Nova Subterra is the first genome for Creatures 2 that not only addresses the problems in the original genome, but expands on it. The Norns of Nova Subterra have four new brain lobes, but only two new chemicals are needed to regulate their bodies. The genome includes a number of changes and advancements, including:
The genome originally began in October 1998, and was developed through February 1999. Originally it began simply as a test to see if the existing genome's design flaws could be improved upon, and then it developed further. Eventually, the testing and breeding that resulted became Project Gemlode, the code name for the genome's development process.
The very first generation of Project Gemlode (which at that time was unnamed) was Andy. Andy had only four changes; one alteration and three new genes. Two of these genes came directly from Cameron, my most likeable C1 Norn ever. A third gene was based on one of those, and the last change was to reduce the output of a tiredness emitter. From those four basic changes came an amiable Norn with slightly improved awareness--a step in the right direction.
The name "Project Gemlode" didn't really come about until an official project page was created, right around the time Carrie (generation 3) was finishing up. The name "Nova Subterra", for the genome itself, breaks down pretty simply: "Nova" for "new", "Subterra" for "underground" (as in, the Norn Underground).
You might consider moving all your other egg files to \Backup, and they'll be there in the unlikely event you'll ever want to use them again. Or, if you feel adventurous, you can even outright delete them. Whatever you do, it's best to get the existing eggs out of the \Egg directory so that the only eggs that appear in the hatchery are Nova Subterra eggs.
While Nova Subterra Norns can interact with other Norns without a problem, they should not be allowed to interbreed. Hybrid Norns will usually be prone to early death due to genetic errors. This genetic incompatibility could not be helped, and was avoided as well as possible. Cyberlife, however, decided that speciation was more important than compatibility.
It should also be noted that interbreeding with Canny Norns is likely to be especially bad. The extra lobes and numerous drive changes in Canny Norns are likely to interfere in some way with the extra lobes in Nova Subterra Norns, and the resulting clash could easily and often result in Norns that die shortly after birth.
Canny Norns have all the same interbreeding problems that Nova Subterra does anyway. This genome is a tad safer to interbreed, in my opinion, as many of the changes are less crucial to the genome, but I wouldn't recommend it. Those changes that are crucial, namely the extra lobes and changes made to support them, need to pass down as a set or not at all in order to provide a viable Norn.
The key things to note in the brain map are:
These lobes are comparable in purpose--but not in operation--to the Canny Norns' Search and Choice lobes. Canny Norns were released while these two lobes were still under development.
Afterdrive and Change Lobes
Afterdrive is sort of an emotion-memory lobe. It keeps pretty close to the status of the drive lobe, but if a drive goes up or down suddenly (if, for example, a Norn eats, plays, or is patted or slapped), the afterdrive lobe won't go up or down with it. Instead, the afterdrive lobe will change slowly to match the drive lobe. During the time it takes for the afterdrive lobe to sync back up, the change lobe can detect the difference in their outputs. The change lobe will spike in output if a drive is reduced suddenly, and the Norn uses this lobe as an additional sense, thus allowing Norns to notice decreases in drives.
The purpose and function of these two lobes is very similar to the Canny Echo lobe, and was basically inspired by the improvement in intelligence that the Echo lobe brought about in Canny Norns. The key difference is that the Canny version needs many genetic changes to work, and this one only needs two (not including the changes to the perception and concept lobes).
If you play close attention to your Norns' biochemistry, you should be aware of two chemicals added to the Nova Subterra genome: SeaASH1 (chemical 190) and SeaASH2 (chemical 191). These are used to regulate the search and sub-attention lobes, and function the same way that DecASH1 and DecASH2 do in the decision lobe. SeaASH1 affects the reward dendrites in the search lobe, and SeaASH2 affects the punishment dendrites in the sub-attention lobe.
