[This autumn, I contacted Iathus of Scara, who served as the Morsulus Herald of the SCA from 1989 through 1992, and asked him some questions about that experience. He graciously wrote up two lengthy documents narrating that history, which I have integrated below in what I believe is chronological order. The spots where I’ve stitched material from those two documents together are marked with “[…]”. I’ve also added a few missing words or bits of context in square brackets, redacted a few mundane names, and tweaked the punctuation and formatting for ease of reading. I am so grateful to Iathus for writing up this history and giving me permission to share it here. — Mathghamhain]
Once upon a time, long ago in a Kingdom not so far away….
Renfield and the Early Armorial Database
[At the start of the 1980s,] Renfield Wander Scribe (West Kingdom in San Jose, CA) was maintaining the LoARs for the Laurel Office and printed an Armorial compiled from the historical LoARs in 1979. Renfield mailed out copies (on 11×17 fan-fold paper) to a few interested SCA Heraldry/computer jocks (“The Working Group”).
When he started running into the limitations of Magnum [the database that had been used for the armorial in the 1970s] and the cost of a commercial Tymshare account. Renfield transferred the Armorial data to a Humongous Hard Drive (512 Meg, as I recall) and started to massage the data with an Apple II. AppleDos had difficulties coping with that much disk-space, so he and Waldt [von Markheim] started exploring CP/M.
Then Waldt […] found a job with a consulting firm, they had a contract in the Denver Office, and he and Alison moved to Caerthe. A certain amount of email and telephone collaboration ensued.
Becoming a Herald, Joining Haus Markheim
I had been a field herald for three or four years, in the Barony of Caerthe, Principality of The Outlands, Kingdom of Atenveldt. I eventually became the Aspen Pursuivant. […]
As Aspen, I helped out folks researching existing Blazon, using the resources of the Denver Public Library, they had a circulation copy of Papworth and a couple of other books (Heraldry for Dummies, basically, except the For Dummies books didn’t appear until several years later.)
Walft and Alison von Markheim moved to Denver from Caid n the summer of ’78, Alison started a bi-monthly heraldry class and submission research meeting. I eventually became part of Haus Markheim and things went from there. Waldt and I both worked with computers, so we talked a lot of shop at the dinner table. Alison and I were book and field heralds, and we also talked shop at the dinner table. Things bled over into each other. A lot.
Moving West, Becoming the Sea Wolf Herald
[Then] the household moved to The West. […] See, there was this snow storm in December of ‘80, starting on the 23rd, and didn’t end until the 7th of January. The snow didn’t get melt off the street before the next storm in the train came through. So, PubWorks just plowed snow off the streets and dumped it into piles in the City parks. The piles were still there for the Mayoral election in early May.
I got tired of re-shoveling the walk after the snow plow made yet another pass up or down the street and shoved the resulting slush/sand/snow mixture off to the side, and there went the path I shoveled out yesterday. I called a friend in California and said, “Dave, you said if I wanted to change jobs, I should talk to you. Where do you want me to send the resume?” I started working for Kaiser the following August 28th. […]
When Haus Markheim moved to the Mists, the then-Sea Wolf looked at three large voices, handed me a copy of the field litany and said “Read this. Go Be Loud. Try not to stumble over names.“ A couple of years later when Gerstan decided to retire, I got drafted [as the next Sea Wolf Herald].
Failure of the “Humongous Hard Drive”
Sometime in late ’80 or early ’81 Renfield’s hard-drive took a hit and quit.
[At] Worldcon in August , in Denver CO, we — Waldt von Markheim, me, Alison von Markheim, Baldwin of Erebor, and Renfield — all sat around in the front parlor, drank Earl Gray tea and talked shop, both Heraldic and Computer. It was eventually decided to try to resurrect the hard-drive and, as Plan B, scan a paper copy of the “Armorial of June, 1979”.
Waldt, Renfield and I (gopher and general extra set of hands) worked on the “dead” hard-drive during the week following WorldCon. Some data was recovered and sent to the Working Group.
A couple of months later Baldwin (Outlands, Las Cruces, NM) produced an ASCII tape of the “Armorial of June 1979”. I started transcribing the LOARs from mid-1979 [to] be ready for update to a digital Armorial.
Working at Kaiser and the “Sys Shire”
During all this time I was working for Kaiser Foundation Northern-California (Hospitals and Medical Facilities) as a system Programmer. This gave me freedom to store “load test data” on the main-frame systems and occasional run a 1000+ page print job. […]
I used IBM mainframes for a couple of reasons: They were handy and I was one of their keepers. IBM main frames are designed to deal [with] large files (You can handle a lot of data with 64-bit addressing, and that is before you start virtualizating the data-space.)
