Construction and Fabrication is how most of the iMorph and Clanking Masses indentures are employed. I suspect that the time required to nano-facture most items (p.285) is wildly optimistic at 1 hour per cost category. However, I believe that this time cost assumes the 40-125 Liter restriction of a Desktop sized Fabber or CM. So in many cases you'll only be able to build components and there's some assembly required. The book does reference Fabbers as large as a 10 meter cube, but it's probably safe to assume that as the size factor increases so does the time. So if a desktop unit can fabricate 125 liters of smart clothing in 2 hours it would take 16000 hours (two years) for the largest industrial unit to fabricate a thousand cubic meters of Solar Sail. (neatly folded of course) Even though you could proabably build a whole car or small ship inside the largest industrial Fabber it's probably faster to nano-facture components in multiple smaller fabbers and have your slaves assemble them.

I totally agree, traditional fabrication methods are likely to be massively more efficient than nano-fabrication for items that can be manufactured in traditional ways. It takes an hour to manufacture one Trivial item with a CM or Protean Swarm. A traditional assembly line could probably make tens of thousands of the same thing in the same time. If I were going to use nano in large scale fabrication I would use it to build the assembly line.

Well I'm not sure where it's written but I remember that there is a new Nano way of building habitats that's basically seeding an asteroid with the nanos and watch it grow. They guessimate about 3 years but it requires minimum supervision. So if 3 years is a good span for a standalone state of the art habitat I'd say that a good sized scum barge-like ship (housing about 20K) might be around the same time. If you divide the project time by the order of magnitude a 10K habitat 1.5 years. Then again compared to today when a supercarrier type aircraft carrier takes about a 3 to 4 years to build and a luxury liner takes 5+ year from drawing board to shakedown cruise it hasn't changed that much.

One of the big things that can speed the construction of major projects will be the number of intelligent and semi-intelligent minds that can be thrown at the operation. With nanotech, industrial robots and enough competent intelligences around, you can break down an immense plan into thousands or even millions of discrete parts. This gives a huge edge to any society capable of organizing at this level, by whatever means. The Jovian Junta, the Morningstar Constellation, the Titanian Commonwealth and even the Planetary Consortium. If you have citizens, or workers, enough to throw at something, and resources that are easy enough to come by, you can get it built. Remember that there are millions of infugees available to the hypercorps if they decide to move individually or collectively on something, and most probably have few qualms about forking or even alpha forking their indentured servants "as necessary." It's a fair bet that all of the major polities have incredibly impressive defenses against overt attack, given the nature of the environment they live in, the impact of high-yield fusion and moderately advanced nanotech, and the ongoing TITAN menace.

Considering that in 10 years the EP society has been able to build up infrastructure for 500 million citizens, their building time must be very impressive. While I wouldn't say that a destroyer is build in a week or a month, half a year or a year would be reasonable. Consider also that besides nanotechnology(which in itselfs as it works in EP isn't a big game changer regarding time of construction but rather complexity), the EP society has machines and synths and infomorphs that can work 24hours/day(simply speaking) with extreme efficiency. All complex and detailed work within structures can be performed by nano, while the available oversight of construction is also of magnitude larger than available to us. Large scale projects require of course degree of organisation and resources not available on regular basis to individual groups but those that are in hands of organised state-like entities. Thus Titanian Commonwealth has been mentioned as engaging in several infrastructure of large scale(Scoop for getting antimatter which only partially performed up to expectations), Planetary Consortium(terraforming of Mars)/ LLA(making interior of Moon habitable, acquiring water resources), Morningstar Coalition(terreforming of Venus atmosphere). Jovian Republic probably engages in large scale military projects and energy infrastructure due to its location.

"Sealing of the pieces." Quincy, if you and I are thinking of the same thing, that's a really good point. even outside the Fabber, Nano-tech will probably still be applied in assembly to make the job quicker and stronger. I spend a good portion of my time at work welding. (SMAW 3g/4g) On a good day, if i'm cleaning my own slag I can weld 2-3"/ minute. With metamaterials and other exotics my job goes away. I envision most field fabrication of large parts to involve something like 'Glueing' things together. Except the 'glue' is nanites that create an actual molecular bond between parts. And because fabrication in the shop can be controled down to the molecule the fitment of the parts should be perfect. That's going to be Damn Cool! Which brings up another point. In structural steel getting the fitment right happens a depressingly low percentage of the time. It's rarely a problem on the fabrication side. It's almost always a problem on the design side. From my perspective, in the field, 80% of engineers are fucking useless CADjockeys with little to no innate mechanical ability or ability to internaly visualize spacial relationships. 15% are older and capable of doing arithmetic and making judgements independent of what their software tells them but still have little mechanical ability. 5% are really pretty bright, are capable of predicting issues and looking for alternate solutions and CARE, or they've had significant field experience. My point: Throwing more minds at a project quickly reaches a point of diminshing returns. Double that if you're useing indentured minds. You can only break an engineering project down into so many parts before the Job of managing becomes more complex than the Job of engineering. At that point your management software is complex enough to do the engineering too, I suspect.