Other Genetic Changes
Since the remaining changes are all scattered about the genome, it makes sense just to make a quick list:
Cameron's special emitter was supposed to emit NFP decrease (happiness increase, essentially) whenever an audible event occured, as he was supposed to be a music lover. Instead, however, when I hex-edited the original gene, I made two mistakes. The first mistake was to tie the gene to the wrong kind of event, and this I fixed. The second mistake was that instead of typing out the code for decimal 35 (NFP decrease in Creatures 1), I typed 35 in hexadecimal, which in C1 is chemical 53, ConASH. ConASH is a brain chemical that affects the concept lobe, causing old, useless memories to be discarded in favor of new ones.
When I made Andy, I included this gene as-is. I altered the chemical number to match ConASH in C2, and let it go. And the results were amazing.
Andy's personality may also have been due to one of Cameron's instinct genes, but he was much more aware than an original-genome Norn, which could only have been caused by the ConASH emitter. This emitter may have also helped lessen the effects of OHSS, though the extent of OHSS in Andy was never fully known.
The irony about this emitter is that it's counter-intuitive. It should be causing the Norns to forget things, but it doesn't seem to do that. Even Cyberlife was baffled by this gene; when they saw the Project Gemlode page a few generations later, their genetic engineers commented on this gene specifically, saying that it would likely do no good. And really, they're right in that it shouldn't work; but it does.
The second generation, Becky, had a serious intelligence problem. (Becky was also a wall-bonker, because I screwed up the stimulus gene to prevent it, rewarding her for bonking walls. This was fixed in Carrie.) The main gene I was interested in in Becky was straight from Robert Reid's genome: an expanded concept lobe 65% larger than normal. Robert Reid had a lot of success with that gene, but I didn't. Becky was an idiot, and OHSS set in at half an hour old.
This was the key to realizing that OHSS was in the concept lobe, however. Chris Double later pointed out that this was probably too large a size to use, but that some increase in the concept lobe's size would still be a good idea. So I enlarged the lobe by only 35% in Carrie, and that did the trick. From then on, OHSS was cured.
A Project is Born
Carrie's tests were short and sweet: Carrie was far superior to both Becky and Andy in every way. Unlike the first two generations, there was no real difference in intelligence between male and female. (In Andy, the males were smarter. In Becky, males were dumber, but then "dumber" in Becky was like deciding whether Microsoft or Disney is wealthier than the other.)
But best of all, OHSS was cured, and the genome suddenly looked more promising than ever before. Thus, I started coming up with names for the genome. SteerPike suggested "Subterra", which led to "Nova Subterra". (And fortunately so, since my other names were all bad.) From there I decided to make an official project out of it, as I hadn't done since the first days of the Norn Underground. So as Carrie finished testing, I created the official Project Gemlode page, the name Gemlode being chosen for its obvious connection to my site's theme.
Visitors to the project page (of which there were few, as the URL was kept secret at the time) saw regular updates on the status of the genome, the current generation and its results, and plans for the next generation. Details of the genome's construction were there.
The Story of Dan
Dan, the fourth generation, was exactly what I'd wanted to avoid in Project Gemlode. It wasn't that he was a failure, but that so many of his early versions were. Knowing that the new lobes, search and sub-attention (and, early on, a third lobe called Object Memory), would be difficult to get right on the first or likely the fifth try, I said Dan was in a volatile testing stage--he would have several prototypes tested before a final version was chosen. However, most of the early models were spectacular failures due to my own incomplete knowledge of the brain. To make matters worse, around that time a lot of the people testing the genome were quite busy.
It took five different tries at a search lobe model before one was finally chosen as the best way of doing it. And in the process, fifteen sub-genomes of Dan were made: Dan1a, Dan1b, Dan1c... all the way up through Dan1o, which was chosen as the final Dan genome.
The volatile testing stage took four weeks; and after that, Dan still had to go through regular testing like every other generation. In the end, the result was fairly good, but it had stalled out the project for so long that it cast doubts on the idea of this particular search lobe model working. Not until Ellie were the new lobes shown to be pretty good, and even then the benefits to Norn intelligence were doubtful until Finn appeared, and we got to see two lobe sets in action together.