I was part of the Systems Programming and Support Group, and eight of the nine people were in the SCA. (We thought about registering the Shire name as “Sys Shire”, the Sys-Prog security group name was SYSSHR. We had enough people to constitute a Shire, but enough Peers to be at least a Barony — three Knights, a Master at Arms, four Pelicans, three Laurels, several Court Barons, a Count (Count, Count – he refused to be made a Duke), and me with only an A[o]A to my name — I was the token peasant. We had obviously had several multiple peerages….
I don’t remember if we got around to registering as a device “Billety Or and sable, a corner sinister sable” – a stylized 80 column punch-card — or a badge “Barry Vert and Argent” – 11×17 inch green-bar paper. We never did quite figure out how to add the perforation strips down each side (…between pales argent mullety sable…?)
Building the Ordinary
[From 1982 through 1984] I continued transcribing LoARs for the Vesper Herald’s office [and] I started thinking about how to extract an Ordinary from the Armorial file.
[Around 1984] I load[ed] the tape [of the 1979 Armorial] onto a file on my mainframe and took a first cut at a hand index, as a proof of concept. Having an ID in the SYSPROG group was so handy.
I started integrating the LoARs with the tape of the ’79 Armorial, and building them into my index structure.
Late September 1984 [saw] first light for An Ordinary of SCA Heraldry: three copies, plus one for the Laurel office. [It had] limited circulation […] at the Principality of the Mists Heraldry meetings. Lots of proof-reading, editorial commentary, corrections, and re-circulation, rinse and repeat.
Inside the Ordinary Database
[To update the database] I semi-reformatted the LoARs, extracting only the name and blazon data from the mass of commentary, and determined the index terms off of that. The resulting file consisted of Name, “|”, Blazon, “|”, plus one or more indexing terms and that got merged into the Ordinary master file. […] (The “pipe” symbol was chosen because it [did not] occur in the text data.) […]
Printing the Ordinary consisted of a pass through the master building an intermediate file containing index-term, “|”, name, “|”, blazon […] then sorting on index-term, and printing the resultant file. Using a main-frame that could address virtual files measured in giga-bytes made this feasible.
I’d run an extract in the background for three or four hours, and then build a print tape, also as a background task. Then I would clear time on the high-speed printer (300 pages per minute) with the printer folks and print my master copy: “one copy, one-side only, non-collated, drilled, or bound. estimated page count 1,200”. Half an hour later, I’m ready to take me master copy off to my local print shop for a run of “1200 single sided masters, printed duplex, three hole drilled, 200 copies.”
Indexing for the Ordinary
I tried to index by “visual difference”. Primary charge, surrounding charges, secondary charges, line of modification, etc. i.e. “Argent, a shake-fork wavy sable between three martlets; on a chief embattled gules, two towers Argent” would be indexed under “Shake-fork”, “Bird, Martlet”. “Chief”, “Chief, Embattled”, “Tower, Two or More” but not “Field, Argent” or “Chief Gules”.
There weren’t enough shake-forks to make it worth splitting them up in to “Shake-fork, Plain,” and “Shake-fork, Modified.” Any indexing bucket that grew too large became a candidate for splitting into sub-groups (X; X, Two or More…).
One of the internal reports I ran for my own edification was index-term by count. That gave me a feel for when am index-bucket was getting unwieldy, and also gave the design-heralds a list of little used charges that they could suggest.
There are few “field only” blazons, particularly in badges (“Argent, pelletty” or “Checky sable and Or”, for example); but for the most part the field was not indexed separately.
I was using “Chevron”, Chevron, Two”, “Chevron 3 or More”. “Chevron, Inverted”. “Chevron, Wavy”, you get the pattern. I suspect that “Chevron*7” came out of the translation of my files into a “real database”, since the comma is a usual field separator in DB-land (FIND ALL WHERE CHARGE=CHEVRON AND CHEVRON, INVERTED generates “format error, missing value”). And once things like that get baked into the System and get institutionalized, people forget the original motivation. […]
There was a project running for a while to see if the indexing process could be automated. Waldt and Renfield spent some time early on a BNF grammar, (there is an article or two in Vert & Or), but due to the vagaries of the language that went by the way. I gathered it was a chicken-egg problem. You can build a parser for the Grammar of Blazon, but unless the blazon is written in a particularly constrained form, the decision trees get humongous and runtime[s] go exponential.
Separating the Ordinary and Armorial
[In 1985, the] Armorial project went to the East coast (Alban St. Albans, East), a political decision to put a staff position outside of the West Kingdom (according to Eowyn Amber Drake who was on Karina’s staff at the time).
The Eastern job retained the title Morsulus, and I got to pick my title. I settled on Golem Herald. As I recall, “Abacus Herald” was taken and “Computarius” was just right out.
I sent a copy of the June ’79 tape and some twenty-five diskettes of LoAR transcriptions off to Alban in late November(?) of ’85.