There are some substantial differences between Eclipse Phase and the real world, besides the obvious advances in nanotech. One, people are, on average, much smarter, healthier, agile, etc, and better educated as well. Two, with the vast number of infugees in storage, many polities have their pick of extremely talented people to restore. Three, even less enthusiastic indentured workers can be made so using basic psychosurgery or other forms of influence, or made more talented using skillsofts. Four, infomorphs operate at a few times normal human speeds under even normal conditions, and can go into simulspaces where the time differential can be sped up to a multiple of 60 -- thus enabling an infomorph scientist or engineer to pull off enormous amounts of work in hardly any time whatsoever. Five, plenty of AIs and AGIs exist that can handle basic details, not to mention synths and industrial robots. And so forth. How many pieces you choose to believe a major project has been broken down into is your option... there's an argument to be made that with the inclusion of nanites, there are easily billions if not trillions of discrete elements involved. I still suspect, given the technology, resources and limitations described in Eclipse Phase, that many of the stronger powers are capable of immense construction projects, and are likely engaged in them all the time. Remember, though, that many of their key works may be focused on basic aspects of their infrastructure -- energy, manufacturing, resource acquisition, habitations, communications and computronium -- as much as on defense alone.

OneTrikPony wrote:

I like to believe that the people on the design team (Architards and CADjockys in the vernacular ;) ), are on average smarter than me. But the complexity of thier job exceeds mine to the point that they're wong as often as I am.

I think this is a very important observation. In EP many projects are not so much materials limited as complexity limited: getting the myriad parts to work together without serious bugs is very hard, and testing all interactions suffers from combinatorial explosions. And thanks to apparent ease of design, ease of assembly and high ambition level devices are at the edge of feasibility. It is a lot like software. Many big software projects show enormous cost-overruns (and embarrassing total failures are not uncommon), often because they were slightly too complex. Many of the most successful projects build on smaller pieces instead. Just consider what medichines are supposed to do - diagnose, repair and protect an entire transhuman body. That [cost: low] gear is the pinnacle of a century of medical engineering, millions of man-years of work and likely many billion man-years of testing. A cut-down simpler version that merely fight bacterial infections or diagnoses problems would likely be much easier to design, build and use. Most likely current medichines emerged from a myriad of smaller specialized tools and projects, which were later integrated into usable packages. A lot like web browsers - and there are still plenty of exploits, misfeatures and plain bugs around.

I agree with Dry Observer about how things ought to be done - modularity, using hightech to make very simple and robust parts, built in upgradeability, reusable well-tested parts, extensive testing etc. This is more or less what I was told in our first software engineering class in the early 90's: these considerations are the way to make good software (and object oriented programming will take us there!). In fact, my teachers were probably told roughly the same thing in the early 80's or even 70's - these are the hard-won insights of software. Even my old teacher who had once had his foot crushed by a falling byte from a 50's computer attempted modularity and re-use in his punched tape machine code ("just tape the subroutines together"). But as we see every day, just because we know how we ought to program doesn't mean our software (or hardware) behaves like that. It is hard to really make good software. Economic and practical pressures favour fast development, having many and cool features, leaving much of the testing to users through betas and bug reports and solving problems well enough, not perfectly. The result is a lot of exploitable weaknesses, bugs and plain useless software that nobody really understands. Things get even worse when systems are distributed and run in parallel - human minds do not seem to be good at writing multitasking code. In EP the same thing is true for everything. Matter can be engineered just like software, and unfortunately just as badly. There is at least a century more experience about what not to do, but also a century more of imperfect solutions. In the 60 years since my teacher used a screwdriver to flip a recalcitrant bit we have moved up about 5 or 6 levels of abstraction in programming (hardware to machine code to assembly to high-level languages (in themselves evolving from direct stuff like Fortran to abstracted languages like Python and Haskell) to very high-level languages and developer tools that solves constraints and tasks). In a century I expect at least an equal number of levels, many of which are AI. EP engineering is to a large degree telling systems what they ought to be, and the tough part is having a dialogue that manages to find and straighten out problems and uncertainties before they happen. But there is so much room for software/hardware bloat (why do the mandatory recycling-tag nanites include translation libraries? does my cogwheel really need a firewall, and how do I get the software to leave it out?), unexpected interactions (medichines meeting enhancement nanites, defense sensors seeing a very remote and very weird Factor spaceship) and sloppiness (sure, *you* are a great programmer, but one of the sub-sub-AIs implementing the garbage collection part of your design is a rip-off - the error will not show up until the device has been used continuously for one year, then it will suddenly crash mysteriously). Making a really simple and robust rail-gun is possible, in the same way as making a simple and robust word processor is possible today. But most word processors in use are bloated, buggy systems that do a hundred non-word processing things too. I expect the same thing to be true about the rail-gun to some extent.

