The early days of DS36 were very shaky (we'd start the talker when we wanted to talk to our friends and kill it shortly thereafter), mostly due to problems with A/IX's compiler. I'm not certain whether they were bugs in NUTS 2.3.0's code or bugs in the compiler itself, but you couldn't use the .who command without a core dump.
That was compounded by the fact that I'd only programmed in BASIC up to that point. Over Christmas of my senior year, I believe, I took the code to my Grandparents house, where I created additional functions to mimic the ones that were there, only doing slightly different things. Over time, as I got the feel of the language, my coding started looking less like the original functions, but still fairly simple.
The Wonder Years
If the Unix1 era was infancy, the Mars era was the talker's childhood.
I arrived at UTM the next year, and began running the talker out of my
own account... on the newly created student box, Mars. During this era,
Deep Space 36 grew by leaps and bounds, reaching a regular user base of
around 50 people and a total user base of well over 200, though the
exact number cannot be determined because the old code didn't screen
very well for people with multiple characters.... The code also changed,
thanks especially to one of our wizards, Hawk, who provided me with
_huge_ lists of possible new commands, mostly things she had seen on
other talkers. Since I hadn't ever seen many of the commands, I, of
course, implemented them blindly, usually a little differently than
other talkers, hopefully a little better in some cases (like the afklog
command). And as always, there were a few that were existing commands
with different names. Regardless, had it not been for Hawk and others
like her, Deep Space 36 would not be the talker it is today.The Mars Era
The Hacker Years
Mars was a slow machine by today's standards (486 of some sort),
yet fast enough that it didn't really mind the extra load. Then,
as more and more users began using the system, the admins began
looking at what users were doing a little more carefully. Thus,
about a year after it began on Unix1, Deep Space 36 reached
puberty, and it hit hard. Deep Space 36 was forced underground
for a while, with the receipt of a message from the sysadmin
entitled Guys and Gals Stop Yer Talkers. (I still have it, and
might even put the discussion online as a historical document,
if I can get permission to do so.)
Thus, the Hacker years began. At first, I assumed the talker was
probably dead, but then I found a web page about free unix on
Macintosh computers. It was an outdated link about MacBSD, which
said that MacBSD had merged with NetBSD and provided a link to
NetBSD's home page. Later, when IIsi support became part of a
standard distribution, I was there. That Mac IIsi's NetBSD drive
died some time ago, though now my PowerBook 145 is partially
working, running somewhere between NetBSD 1.2.1 and 1.3.The NetBSD Era
The NetBSD, or offline era, while seemingly a death sentence, was actually a blessing in disguise. In just under a year, I fixed numerous bugs in my code, as well as adding support for multiple worlds. (See Coding: In His Own Words for more details.)
The Bleeding-Edge Years
Sadly, NetBSD-mac68k did not (and still doesn't) support Asanté
EN/SC Ethernet adaptors, because of Asanté's refusal to provide
any technical information. Since I didn't have a way of leaving my
modem on campus, where the IIsi is (my father's office), and didn't
have any idea how to set up PPP under *BSD, anyway, I was rather
elated when I accidentally stumbled across the announcement of
a version of Linux, hosted on a Mach 3 microkernel, to be made available
for PowerMacs. After much fun trying to get it to work on my AV model
PowerMac, things finally started working, and everything began to change.
First of all, setting up PPP was easy under Linux (it's pretty easy
under NetBSD, too, now that I have a clue), and in a few short days,
I was able to actually alpha-test the New Deep Space 36 on the internet.The MkLinux Era
Thus the talker reached its Gen-X era, that time when everything is bleeding-edge beta -- or alpha. To make a long story short, this was mostly a bug fixing period, working out all the new bugs that were introduced by the WORLDS support, creating new worlds, and cleaning up old ones. It was during this time that the UTM world was separated from the station, connected only by a portal (now a bus).
However, one other major code update occurred during this period -- the mail code. Written primarily in the summer of 1996 (using the PB145, via a serial console) it was, of course, extremely buggy. It has since been extensively modified, but the bulk of that code remains, commented out, as a reminder that things don't always work right the first time around.
The Golden Years
Throughout all this, I managed to volunteer for the GlobeGate project.
It all began when Dr. Bob Peckham (Tennessee Bob) sent out a list of
links to a bunch of his students and other folks. Being fairly proficient
at C by this time, I set about writing a program to convert this list
into HTML hot links, and put up a web page containing the links. After
that, I began creating "Dave's French Links", which later evolved into
a part of the GlobeGate project.The GlobeGate Era
Eventually, I managed to convince the Department of Modern Foreign Languages that things would be much better if they had their own un*x box for the GlobeGate web pages. They were, at that time, being served by an IBM running a Windows-based server, serving files off of a Novell File Server, a combination that proved somewhat slow and problematic, especially compared with a dedicated box that serves both functions.
Thus, the talker finally reached its golden years. When the department purchased a small lab of PowerMac all-in-one's, they decided to also buy the computer store's last 7600/120 (the display model). At about the same time, the bleeding edge PCI PowerMac support for MkLinux was completed. After several emails back and forth to one of the developers, and a SCSI ID change, it sprang to life.
Since we didn't have a permanent IP number, the bootp server assigned it (randomly) the IP number 192.239.150.168. I punched it into the unix side of things, and it just stuck. Shortly thereafter, we requested the domain name globegate.utm.edu. At that point, the talker had been transferred over and was in preliminary testing on its first permanent internet connection in a little over a year.
Since then, the talker has been drastically updated. Over the last few months, hundreds of bugs have been fixed. New code is added to the talker almost every day. In fact, the talker's source code is almost six times the size of the original, and takes a whopping 6 Megs of memory (though only about 800k is generally resident in physical RAM. The talker web pages have also been drastically upgraded, from the first DS36 page, which was little more than a "Hello from Jkurtz" page, to the current page, which contains this history along with a code history, and even the start of some graphics. There are many projects in the works, from support for the talker Netlink protocol (or a superset of the same), to a web interface for chatting with your web browser... all right here on the New Deep Space 36.