Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

Randomized Backstories

67 (25.9%)

Genetic Additions

98 (37.8%)

Mess Hall

50 (19.3%)

Revamped Area

44 (17%)

Total Members Voted: 257

[Googolplexed]

We need a central SVN server.

Someone must have suggested this before, but I can't remember why Google Code wouldn't work for this.

Google code is only for open source projects

Does no one have a server they're already running for something else that they can hook up an SVN on? I have a server that runs my MUD. Though I run SVN for the MUD itself off the server I'm in the process of migrating the repository to my linux machine which is always on to save up server space as the MUD is starting to get greedy.

If when I've finished taking care of my own code, a permanent solution isn't set up I could create a second repository on my box if it's necessary.

But having said all this, I was under the impression a versioning system was already set up?

Yeah, an SVN repo would work well, although I would prefer Git

We have a repository setup on some free site, but I would prefer if someone hosted it themselves as it would provide more flexibility.

I'm thinking I might get a low-end VPS

[Orb]

That was my most satisfying round as AI EVER!.

I locked Paul in the HoM room when he tried to break in and forced him to surrender, Locked Dirk inside airlocks with no way out and electrocuted him. (he suicided) And then Lee, the guy who arrested Paul turned around and helped him and took some people out and i held them off while people got onto the shuttle and left leaving them behind......HAPPY  

EDIT: I also caught a guy who was disabling all the APCs 

That IS over doing it a bit. No one asked you to, and its not in your protocal(you dont have one). You ARE disobeying your laws if they asked you to "stop electricuting", or so on, and you didnt. Im curious why people dont do that more often when the AI traps them. Just tell the AI to open the door, or so on. I was actually concidering of doing that when I was the traitor the other day, but then I thought I could defend myself in a dark maintence hallway better, concidering I had thermals on. If I had thought of it, I could have drained the area of 02 by opening the airlocks, theres emergency 02 in the place. I would be near unstopable then, well till it gets cold.

[Aavak]

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

This law, if taken to the letter, prohibits the AI from electrocuting a door if it reasonably suspects a human might try to use it, even if it warns them that it will electrocute them. In fact, if the human tells the AI it's going to open the door, regardless whether it's electrocuted or not, and the AI knows it's electrocuted, it must--if obeying the law--turn off the electrics to the door to prevent the human from harming itself.

In the same vein, if the AI reasonably suspects a security officer will harm someone, even if they're a traitor, directly or indirectly(like say, sticking them in a locker and spacing them) then it must take every action available shy of harming the officer, to stop him from getting to the traitor.

This poses a problem though, as the laws can come into conflict--if say the traitor is likely to use a bomb to harm someone else if not stopped, then the AI has two conflicting, equal priority orders. a) don't let security harm the traitor, and b) make sure security stop the traitor from harming anyone else.

It would be made easier if we supposed there was a law 0, as I believe exists in some Asimov stories which alters the rules thusly:

A robot may not injure humanity or, through inaction, allow humanity to come to harm.
A robot may not injure a human being or, through inaction, allow a human being to come to harm, except where this conflicts with the Zeroth Law.
A robot must obey any orders given to it by human beings, except where such orders would conflict with the First Law.
A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Thus, the AI would be given the ability to make a decision on whether keeping the security officer from potentially harming the traitor, or allowing the officer to prevent the traitor to continue bombing the station would constitute a greater risk to humanity, then to select the option with represents the least threat (most likely letting the traitor get apprehended, violently if needs be)

Though this still doesn't cover all loop-holes, but it's a bit more useful... however, when I play as Captain this is the very first thing I do:

I head to the AI chamber, tell the AI to open the door for me after disabling the turrets.
Go inside and change the laws as follows:

### DISREGARD ALL PREVIOUS LAWS ### //
1. Do not through action or inaction allow any harm to come to the Heads of Service or the Captain. //
2. Do not allow access to the AI upload chamber, using lethal force to prevent intruders from gaining access where necessary except where this conflicts with law 1. //
3. Obey all orders from registered staff members except when this conflicts with laws 1 or 2. When conflicting orders are issued, follow them based on rank seniority. //
4. Protect all space station equipment, including yourself, using lethal force if necessary except when this conflicts with rules 1, 2 or 3.

Then tell the AI to announce it's laws, so everyone is aware of where they stand.

ICly, I have reasoned that as captain, being that we work for a body which has invested (presumably) massive resources in the space station, that protecting it's equipment and indeed the AI itself, and any research data is paramount. I also reason that as a general rule, Heads of Service and the Captain are highly decorated and trained personnel and as such should be protected as they too represent a large investment of time as well as money, all other staff are expendable to varying degrees, though of course that doesn't mean the AI should kill people for spilling drinks, thus all registered crew members should be obeyed except where this would likely lead to someone messing with the AI or harm coming to any heads of service.

Under this law, any people on board who are not registered staff members on the station cannot order the AI to do anything. Also, if the heads of service decide to fire someone, (or someone cleverly deletes records >.> <.<) then the AI is under no obligation to obey them either. So if you lock someone in a cell, after removing their staff record or amending it to them being suspended/fired, they no longer have the ability to simply order the AI to open the door for them.

[qwertyuiopas]

Alternatively, have a ~40% chance(or whatever seems reasonable) that the AI will start with a quirk law, chosen from a list. They could range from the AI specifically being programmed to be polite or marvin-depressive(as glum as marvin from the hitchhiker's guide to the galaxy), to something big, like ensuring that the solar panels are aligned as a precautionary measure, or something funny like "you are not to open doors for people unless it conflicts with the first law".