I got back a “What the Hell” note. Alban [had] assumed that: a) everything was machine readable by his IBM PC and b) that I was integrating the LoAR updates into the Armorial on a monthly basis. He was perturbed to find that most of the data was on a 1200 foot tape-reel, encoded in ASCII at 200 bpi. He wanted me to reformat the data onto 5 ½ inch floppy disks. I pointed out that 300 or so floppies was not in my budget. Never heard anything more about that. […]
Initially we were both updating different files, and Alban was planning to send me a printed master to use to print the Armorial once or twice a year. It became evident within a few months that there were typos in the two different transcriptions that didn’t correlate. I don’t know if Alban was trying to merge the monthly LoARs from machine-readable media or re-entering the names and blazons (which is what I suspect was happening for the first few months of his tenure). […]
For the first year or so, I was sending my extracted Name/Blazon from the LoARs to Alban. Eventually he got on the electronic-distribution list for the LoARs and I quit sending diskettes.
Distributing the Ordinary
1986 [was] a busy year.
[I took] a trip to the HeraldiCon in Los Angeles (I forget which Urb; Hawthorne perhaps) in March, with 25 copies of the Ordinary. They sold out inside of the first hour. I got a lot of “Why doesn’t this include the Armorial?”, and “I found an error in the Armorial!” and “I need this typo in the Armorial FIXed! Right Now!!” I pointed out that the Office of the Morsulus Herald was responsible for the Armorial side of things, and “since he wasn’t here, this is his email address”.
I mentioned getting asked to make corrections to the Armorial during a hall-way conversation with Mistress Eowyn Amber Drake. That is when she allowed that splitting up the position had been a political decisions and not been well thought out. (“It was a mistake” was the phrase she used.)
More corrections and emendations to the Ordinary data, and a number of changes to the index of terms happened as a result of feed back from HeraldiCon. Splitting up entries with multiple occurrences of a charge (X; “X, two”; “X, three or more”), additional charges that were in Papworth but hadn’t been used in SCA armory, etc.
[I took a] trip to Twenty Year (in June) with 200 copies of the Ordinary in the trailer. Sold out inside of the first day. Orders taken for a second print run when I got back to Oakland. [Then a] trip to Pennsic (in August) with another 200 copies of the Ordinary in the trailer; again sold out the first day.
[After that was the] start of mail-order sales of the Ordinary through Free Trumpet Press [and] quarterly updates. […] [In 1987] I started producing annual updated Ordinaries with the previous years LoARs merged into the body of text.
Requests for Electronic Distribution
[1986 also saw the] first of the requests to furnish machine readable copy of the Ordinary. I punted the request upstream to Laurel. Laurel decided that this was not in the Society’s best interest, so I refused, politely.
[There were repeated] requests/demands that I provide a data-base schema and a data dump of the Ordinary. I eventually had to point out to Folks that the “database” were several flat text files […] currently almost five mega-bytes and growing at 1% per month. If they wanted a structured database format, that would take both clearance from Laurel and a rather lot of dog work, were they volunteering?
[Over the next few years, there was] more rumbling about getting the Ordinary online. Feasibility discussions pointed out that data-base solutions based on IBM PC or Apple machines couldn’t support the volume of data, as I recall there was a four megabyte limit for all of the affordable (read free) software packages, and a commercial license for a real DB system was not in the budget. Besides the software issues, the hardware to support online access would be expensive; server-class machines were running $15K and up at the time.
Reuniting The Ordinary and Armorial
When I became Morsulus [in 1989], I continued to use the Ordinary master-file as the basis for the Armorial, that way the typos became correlated [and only had to be fixed in one place]. [It’s unclear what became of the database that had been maintained by Alban St. Albans.]
In early 1990 Kaiser centralized operations into the computer center in Southern California, as a cost-cutting move. Since I don’t like to breathe air that I can see, I transferred internally and took a position in E-Mail Support, which stayed in the Bay area. Of the 20 people supporting the Kaiser main-frame systems, 19 retired, transferred inside of Kaiser, or quit out outright. Another political decision with adverse short-term effects.
About this time I started to notice burn-out setting in. Iulstan [Sigewealding] was interested in taking the Ordinary into the 20th century, since server hardware was getting very cheap, and there were several Unix database systems could handle the amount of data. So [in] 1992 [the role of Morsulus was] turn[ed] over to Iulstan.
Due to external factors, a divorce among them, I pretty much lost track of the SCA.
I started working as a consultant (Quick Gun for hire, specializing in Perl, Python, Shell, and the Unix Tool Kit. Oh, and I speak Main-Frame as well.)
I eventually retired, caught up on my sleep, started spending more time walking the dog, and generally not doing to much of anything IT-ish.
And then a letter arrived asking me what I remembered about the O&A Project…