Although Emy was a little disappointed that generation five wasn't called "Emily" (sorry, Emy), and I got regular chiding for it, Ellie went off without a hitch. She compared well to other known genomes. Canny results were coming in that showed that, as I suspected, they still suffered from One-Hour Stupidity Syndrome, but it was masked a tad by their increased intelligence.
A pattern was noticed in Project Gemlode that the odd generations were all better--Andy and Carrie being prime specimens vs. Becky and Dan. Ellie was no exception due to her quick success, and to some improvements on Carrie. In hopes of breaking the even-gen curse, I toyed with new brain ideas; I wanted to add a form of sequential memory to Norns, which ought to create something of a rocket scientist among the species, but that idea was rejected the more it began to look like another Dan all over again.
At about this time, rumors began to fly--intentionally. The secret of Project Gemlode was being kept too well, so the acronym "PG" was allowed to slip, and spread in a most Machiavellian fashion. It wasn't until Finn was in testing that the full name of the project became known.
Of all the changes in Nova Subterra, the afterdrive lobe is one of the ones I'm most proud of. The goal, when I set out to create Finn, was to further increase Norn intelligence; the Canny Echo lobe looked like it worked well in enhancing intelligence, and for good reason: It's a known principle in neural nets that providing more views of the same input data (in this case drives and emotions) improves the quality of output and decision-making processes. So, I wanted to do the same thing. At first I toyed with the idea of implementing Canny reduction chemicals on a smaller scale--only about six drives' worth. In the end I threw that idea out and looked at other methods. I'd thought of making the Sandrabellum perceptible, but really, what Norn needs to know how its insulin and protein levels are doing?
Somewhere in there, I got the idea that something similar (but far from identical) to the Canny Echo lobe could be made by hooking two lobes together. Since I wanted Norns to notice drive reductions, I made one lobe that kept track of drive status but wouldn't change right away, and another that would notice any discrepancy between the two. Best of all, I could do it without any other changes besides making the perception and concept lobes a little bigger to handle the input. No receptors needed, no reactions, no new chemicals. And the lobes were so easy to make that they worked on the first try, no volatile testing required.
Results met my expectations: Finn broke the even-gen curse, and Project Gemlode was ready to move on to the final generation. The rumor mill had begun to run full-tilt, and I prepared for the release; this document was being written as Gwen just began testing.
Gwen's testing was coming along nicely, Finn's anger problem clearly solved, when it finally came out that the other problem from Finn--the fact that the Norns didn't mate as well--was no better despite several changes to that end. This news came on February 11, 1999, nine days before the planned release--hardly adequate time to start a new generation.
In desperation, I pushed Gwen into a sudden volatile testing stage, and within two hours I'd rattled off four new sub-genomes, three of which were tested. Mike and Sarah, who had reported that the mating was still a problem (they were the first to get Gwen to mating age, and had had large success with other genomes), were also kind enough to test Gwen2, Gwen3, and Gwen5. Just as I was about to update the Project Gemlode page with the news that Gwen was in volatile testing and generation H was in standby status, the news came in that a Gwen5 was pregnant.
So, a potential disaster was averted. Gwen was moved back into stable testing, with Gwen5 chosen as the new genome to test. The genome was back on track for release.... or so I thought.
Wednesday, February 17, three days away from release: Julianne's report came in, with the news that Gwen5 wouldn't breed. There had only been two eggs produced by Gwen5s in the entire testing, and none in a wolfling run, so Gwen was pushed back into a volatile stage. It took a while to get a good breeding pair, and I ended up having to pull out all the stops; Gwen10 bred very well, but as a result traveling skills were ruined. So more work was done from that point to restore a balance between traveling and mating.