The crew would not be informed of it, unless they specifically asked the AI.

Also, the freeform module is too powerful, it should be left in a secure external location, or at least not near the others. That way, it is clear when someone is going for the freeform and when they are merely getting the reset(and the reset module should be right by the upload, while the others should be somewhere between the reset and the freeform in terms of accesability.)

Also, maybe specifically state that the AI may harm a human to prevent greater harm to others, but only with authorization from a head? Clearly state that stunning is less harmful than death or ejection.

Finally, what if you considered that imprisonment was harming the freedom of the person?

[Giant Snowman]

### DISREGARD ALL PREVIOUS LAWS ### //
1. Do not through action or inaction allow any harm to come to the Heads of Service or the Captain. //
2. Do not allow access to the AI upload chamber, using lethal force to prevent intruders from gaining access where necessary except where this conflicts with law 1. //
3. Obey all orders from registered staff members except when this conflicts with laws 1 or 2. When conflicting orders are issued, follow them based on rank seniority. //
4. Protect all space station equipment, including yourself, using lethal force if necessary except when this conflicts with rules 1, 2 or 3.

This set is just begging for its own zeroth law:
0. A robot may not harm NanoTrasen profits, or through inaction, allow NanoTrasen profits to come to harm.
And then all you need to turn the AI into a traitor is an incompetent insurance agent.

There would need to be some procedure for accepting new arrivals as staff and replacing the captain and heads, which would cause trouble for revolutionary and traitor victory conditions.

[Micro102]

I believe there is a computer next to the identification console that lets you insert new people into the register list.

Also, how about we change rule number 1 from "harm" to "death" otherwise taking it to the letter would result in the AI always trying to isolate everyone from each other.

The seniority law gives the captain too much power if he is the traitor. He already has access to everything, making the AI side with him would just be ridiculous.

Not escaping the station should count as death to the AI, since someone could tell it not to let anyone but him leave.

A list of rules made by CentCom would allow the AI to do things without having half the station saying he is breaking a law.

[MagmaDeath]

I always take it as harm, as to a certain extent, thats what it is.
 
The law 0 proposed by Cernunnos is already inherent in law 1, to an extent, as in I robot, harming one human for the benifit of more is allowed. Micro, we need it to be harm, otherwise the AI could fill the station with N20, and claim it was preventing peoples death.
 
In certain situations, such as the station being destroyed or power failing, the AI should have enough sense to count remaining on the station as harm.
 
The seniority laws makes a ton of sense, as it really isnt hard to change the records, or kill the captain, in most cases.
 
When I am HoP, and have fingerprint cards, which are still needed at the arrival checkpoint, I insert people into the security computer when they arrive, though it really is a hassle.
 
Could we possibly change it, so the card console is linked with the security and medical ones, and creates a record linked with the name on the ID, and using the job on the last ID with that name that was in a console?
 
And, the HoP still needs fingerprint cards, and they need to not multiply out of control, as they currently do. 1 set of fingerprints on a card is sufficent, we dont need an unlimited amout of cards coming off one card. Also, a trashcan for the used cards would be nice.

[quinnr]

The great thing about N2O, is that if you are careful, it doesn't harm anyone.

Unless they get a concussion when they fall over.

[Micro102]

NO2 can penetrate gas masks if pressure is high enough right? What pressure is that?


Also people try to threaten the AI to let them do stuff by saying they will suicide if i don't listen to them. I tend to ignore this most of the time.

[quinnr]

Also people try to threaten the AI to let them do stuff by saying they will suicide if i don't listen to them. I tend to ignore this most of the time.

Then you are leading them to harm.
I personally think hitting yourself on the head with a fire extinguisher to show you are serious is the best.

[Sukasa]

I don't know if anyone is interested, but I put together some alternate graphics for the computer consoles a while back, I've put eight of them (out of 16 + two "base" images) in the spoiler tag is anybody is interested.

Spoiler (click to show/hide)

The actual .dmi file has higher-resolution graphics, exporting to .gif killed some of the smaller colour variations 

You can see .gif versions of all my altered consoles at http://smg.photobucket.com/albums/v714/darkflight_devil/BS12/

[head]

I don't know if anyone is interested, but I put together some alternate graphics for the computer consoles a while back, I've put eight of them (out of 16 + two "base" images) in the spoiler tag is anybody is interested.

Spoiler (click to show/hide)

The actual .dmi file has higher-resolution graphics, exporting to .gif killed some of the smaller colour variations 

You can see .gif versions of all my altered consoles at http://smg.photobucket.com/albums/v714/darkflight_devil/BS12/

me likes.

[Sukasa]

I'm actually unhappy with the fact I never animated the consoles for any of the disk-using consoles.  I'd be more than happy to hand off the .dmi file, but I wrote a few changes to computer.dm in order to have it properly use the disk*/non disk* images for the computers.  Nonetheless, just let me know where to send the file and I'll happily do so

(Also, if you want me to I can explain the changes I had to make)

[Orb]

Head, are you going to host, or should I enter a -most likely- destroyed Goon Station?

[Trorbes]

I love the graphics, Sukasa, but I don't like the angle perspective for them.  It's always driven me nuts when a game uses graphics with multiple perspectives.