A JediMUD history would not be complete without Jeremy's account. This material was excepted from the Circle 3.0 site and from personal email exchange with Jeremy.
The Birth of Rasmussen (Ras)[]
I was introduced to MUDding in my first semester at Johns Hopkins at the tender age of 17 when I noticed some friends playing an interactive, multi-user game -- WhatMUD , a MUD run at Hopkins by undergraduates Dave Reed and Justin Chandler . I began to play, choosing "Rasmussen" as my character's name (after the character Berlinghoff Rasmussen, star of that week's new 'Star Trek: The Next Generation' episode.) After WhatMUD was forced to close in November of 1991 due to a hardware failure, I found a copy of the DikuMUD source code and began developing my own MUD, finding that working "behind the scenes" at creating a MUD was far more enjoyable and interesting than playing one. It was also a nice distraction from the usual college life of class work, planning my personal budget, and not going on dates.
After a couple of weeks of learning how the DikuMUD source code was organized and adding a couple of elementary new features, my fledgling MUD was (covertly) tested for the first time by several of my friends and me. Dubbed 'CircleMUD' in honor of the DECstation on which it was running (circle.cs.jhu.edu ), the MUD seemed to work well and (still covertly) was opened to players on a very limited basis.
A Circle is Born[]
Development and limited play-testing continued for several weeks, and Circle began to get a fair amount of word-of-mouth popularity on the Hopkins campus; between 8 and 12 people could often be found logged in at once. At this point (January 1992), I decided to take the big plunge and inform the Computer Science system administrator (at that time, Tim Stearns ) of CircleMUD's existence, knowing it would either lead to the MUD's immediate demise or open the doors to running it in the open without having to hide it.
Happily, Tim responded positively, declaring that Circle would be welcome in the Computer Science department as long as I made sure it did not use more than 2.5MB of disk space. (A disk cleanup might have helped.) With that fateful email, I happily declared in Circle's MOTD (Message of The Day -- the message users see when logging in), "CIRCLEMUD IS HERE TO STAY!!" officially opening Circle to players on a full-time basis.
At first, Circle was plagued by the same problems many other MUDs have -- instabilities in the code and a lack of features. However, my advantage was that I was a programmer, whereas many other MUD administrators of that era (and even today) were solely game-players, not programmers, and weren't capable of fixing many of the more elusive bugs. Circle soon became an exceptionally stable and bug-free MUD -- a reputation which I think it still maintains to this day. I was also able to create other special features such as a MUD Mail system (commonplace today, but rare at the time.)
I never advertised CircleMUD in any public forum such as the DikuMUD newsgroup. Through word-of-mouth advertising alone, Circle became moderately popular; it became standard to find 30 to 40 people playing during the afternoon. The operating system under which Circle was running (Ultrix 4.0 on a DECstation 3100 ) limited the MUD to about 58 players -- a limit which was reached several times over the course of Circle's existence.
Nasty Politics[]
However, the dark side of MUDding soon reared its ugly head, as political battles ensued in Circle's later months of operation. I had appointed several players as "GODS" (i.e. administrators) to help the "mortals" (i.e., players) with problems, and immediately received a tremendous amount grief: some of the Gods opposed my policies and others disagreed with which Gods should receive promotions. ? Some "Immortals" (i.e., players who have received enough points to 'retire' from active play and are granted special privileges) also loudly opposed various policies. One of these IMMs likened me to Hitler!
Not all Gods and Immortals were displeased: many of them became close friends in "RL" (Real Life -- i.e., outside the MUD). However, finding that I was devoting more time to politics than coding, faced with the imminent loss of the circle machine as a site on which to run the MUD, and given an offer to work solely as a coder (and not as a political administrator) on a new MUD running at Hopkins called JediMUD, I chose to permanently bring CircleMUD down on August 26, 1992.
The decision disappointed many of Circle's players; at least eight of them offered to host Circle. However, not wanting to deal with the political problems of running a MUD on my own, or dealing with a "foreign" site to run my MUD from, I instead archived the CircleMUD code and became a full-time coder for JediMUD at the beginning of September, 1992.