The night of February 19, less than 24 hours to release, Gwen16 was developed, and it was Gwen16 that ultimately proved to be the best of both worlds. In the process, however, she had more sub-genomes than Dan (who only had fifteen), albeit developed and tested far more quickly. Gwen16's success was verified by Mike Raiford, Emy, and AntiNorn; and Denise Voskuil, dutifully testing all the genomes to that point, suffered a system crash for Gwen's sake. But in the end, Gwen16 was a success.
Q: How many generations did it take to
develop these Norns?
A: Seven. The generations, in order, were Andy, Becky, Carrie, Dan, Ellie, Finn, and Gwen. Dan had to go through about fifteen sub-genomes before he could move on, and Gwen had sixteen, so in all, 36 different GEN files were made and tested for the project.
Q: Aren't you making C2 too easy by
making this genome?
A: I don't think so, no. These Norns are just as capable of injury or suffering disease as any others, and have a lot of the same frailty. While they were made tougher for pregnancy, and they're not as likely to starve to death or jump off a cliff, this isn't so much making it easy as it is making it more reasonable. A Norn should have a decent chance of surviving on its own, shine when attended to, and die when its time comes. But beyond that, things like constantly high fertility and immortality have never interested me in Norns; now those would be making it too easy.
Q: Why won't Nova Subterra Norns
interbreed well with other kinds of Norns?
A: Despite my feelings to the contrary, Cyberlife believes strongly that the process of speciation is somehow important to Creatures and Creatures 2. This makes sense to an extent, but not, I think, to that extent. Nova Subterra was intended to be compatible with other genomes, and I went to great lengths to keep it that way, but in the end it turned out to be impossible. The real problems come from the extra lobes and their support genes, which are unfortunately handled by the game engine in the same manner as things like reactions and poses (which are far less critical--if only half make it into a child, no big deal; not so with brain-crucial receptors and the lobe connections themselves). It may be possible to get stable, viable Norns that are hybrids of the original genome, but it's doubtful; and hybrids with other third-party genomes are bound to be disastrous.
Q: What causes OHSS in regular
A: The concept lobe is too small for the size of the perception lobe. Increasing concept lobe size slightly cured OHSS, but increasing it too much was found to be a bad thing.
Q: Both Canny Norns and Nova Subterra
Norns have a lobe called Search. Is Nova Subterra's lobe based on the Canny
A: No. The search lobe was implemented in Dan, who was still early in the volatile testing stage when Canny Norns were released. Until that time, I had no idea that Lis' genome would be using such a concept (much less the same name). When I learned about it, I discovered that our versions of the search lobe were completely different.
Q: What about the Change and Afterdrive
lobes? Are they based on the Canny Echo lobe?
A: In concept, yes, but not in design. After Ellie completed testing, there seemed to be a certain mental gap between Project Gemlode's Norns and the Canny genome, despite Gemlode having the cure for OHSS already in hand (i.e., the Canny Norns were smarter). The principle behind the Canny Echo lobe was good--more useful details lead to smarter Norns--but I didn't like the implementation (no offense, Lis). As I considered adding seventeen chemicals and the reactions and receptors needed to use them impractical (not to mention unoriginal since Canny Norns had already gone that route), I looked for an alternative means of producing a similar effect. This I finally realized when I dreamed up the afterdrive lobe, with which I could perform much the same feat with only two lobes, and no support genes to speak of. Thus Finn, the next generation, was incredibly bright.
Q: Is there anything special I need to
know about the new chemicals in Nova Subterra?
A: Not really. If you understand how DecASH1 and DecASH2 work, it suffices to say that SeaASH1 and SeaASH2 (chemicals 190 and 191) work the same way, but for attention instead of decisions. I would have avoided using them altogether if they weren't necessary to operate the search and sub-attention lobes properly. They're certainly nothing that you have to monitor for your Norns' overall health.
Q: Were the different generations of
Project Gemlode named for anybody in particular?
