View Single Post
  #22  
Old June 26th, 2010, 03:38 AM
OjnoTheRed's Avatar
OjnoTheRed OjnoTheRed is offline
Citizen: SOC-13
 
Join Date: Dec 2009
Location: Melbourne
Posts: 765
Gallery : 17
Visit OjnoTheRed's Blog
OjnoTheRed Citizen+OjnoTheRed Citizen+
Default

E.D. Quibell and Seriayne, thanks heaps for the replies.

If there is a draft copy of DGP's proposed integration of Robot brains into MT Craft Design and it can be distributed, obviously I'd love a copy as well. Would perfect this thread!

Aramis above also mentioned the 250 CP / point of INT, thanks for finding the reference. I got out my copy of 101 vehicles and found it as well, finally - had it all the time, of course.

But I'm not sure I can agree we can assign that many CP to a robot brain. DGP's reasoning is this (quoting from 101 vehicles):

Notice that if 250 CP is intelligence 1, then a TL16, Model 11 computer (CP multiplier of 200) is almost intelligence 1. A TL17, Model 12 computer (CP Multiple of 1,000) is intelligence 4. Thus, you can see, that by TL17 starship computers become truly intelligent.


The problem with this statement is two-fold. Firstly, INT and EDU ratings for robot brains occur from TL11 - that is, we are not saying that they are self-aware and sentient, but that for game purposes we can rate INT and EDU for skills and tasks. This makes using INT points as a multiplier by comparing them to a TL16 computer problematic because the difference when we get to TL16 is not the processing power so much as true sentience comes into being.

The second problem is that a TL16 Model 12 computer with one Holographic Linked control panel comes in at 20.02 kL, 5.01 tonnes, MCr43.0005, and produces 300 control points, rated at INT 1.

A TL12 robot brain with 5 Parallel CPU's and 11 Standard Storage units comes in at 0.008 kL, 0.0011 tonnes, MCr0.057250 and is also rated at INT 1 and EDU 1 - and hence produces 250 control points.

For controlling any craft of any size, it would be insane not to ditch on-board computers by TL12 or TL13 and replace them with robot brains in nearly every case - cheaper, lighter, more compact and consume less power by an order of more than 1,000 times. I think this distorts the intended spirit of the design system.

I think the way around this to take the view that robot brains being highly specialised decision makers need electronics to interface with the rest of the craft, but do not need the volume requirements indicated by holographic linked controls - a heavy discount needs to be applied for the ability to directly interface with the brain, not through displays, keyboards, or the need to customise controls, etc. etc. I also think that while brains should count as having a CP multiplier, 250 is too much. The computing power in the volumes and masses represented on the Craft Computer table represents all of the networked cabling and server power needed to co-ordinate the vaaast amounts of control for a craft (especially a starship).

But having said all of that, the reverse problem is why not build starships with built in robot brains that pilot, navigate, and control weapons? In my view, you can, but the pay-off is cost and (indirectly) weight: hence the x5 price multiplier and x2 mass multiplier to the hull for any robot brain being installed. That means we can create giant 10,000 displacement ton unmanned robots armed with Meson Spinal weapons, intelligence 15, education 15, and Gunnery-5, Pilot-5 but they are going to be enormously expensive. Or we can instead built our standard battle riders but crew them with robots that interface with the computer - much less expensive and more flexible (the robots can bail along with sentients - gee, there's a good idea for a story, two robots bail in a lifepod carrying information important for carrying on the fight against an Imperial government ...).

This goes to the heart of "what is a robot?". I think it is a craft with a robot brain that as a consequence can be given orders and operate independently of sentient control consistent with those orders.
Reply With Quote