It should here be noted that any individual who wishes to start his or her own MUD must first find a "code base" -- i.e., a currently existing MUD on which he or she can build. The original DikuMUD, released in 1990, was not desirable to many people because it had many bugs and did not have any "modern" MUD features which most MUDs of the day already had. Most MUD implementors kept their code closely guarded, not wanting to work for hours to create special new features only to have some other MUD use them.
Towers of Babel[]
The few MUDs which had released their code publically were better than the original DikuMUD, but were often difficult to find (you usually had to get them from someone who knew someone who knew someone who had a copy), and some were quite old and did not have many modern features. At least one publically available MUD, SillyMUD, did have many fancy features, but was large and unwieldy. In addition, many people did not want to start with a MUD which had fancy features -- they wanted to start with a very "plain" MUD, so as to be able to add their own fancy features based around their own vision of how a MUD "should" be.
Another publically available MUD, Merc DikuMUD , was compact an efficient, but had a very different "look and feel" than the original DikuMUD -- in fact, much of the DikuMUD code and area files were not even compatible with Merc, turning many MUD implementors away from it.
All of the publically available code, with the notable exception of Merc, also suffered from the problem of portability: if the person who released it was using a Sun running SunOS, and your machine an SGI running Irix, you were in for some major headaches getting the new code to work on your system.
Open Source MUD[]
In May of 1993, in light of all the above circumstances, in addition to the constant pleas on the USENET newsgroup rec.games.mud.diku from people asking where they could get a nice copy of DikuMUD source code, I realized that there was a big niche waiting to be filled in the DikuMUD world. A well-written, stable, bug-free, publically available and easily accessible DikuMUD code base was needed. A code base which was fancy enough to have the standard features which most contemporary DikuMUDs players and implementors expected, yet basic enough to allow each implementor huge latitude in customization, and which which would be easily portable to many different operating systems and hardware platforms. I decided that with a little work, CircleMUD would be perfect to fill that niche.
I pulled the original CircleMUD out of its archive, and for many weeks during the summer of 1993, devoted almost all of my spare time to modernizing and improving Circle. I infused Circle with much of the Diku code I'd written while working for JediMUD, in addition to dozens of other new features and optimizations.
CircleMUD 2.0 , the first public release of Circle, was quickly becoming a reality. Circle 2.0 was specifically written to be very small and efficient -- indeed, the basic system used only about 2 megabytes of memory (an unprecedented amount at the time -- most MUDs of its size used 6, 8, or even 10 megs.) Circle 2.0 was specifically designed to be flexible, easy to expand, and easy to debug -- a starting point upon which other MUD implementors could easily turn their dreams of what a MUD should look like into reality.
CircleMUD Today[]
Since Circle v2.00, many new features have been added; much of the code has been revised or completely rewritten to be smaller, easier to understand and modify, more efficient, and more functional. The most current version is v2.20; the release of CircleMUD 3.0, eagerly awaited by much of the Circle and Diku communities, is scheduled for release later in 1996.
Since its original release, the various releases of CircleMUD have been downloaded tens of thousands of times from its official FTP archive by users from every corner of the United States and from over 65 foreign countries. It has also been downloaded an unknown number of times from dozens of other FTP sites in the U.S. and abroad which started carrying it (most of them through Linux mirroring). A CircleMUD mailing list, with a readership of about 350 people in May of 1996, was established for users of Circle to discuss their experiences, toss around ideas, exchange code, and propose new directions for future versions of Circle.
Circle was designed to be highly portable on as many Unix systems as possible; it is known to run under Ultrix, SunOS, Solaris, IRIX, AIX, Linux, HP/UX, ConvexOS, and other versions of Unix using the GNU autoconf package. Using the WinSock API, it is even possible to compile and run CircleMUD under Windows 95 and Windows NT! In addition, CircleMUD users have also ported Circle to the Amiga and OS/2.
Hundreds of MUDs around the world are based on CircleMUD. New Circle-based MUDs pop up almost every day, and I've recieved thousands of pieces of mail from thrilled users who love the code. Some use it to run huge MUDs intended to attract an international audience; others just use it to hack on at home and maybe share with their friends or schoolmates. It makes me very happy to know that I've contributed to the total amount of coolness inherent to the Universe, even if CircleMUD really is a small contribution in the grand scheme of things.