"Making of UO" articles at MMORPG.com
Read them here: Ultima Online UO General Article: The Making of a Classic Part 1 and part two.
I wasn't able to really sit down with Adam Tingle, the author, but he did run around the blog archives a fair amount. There's some inaccuracies here and there, but it's a decent overview.
Throughout 1979 Garriott would design his computer role-playing game, revising it, adding to it, showing his friends, and finally when "D&D 28b" was finished, he renamed it Aklabeth…
It's "Akalabeth" not "Aklabeth" — you can actually play it on your iOS device these days.
Together [Starr and Richard] started to knock ideas back and forth, based upon these new experiences, and came up with the premise of a multiplayer Ultima, or "Multima" as they affectionately termed it.
My understanding is that talk of "Multima" had been going on for quite some time, including discussions with Sierra (back when they ran the ImagiNation Network). This also leaves out the role of Ken Demarest, who was a moving force in getting the project started.
1996: So now that they had the art style, the funding, and the engine – they needed a direction for the game to head in…
Technically, we didn't have the engine at the point the article states; the client was basically rewritten in 1995-96. Rick Delashmit had been there for a few months when my wife and I joined the project on Sept 1st 1995; other key early folks such as Scott "Grimli" Phillips and Edmond Meinfelder also joined in August to September of 95. That's also around when Ken Demarest left, and Jim Greer — best known today for founding Kongregate.
I think I have told this story before, but the whole "dragons eating deer" example came from the design samples that my wife and I sent in as part of our job applications. We showed up on the first day and were taken aback when we were told that was how the game was going to work… So at least that much of the notion of "what the game was going to be" was set in 1995…
That crazy resource system stuff, particularly some of the AI, did in fact work in the alpha test. It led to rabbits that had levelled up and were capable of taking out wolves — or advanced players. We found this intensely amusing, and quoted Monty Python at each other whenever it came up.
Being as most of the team in charge of UO were coming from single-player games, with very few MUD veterans involved in the process…
This is just not really right. At least on the game dev team. From that September team, Kristen and Rick and I came from DikuMUDs. Edmond came from MUSH and MOO backgrounds. Scott and a tad later Jeff Posey came from LPMUDs. We had Andrew Morris, who was the original lead designer, who was a veteran of U7 and U8. And of course, our first artist, Micael Priest (most famous for his amazing poster art for bands in the 70s) wasn't an online gamer either.
Later, as the team grew and absorbed a lot of folks from U9 (which was suspended for a while) there were plenty of non-online folks on the team. But the basic premises of UO were definitely set by folks from MUDs.
…the idea of GMs taking an active dynamic role, never materialised as initially intended…
The article says that the idea of having GMs take active roles in running events never panned out… but those who recall the Seer program and the many phenomenal live events that were run know that in fact, this did happen quite a lot.
By the time the alpha had ended, the Origin team had collected enough data, were able to fix bugs, glitches and exploits, and finally the home stretch was seen bobbing along the horizon…
This leaves out one dramatic and important step in UO's history. The alpha was not an MMO in the "really massive" sense of the word. It supported the same sort of concurrency as Meridian 59 did — 250 or so. In between the alpha and the beta, the server was rewritten to allow for 2500-3000 concurrent players per shard. In order to do this, a whole bunch of new technology had to be invented for creating seamless borders between adjacent maps. These borders would prove to be a source of bugs for years (most dupe bugs made use of race conditions when moving across server lines).
Vogel would later admit "We were pressured on time. I wish we'd have had a little bit more time."
All game dev teams say that, right? 
In UO's case, the time pressure was fairly extreme towards the end. After the huge reaction to the beta, all the eyes of the press were on us. A big meeting was scheduled to decide whether to ship — on a date that would make the all-important Christmas holiday sales. Nobody thought the game was ready to ship, but some higher-ups came around and helpfully told us that saying so in a meeting might be a career-limiting move. When the big "go-no-go" meeting was held, everyone voted yes except the QA guy.
All in all, if you go from when a team was put on the game (as opposed to just Rick & Starr), it was around Sept 95 to Sept 97 to make UO. Except that everything made up to the alpha test was thrown away and started over. So it's really more like May 1996 to September 1997…
Vogel also often doesn't get enough credit for putting in place all the vital things that simply were outside of the dev team's scope, like, oh, billing and customer service.
Explaining events later, Rainz describes "I just cast the scroll on the bridge and waited to see what would happen. Someone made the comment 'hehe nice try', I expected to be struck down, instead I heard a loud death grunt as British slumped to his death". Accidentally, this Internet consultant has just committed the most infamous act in gaming history.
I was busy coding something and missed this altogether. Scott rushed into my office and said "Did you see? They killed Lord British!" At first I wasn't sure if he meant in the game or not…
Interestingly UO garnered an initial negative response from the press, Gamespot giving the game 49%
Both UO and SWG had the mixed blessing of being Coaster of the Year from Computer Gaming World… and also winning a variety of "best of" awards.
As time progressed Origin patched, refined, and grew the game in ways they saw fit, adding in a reputation system to calm down the rampant PKing
Somehow the article manages to then skip ahead to 1999, thereby missing out on discussing the great Trammel/Felucca split.
Anyway, a nice walk down memory lane, and perhaps the articles have stuff in them people have not heard before.
