Author Topic: The randomizer  (Read 12202 times)

Offline spike43884

  • Bot Overlord
  • ****
  • Posts: 656
    • View Profile
Re: The randomizer
« Reply #15 on: November 27, 2014, 01:56:19 PM »
Yea looks like you have serious case of the pride and the fall going on.  :)

I don't claim to know anything, and I've shared ALL of my code on these forums, to get fixed, or to aid others.
And I might link to that post you linked to a few comments ago, Because I won't bother with the SAME long reply again http://forum.darwinbots.com/index.php/topic,6656.0.html
Autism can allow so much joy, and at the same time sadness to be seen. Our world is weird, and full of contradiction everywhere, yet somehow at moments seems to come together, and make near perfect sense.

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
Re: The randomizer
« Reply #16 on: December 02, 2014, 09:30:52 PM »
Not until I actually care to explain.

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
Re: The randomizer
« Reply #17 on: November 01, 2015, 05:17:57 PM »
There is ALOT of stuff in DB2 that uses random numbers. Everything from shots to the way the DNA mutates to weather patterns. At first I was grabbing random numbers from a service similar to random.org but had a free API. It worked ok but a bit slow. I wanted a way to use more common data sitting on the internet. I ended up compressing pictures that I get from a set of URLs using my portable hack of 7z and feeding that trough the randomizer. Most of the code is in the beginning of MDIform1 if anyone wants to take a look. I needed a place to host the inet control, that is why it is in a form instead of a module.

Offline Testlund

  • Bot God
  • *****
  • Posts: 1574
    • View Profile
Re: The randomizer
« Reply #18 on: March 28, 2016, 08:04:01 PM »
So is the Randomizer finally working now?
Has anybody seen any interesting result when using this?
Is the randomizer only running the first time when you start DB?
What are all those .bin files for?
What's the maximum simulation size that the Randomizer can handle?
« Last Edit: March 28, 2016, 08:34:32 PM by Testlund »
The internet is corrupt and controlled by criminally minded people.

Offline Testlund

  • Bot God
  • *****
  • Posts: 1574
    • View Profile
Re: The randomizer
« Reply #19 on: March 29, 2016, 07:39:59 AM »
Is the randomizer only running the first time when you start DB?

This question I can answer myself. The Randomizer will try to connect every once in awhile and if the computer is not connected it will cause DB to freeze permanently.  :glare:
The internet is corrupt and controlled by criminally minded people.

Offline Botsareus

  • Society makes it all backwards - there is a good reason for that
  • Bot God
  • *****
  • Posts: 4483
    • View Profile
Re: The randomizer
« Reply #20 on: March 29, 2016, 05:30:28 PM »
Quote
The Randomizer will try to connect every once in awhile and if the computer is not connected it will cause DB to freeze permanently.

I was hoping users will figure out that if it says "seeding randomizer" that they need an active internet connection. As soon as you reestablish an internet connection the program should resume. Remember this is an alpha feature, and probably will remain in alpha as after I have a way to track bugs better I will probably switch to a simple scheme again. (read more below) With the simpler randomness source I will be able to figure out if there is an internet connection easier because it is a simple html request.
Quote
So is the Randomizer finally working now?

The randomizer was always working really. I just thought I'd like to experiment with it. I initially used a "Quantum Random Number Server" for it. That worked well, however, I was not able to run it in safe mode because safemode needs consistent data to be able to recreate a sequence of events within a simulation. I am using bytes from compressed pictures instead.

Let me also note that I am getting picture data from all kinds of websites, some not as clean as others. I did not really care because it is just data. But a simpler scheme will definitely fix this. Once again I explain that this is a really early alpha feature, just a little experiment I had going (this leads to my next pointer)

Quote
Has anybody seen any interesting result when using this?

The results are definitely "more interesting" but a little less "random" so generally it takes longer for robots to evolve, but when they do, the mutation is way more interesting imo. My zerobot was evolved using this randomizer. Also, I am evolving my zerobot further with my survival fitness function also using this randomizer.

However, I had to stop evolving after finding some bugs and improvements that will be published with 2.48.29 soon.

Quote
What are all those .bin files for?

The program collects a bunch of pictures from the internet, compresses them, and splits the data into separate .bin files. Then the randomizer reads one file at a time converting the bytes within the file into randomizer seeds. Safemode is also hooked to this process. The idea is if a crash happens, it will be related to the seeds in a given .bin file. So to recreate a bug I will need the .bin file.

Quote
What's the maximum simulation size that the Randomizer can handle?

Nothing to do with simulation sizes really. Any size.



I am also planning to restart PeterIM. All I really need to figure out is setting up a server and having the program draw nice pretty charts in 1 hr , 1 day and 1 week intervals. I also need access to a security team just incase. So the best time to do it will be in a computer club or from assistance from a professor who has a server to spear. So it is definitely something I am looking forward to doing when I actually get to a uni. Right now I am just studying to retake the "American College Test" I bought the pre-"common-core" version, so hopefully when I take the actual test it will be easier than the test prep.

If I do not make it to the uni. I will probably end up pestering Numsgil to host some kind of IM. Hopefully he will have enough patience left for me to help me out.  :redface:

So yea, as I expected fixing IM is definitely on my radar, and most likely the last feature I will contribute to the vb6 version of DB2. The sad thing is DB2 will probably never really make it out of Beta, vb6 is a difficult language to track complex bugs in. I am embarrassed, but it really did seriously burn me out.



I have some ideas in ways of balancing chloroplasts so there are predictor prey cycles that do not take 10 years to happen. But that is a research topic on its own. I think I gathered enough research here as is to publish at least 2, maybe 3 papers. So if I really ever consider seriously adding brand new features to Darwinbots, it will not happen under vb6.
« Last Edit: March 30, 2016, 10:35:05 AM by Botsareus »