Code center > Suggestions
Eventual improvements to DNA language
Numsgil:
Yes, you can do that. That's just one technique though, and is highly repetitive, and, as you said, real groups of genes aren't controlled by a single hormone.
In real eukaryotic cells, regulation can occur at any of these steps:
In DB, we only have the one level. That's find for modelling simple orgnanisms such as bacteria, but more complex multibots become very burdensome. Attempting to specialize cells into arms, etc. becomes the hardest part, not because the action itself is hard, but because keeping all the conditions consistant is difficult. While I could create simple DNA writing macros that do all the same thing as Noble genes, there's a reason you'd want the macros in the first place.
My philosophy is that if there's something we'd like to have in the DNA language, the bots should have access to it as well. More parts, more complexity. Otherwise we short change the DNA and mutations. Right now we have a lot of advantages over mutations because we know which sysvar A to stick into which sysvar B, how often to use C, etc. Mutations don't have that luxury.
Sleep, Noble genes, and select all came from my attempts at writing a multibot. There's definately need, because it cuts down on repetitive actions and frees the programmer to concentrate on form instead of timing and gene order.
If you can make fire by beating two rocks together, that's great. If I give you a box of matches you can still bang two rocks together to make fire. But if you're smart you'll use the matches and move onto something else (like a wheel!).
That said, there are reasons you may want to run a simpler simulation. I'll probably add something in the options panel that lets you turn on or off the functionality of some DNA commands.
shvarz:
OK, I looked through the links. Basic stuff. BTW, the idea of "Master genes" does not seem to hold under the current data. In majority of cases "Master genes" turned out to be wishful thinking.
I don't agree that we have only one regulation step. You can have tons of regulation:
1. Genes can check for flags set by other genes
2. Genes can check for combinations of flags set by other genes.
3. Genes can use values supplied by other genes.
One can create a very complex regulatory system using our current DNA code. In fact, it can be even more complex than the regulation of eukaryotic genes. When you look at that picture you inserted, remember that all those regulation steps are essentially genes regulating genes. It is very complex, but it is still based on simple rules. And these rules are the same for all genes in an organism.
I understand your frustration with trying to create complex bots. But that's the fun part - if it was easy then there would be no challenge. I am all for adding new commands when they allow something that was not possible to do without them. The "sleep" command seems to be a good example. Without that command it is impossible to achieve the needed effect. But nested conditions are possible. So we should not take shortcuts.
I don't think bots have to have the tools for programming that we have. They don't get confused, they don't make mistakes, they don't feel frustrated and they don't spend their valuable time creating complex DNA. So, I think it is fare for you to create these tools as long as the final result confirms to the standard language.
Numsgil:
I put up a poll, vote! I'm implementing that particular part of it now.
PurpleYouko:
I think assuming a stop whenever it encounters a cond would be great.
Seems much simpler than requiring one.
Carlo:
I just voted for eliminating the stop at the end of the gene. It is useless. Ah, please, Nums, just eliminate the other little stupid thing of the language, that is, the "end" at the end of the dna.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version