A: Not for any actual person, no. Finn, however, was named for a character in William Gibson's classic book Neuromancer. All the other names were chosen at random, because they sounded good.
Q: Are there any plans for a C1 version
of Nova Subterra?
A: No. While some of the genes are certainly possible to include in C1, others aren't. Especially impossible would be the new lobes; C1 apparently doesn't recognize any lobes past ten, and there are nine to begin with there. So there's room for only one more lobe, and it takes two to accomplish either of Nova Subterra's extra brain functions.
Q: Will there be any Grendels or Ettins
based on this genome?
A: Not from me, there won't. I know too little about Grendel and Ettin DNA, and have too little interest in either. But, I wouldn't mind a bit if somebody else wanted to make them.
Q: Is it all right to make genetic
variations of Nova Subterra?
A: Absolutely. I'd like to encourage any genetic engineering project based on the Nova Subterra genome; feel free to change away. It could be that somebody will develop the smartest Norn ever as a result of such a project, and that's a good thing for everybody.
Q: Is anyone allowed to post Nova
Subterra on their page?
A: Not the original files, no--only Norns derived from the genome, such as second generation Norns, etc. The original files can be posted only on the Norn Underground and the sites of other users who helped in the project or were somehow involved in another way--those whose names are listed in the credits. (Also, I wouldn't mind Cyberlife themselves posting the genome online.)
Q: Did AntiNorn torture any of the
Norns he was testing?
A: (I knew someone was bound to ask.) I didn't, and still don't, consider it any of my business. However, I can say that AntiNorn made excellent observations of the Norns that were raised normally or in wolfling runs. He was one of the most helpful testers I had, and actually helped raise the quality of all Norndom. Since it's been said often that C2 Norns torture themselves, participating in Project Gemlode was really a change of pace. If any Norns were tortured on the side, it matters not at all to me; even if they were, I never heard about it.
Just as he helped in the development of Lis Morris's Canny Norns, Chris Double was even more vital to Nova Subterra. The credit for curing OHSS properly belongs to him, as it was his informed suggestion that proved key to fixing it. Moreover, Chris was of invaluable assistance in developing the search and sub-attention lobes later in the Dan generation. And to top it all off, his genome injector was used for injecting every test generation.
Shortly after Andy was created, Robert Reid contacted me out of the blue and offered me the use of parts of his genome for my own purposes. His help was gratefully accepted, and as a result several genes from his project found their way (some in an altered form) into Becky, the next generation. These changes helped curb wall-bonking (eventually) and improved pregnancies. Moreover, his idea of increasing the concept lobe size--even though it turned out to be too large an increase, as Chris Double later observed--led to the eventual discovery of a cure for OHSS.
Julianne and SteerPike
SteerPike was the first person to test out the genome under development, and helped name the genome. His results were especially helpful in the early generations. Julianne's detailed reports for every generation were essential, in many cases serving as the deciding factor on what changes to make for the next generation; she deserves special recognition as such.
Emy and Mike Raiford
Emy and Mike tested out the genome, but they were especially helpful in the rapid-fire volatile testing of Gwen, the final generation of Project Gemlode, and I couldn't very well credit them without a special thank-you after their superb (and quick) work. To be fair, Denise Voskuil and AntiNorn had no small part in Gwen's volatile testing, but Emy and Mike were the ones who at times had to deal with testing four or five genomes over the course of half an hour. (Boy, did I feel guilty every time I'd ship Emy the new Gwen, all the way from Gwen2 to Gwen16; she pretty much tested them all.)
NORNGod, AntiNorn, Cave, Denise Voskuil, and SparkleNorn
For performing a great deal of the testing. Their analyses helped in no small way to push the project forward, even when other testers were busy.
Miff, Kinnison, and mumblebox
For keeping the secret of Project Gemlode so well (and in some cases, helping to keep it badly) while it went on. Mumblebox and Miff, in particular, were of special help in turning the rumor mill prior to the genome's release.