President's Commission on Moon, Mars and Beyond

>> No self replicating nonbiological system exists. Putting such an
>> enormous technology development task in the critical path for lunar
>> exploration all but guarantees failure. The program you propose
>> requires enormous amounts of R&D and long planning timelines. All
>> that's necessary to derail it is a change of administration or a
>> recession.
>
> Beg to differ, no hardware self replicating nonbiological system
> exists,
> simulated systems of this kind are common, we made several in the
> computer
> games company I was working for a couple of years ago.

Simulation is not reality. The fact that you can simulate something
may be helpful in designing the real thing (or it may be nearly
irrelevant), but until there are existence proofs of the desired
technology here on earth it makes no sense to plan to do it on the moon.

>> Why make things more complex than they need to be? If
>> self-replicating
>> factories are possible, build on independent of the return to the
>> moon.
>> That way if it fails the lunar plans are intact. If it succeeds,
>> there will be small scale infrastructure on the moon to help setting
>> up
>> the first factory.
>
> Failure is not an option. :)

No, it's a requirement. Really. In practice things break. Smart
designs allow for this and provide flexibility to deal with the
consequences. The ultimate in flexibility is still a live human being
on site. Technology to replace that human is many decades, if not
centuries, away.

> Fact is IMHO they won't make things more complex, but a lot easier.

No. Imaginary technology doesn't make practical accomplishments
easier. If you want to do something that's possible with existing
technology, you do it with existing technology. Throwing in
requirements for massive R&D programs makes everything more expensive,
and the enterprise as a whole more likely to fail.

> Think of whatever you want to do with a manned mission. Now how much
> easier
> things would be if you had a practically unlimited amount of building
> materials, paved surface, energy power stations and assembly lines
> already
> in place?

All of that infrastructure costs money. Make the R&D program a
prerequisite to a return to the moon and we'll never get there. Don't
put stuff in the critical path unless it absolutely has to be there.

> [...] Forgot the martian rover with his clockwork, self
> replicating systems can make bulldozers, by the hundreds.

With current or near future technology they cannot. You are proposing
something which is simply not within reach for current technology.
Return to the moon and the beginnings of lunar infrastructure are
within reach with technology from the 1960s. Putting as yet
undeveloped technology in the critical path makes no sense at all.

>> I'm not opposed to research on self replicating systems, in fact I
>> would like to see more of it. What I oppose is putting such a
>> difficult task in the critical path for lunar exploration.
>
> Obviously I have done a very poor job at explaining the basic concept
> of
> what "self replicating" means in this context... no research is needed
> in
> the technology itself, only in how to optimize it for shipping to the
> moon.

Now you're talking about something more like a kit with a modest
contribution of local materials, and constant human supervision from
earth. Even so, you need to ship thousands of tons of stuff to the
moon before sending the first human. The alternative is to ship less
than a hundred tons, including the first humans, and then ship the rest
as needed. When lunar manufacturing becomes a viable option, people
can start doing it.

> Nearly everything in the modern industrial world is already made by
> semi
> automatic machines. Look at the assembly robot arms used in the car
> factories, do you think those are made by hand in a middleage kind of
> ironsmith? :) That's what "selfreplicating is about".

Those robot factories are supported by a vast infrastructure that does
not exist on the moon. In principle you could package up the factory
(suitably modified), along with the needed elements of infrastructure
(again, suitably modified), and ship them to the moon. You'd have no
guarantee that they'd work properly in the lunar environment, and if
they didn't you'd have no way to fix them. In addition, the total mass
you'd have to lift is in the thousands of tons. The factories are cost
effective only because they are turning out tens of thousands of
identical items. Early lunar exploration and development doesn't need
tens of thousands of anything. Under ideal circumstances you'd have a
factory churning out very expensive hardware that there is no market
for.

> Forget the 100%
> closure required to imitating a living organism and other similar
> abstract
> philosophical concepts: this is a very down-to-earth (no pun intended!)
> simple concept. Instead of shipping bulldozers and containers full of
> building materials to the moon you send parts of factories and assembly
> lines needed to make bulldozers and the building materials you need,
> as well
> as more of the factories and assembly lines themselves.

Nobody is going to be using bulldozers on the moon for a very long
time. Initially it will be only one or two. There is no need for a
bulldozer factory on the moon at this point or for the foreseeable
future. By the time significant amounts of regolith need to be shifted
around we'll know a lot more about the moon and how to work there, and
we can design a bulldozer to do the job right.

......Andrew