Arenamontanus wrote:

I agree with Dry Observer about how things ought to be done - modularity, using hightech to make very simple and robust parts, built in upgradeability, reusable well-tested parts, extensive testing etc. This is more or less what I was told in our first software engineering class in the early 90's... (Lots of good commentary snipped)

All good points, Arenamontanus. I suppose the assumption I'm making is that Post-Fall, transhumans are really scared of certain kinds of complexity. Which means on the one hand that there's good reason to assume that projects could reach a complexity level where they couldn't be handled by throwing more processing power at it -- not because it's impossible or they don't know how, but because Post-Fall civilizations are truly terrified of slapping that much computronium together in one place and running an incredibly powerful AGI or AGIs on it. Especially a seed AI managing key defense systems. The flip side of that, though, is that military forces also realize the danger of complexity on the smaller scale, not that they're entirely able to get away from it. So it makes sense to have at least some weapons, sensors, communications and power sources that are really difficult to hack, as a matter of survival. As far as making things is concerned, I believe your points are especially well taken. There's a multitude of things in Eclipse Phase that are constantly changing, and usually growing in complexity as they do so. But there are plenty of simple items out there using very basic tech, well known to many users. I think that open-source people (ironically) would be providing military forces with sufficiently rudimentary designs, if no one else were. Basically, some Ruster in a Martian backwater needs a railgun because she lives near the Zone, and it worries her. So she goes to her rep-rap, fab lab, open-source CNC, or whatever, and slaps together a very simple device... just enough to be rugged and get the job done. And having done that, she circulates the blueprints among her close Ruster friends and a few other contacts, thus gaining a touch of @rep. Eventually, the military gets this build or one like it. Meanwhile, some Brinkers are really concerned about lurking Exhuman raiders wandering in, stripping the local hab of everything of value and leaving the asteroid a smoking ruin. Not so much about their lives, because they all have emergency farcasters, but *darn* if rebuilding isn't expensive. Not too convinced of their own ability to counter the brilliant but aggressive Exhumans in a fair fight, they stamp out a huge array of parallel sensors (too many and too simple to hack) and a couple dozen separate parallel-processing data crunchers to assess info. And finally, once an incoming intruder meets certain criteria and fails to hail the hab (and subsequently to change routes of approach, exchange messages, agree to certain ground rules), all heavy weapons are given authority to automatically lock on and fire within a certain distance. If any major military *didn't* have the basics of how to run a hard-to-hack, back-to-basics defense system, the general concepts would be floating around everywhere, and certain minor but technically savvy habs' original (and already outdated) plans would be accessible for a reasonable price. Having said that, I suspect even a lot of open-source programming of a basic nature is a hideous tangle and getting worse, as you imply. But where simplicity is a matter of life and death, and where people who have to build from the ground up anyway have addressed these problems, I suspect a fair number of bare-bones approaches proliferate.

I discussed my airplane/shuttle estimates earlier in the thread with an engineer friend, and he pointed out that besides the complexity issues of modern airplanes there is also a much more cumbersome organisation involved in building them. But most of this organisation is relatively inefficient, dominated by people doing a lot of planning and management (and their hangers-on) rather than actual engineering. Compare Skunkworks with Boeing. As the Fall occurred there was not time to do things properly or by the book, and the real doers went off and did what needed to be done. This typically happens when there is a crisis, and in the case of the Fall the rare places where the managers blocked the necessary fixing simply died or were reduced to abject poverty (parts of the LLA come to mind). What also typically happens after such a crisis is consolidation: by now the managers are taking back power (after all, they are on average better at powergaming than the doers). So while people were hacking together new farcasting protocols, hammering out the Treaty of Uniform Enforcement and manufacturing new habitat domes by the week in the first year after the Fall, by now these processes have slowed. Issues of intellectual property, political viability, safety, quality assessment and ideological purity are allowed to influence the engineering process again.

Rhyx wrote:

Quote:

Arenamontanus: As the Fall occurred there was not time to do things properly or by the book, and the real doers went off and did what needed to be done.

So it seems to me like the "real doers" in crisis situations seem to have a tendency to be totalitarian. This is probably why overinflated management sneaks in after. You can only do the "long hours, no pay" for so long thing before the workers get restless. After that strikes begin and then negotiations begin resulting in security measures, quality control, work norms, unions and other things of the sort that "slow down" actual construction but raise overall standard of living.

I don't think the doers are totalitarian. Rather, the mindset of everyone in a crisis is totally problem oriented: what do we need to do to survive? People who have ideas and do something about it tend to get the others with them. Sure, it is hard work, but it beats the alternative. People are extremely highly motivated when it is about survival. When the pressure lessens, that is where niceties such as pay, social norms and freedom began to reappear as motivations. Looking at how societies have handled real crises most don't go from survival mode to striking for better conditions mode (that seems to happen much later, when they feel entitled to it), but rather there is a gradual more or less subtle ousting of the doers from control. Think of the fate of Winston Churchill after WWII.