the house already helping with that task. I expect that other small organizations may find themselves facing similar issues; perhaps my thoughts or an ensuing TidBITS Talk discussion will help crystallize your thinking on the topic.
Staff — When chatting about who will take responsibility for any new TidBITS project, the discussion always centers at some point on the "hit by a bus" concern. In any small organization, each individual is extremely important, and planning for what should happen if that person suddenly disappears for whatever reason is essential. Our goal in the past has always been to set things up such that any member of the staff, armed with appropriate passwords and internal information, would be able to write and distribute an issue of TidBITS. For many years, in fact, I published TidBITS almost entirely myself, but as our services have increased in number and complexity, it’s become difficult to imagine any one person doing it all.
Nevertheless, I still believe that the short-term possibility of such a single-handed effort remains a valid goal for TidBITS. We do pretty well with the writing and editing portion of TidBITS, and we bring in outside authors when possible for additional knowledge and fresh opinions. And as I’m sure is the case with many publishers, I have a mental short list of people I’d love to suck into the TidBITS staff vortex if funds were suddenly to become unlimited.
More concerning is how our staff interrelates with our technological presence. When no off-the-shelf solutions have been available to solve a given problem, we’ve generally responded by rolling our own – "off-the-wall solutions" if you will. As with our article database and the TidBITS Talk archive, we (and by "we" I mostly mean "Geoff") would design and write the necessary code and back end database. We’re quite happy with our results, but there’s no question that these homegrown solutions require more baby-sitting and maintenance than is ideal. That means that Geoff’s "hit by a bus" quotient is pretty high in the short term; in comparison, my "hit by a bus" quotient is relatively low in the short term, but obviously very high when looking at TidBITS over the long run.
Technology — Minimizing our collective "hit by a bus" quotient leads directly into our second challenge, that of updating our Internet servers to something that doesn’t remember Bill Clinton’s first term as president. Right now the machines we use for the main Web and email server, databases and searching, and for the mailing list are a pair of Power Mac 7600s, a Power Mac 7100, a Performa 6400, and a Power Mac 8500. Plus, we’re sometimes several versions behind on WebSTAR, EIMS, FileMaker, Lasso, and ListSTAR. The sheer age of these Macs and programs doesn’t bother me – there’s nothing wrong with using older technology that meets one’s needs, and what we have now does meet our needs. But at the same time, I’ve started to think more about replacing our elderly systems for a variety of reasons.
Digital entropy. My experience, and I have no empirical evidence to back it up, is that despite their digital nature, both hardware and software systems age in a very analog way. After you set something up, there’s often a short period of break-in, where there are a few unexpected problems. Some you may figure out and fix, but others just go away after a while as the machine gets comfortable with itself. Then there’s a long period of basic stability, or at least predictability, sometimes punctuated by short bursts of instability. But as time goes on, the accumulated cruft of years of basic use and occasional problems builds up to the point where problems start to become more frequent and more random, to the point where major changes become necessary. Again, I can’t point to any specifics here, but my gut says that it’s time to start thinking about the future.
Continued relevance. It would have been impossible for us to keep TidBITS relevant if none of us had upgraded our personal machines to Mac OS X. In a similar vein, we need to be learning about new versions of server software so we can pass on our experiences. That involves testing and running Mac OS X server software, and doing that essentially requires all new server hardware.
Periodic Refresh. Although we’re generally happy with how we’ve designed our systems from both front and back ends, the fact that they’ve evolved slowly over years means we constantly notice things we’d do differently if we were starting over. That might mean replacing certain pieces of software we’ve found to produce bottlenecks, changing processes so any one of us can perform them, and so on. It’s pointless to put any significant effort along those lines into our current systems, providing yet another reason to start looking toward new systems.
Your Opinions Count — Obviously, the fundamental reason we do all of this is to serve our readers better, and as such, I certainly hope people who are interested in our technological challenges and solutions will chime in to TidBITS Talk discussions with thoughts about what we’re doing well, what we could be doing better, ways we might do those things better, and new ideas about what we might do in a new system. Long ago, we even asked readers to develop some sample systems for us with our search engine shootout, in which we chose among a number of excellent search engine systems. Our massively increased data set might preclude such an approach this time, but either way, your opinions are important to us because, let’s face it, you’re the people who will be using these systems.
There’s no question this development effort will be a huge task, but I hope we can all have some fun putting it together. Luckily, we have no specific schedule and if necessary, we’re happy to wait for the necessary versions of some of our old server friends from the past to make the jump to Mac OS X.