Neil Gershenfeld: The beckoning promise of personal fabrication

The Talk (from a slashdot link): [www.ted.com]

My thoughts:

NG seems to believe that fabrication is just a way station to a world of "smart matter" in which our surroundings would be composed of materials combining embedded computational ability with self manipulable structures; imagine a carpet that senses the number of people in the room and grows sufficient chairs, with placement based on time of day and topic of conversation. While this is a fascinating or terrifying idea it does not really bear on what we are doing. This is nanotech stuff and I don't think we can reasonably expect to do a lot of work in that area anytime soon. I also think his 20 year plan is way optimistic.

Where do folks see RepRap in 20 years? 10? 5?

I like his comparison of the FabLab project to the PDP11. In that metaphor we are working on the Altair. Hmmm...

I've read a little about the fablabs and I am ambivalent. $20,000 is a lot of money in Ghana and while I think it's great to give people the ability to shape their own destiny, I wonder if lower tech tools would not have done the same thing in a more robust way. What will happen to these labs in a few years when they need things like replacement lasers or micron tolerance bearings? Will the labs be able to generate the funds to support themselves and will the need to generate such funds alter the community around the tools? Is a bicycle powered jigsaw a good substitute for a laser cutter? Which is better for a town, a fablab or a water treatment plant?
I think that one of the great possibilities for RepRap, if it can move beyond working just with thermoplastics is it's ability to be a tool that makes tools. Thus instead of a $20,000 fablab each village could receive a $1,000 RepRap based kit capable of making the tools for other shops (wood, metal, mechanical, electronic) with limited additional imports (carbide cutters, electric motors). These tools could be manual, powered, automated, or tied into the reprap (fabbervated?) as the owner decides is best and as budget allows.

NG's two killer apps for fab are 1- in the first world, hyper-personalization- products for a market of one; and 2- in the LDC's, the fab divide, giving people the power to create the answers to the problems they face locally. I don't think these are actually different things, they are simply the application of the same method to problems on different tiers of Maslow's hierarchy. The method is the harnessing of information technology to increase peoples ability to create objects by automating repetitive activity and harnessing expert systems as intermediaries between tools and ideas. A fabber allows you to learn one skill (3D modeling) which it then mediates to various tools which you could learn to run yourself after a several year apprenticeship, or you could do the same thing by hand if you were willing to spend long hours doing the same task over and over.

I also think NG fails to appreciate the ecological/economic advantages of making products at the point of use, an advantage that is multiplied if the materials can be produced locally.

The place where I agree with NG unreservedly is his idea that the most important changes these tools will bring are the changes in the communities that surround and use them. Fabbing increases the value of creativity and promotes individuality. I think that it has the ability to fundamentally change the way that society produces value and thus could up end the existing structure of economics, politics, indeed all of society. This does seem to be an anarchistic technology, more so in an Open Source model. Will totalitarian societies try to restrict fabbing? Will future revolutionaries smuggle RepRap parts into Hong Kong?

seeing as how i'm both a reprapper as well as involved in the fablab stuff, i'd have to say that we are in a unique position to take part in exactly what he's preaching.

to put it metaphorically, RepRap could be the engine in the FabLab racecar.

"RepRap could be the engine in the FabLab racecar."

ROTFLMAO! Yeah, and guess who'd get all the credit for the success and all the invitations to do the nice, remunerative keynote speeches. It would be yet another fine example of what MIT does best ... take credit. ><

Edited 2 time(s). Last edit at 02/29/2008 04:02PM by Forrest Higgs.

A wise old hippy once told me that "you can accomplish anything if you don't care who gets the credit."

Yeah, but the rewards, including the backing to finance all those other good ideas you have, are reaped by those who took the credit. Ivy League faculty are the world's best at making it appear that anything worthwhile was their idea.

Edited 1 time(s). Last edit at 02/29/2008 09:45PM by Forrest Higgs.

Maybe we should respond by getting to work designing and building fabrication machines we can build using a RepRap. Say, a laser cutter and a large CNC router.

Would people be interested in this?

In order to agree with Sebastian I should probably state my heresy at the outset.

I don't expect to see a self reproducing machine in the next 20 years, even allowing people to do the assembly. I DO think we can achieve a self reproducing WORKSHOP in 5-10 years given a few bright but not technically skilled folks to do assembly. By workshop I mean a suite of machines, most of which are fabervated, (which is my word for a machine which is automatically controlled based on a computer interpretation of a user designed final product rather than automatically controlled based on user commands) which could be built from a "seed" of two or four original machines (or their components) and a small collection of specialty parts like carbide bits and microcontrollers. Such a workshop should be able to reproduce itself if supplied with materials which are mostly "primitive" or "raw" (that is chunks of scrap aluminum or barrels of pla granules rather than polished guide rods and plastic bushings) and competent workers to do setup and assembly. Such a workshop would be able to also do effective work on a variety of scales and across a variety of materials to address needs outside of it's own creation, again in a fashion that allows it to be used by persons with little knowledge of the mechanical processes involved. Given that....

I am a fan (in an abstract way) of the Gingery books, the "Build a machine shop from scrap" series. They start with building a small aluminum foundry and go through building a metal lathe, a shaper (a type of machine tool for using reciprocating rather than rotary motion), a drill press and finally a milling machine. The lathe at least, has a good following on the net and a reasonable success rate. Most of the tooling for any machine is done on the previous machines, things like carbide bits, motors, and shim stock is bought but 90%+ of the machine is created on site. The machines are manual but upgrading to CNC/fabber should not be a major deal. With a good lathe you can start making guide rods and drive screws, from there the sky is the limit.

I am also a fan of Andrew Wyger's books, which are devoted to making and repairing power and hand tools as well as blacksmithing. He is a disciple of handmade but he has plenty of hints on adapting and repurposing.

My goals for the machine I am working on include a milling head to make forms for lost foam casting. Lost foam is an investment casting method where a positive of the item is made out of styrene foam (or perhaps CAPA?) with gates and risers attached; this is buried in packed casting sand and then molten metal is poured into it vaporizing the foam and replacing it with metal. The process is stinky but basically impossible for a normally intelligent and attentive person to mess up. (It is widely used by college students, not a group known for their caution or forethought) Aluminum foundry work can be amazingly low tech ( charcoal bricquets and tin cans) and is reasonably safe. Casting is an amazingly powerful technology which, I think, would work perfectly to extend RepRap's capabilities. It is hard to imagine many useful tools made from HDPP.

So if we assume our 4th generation reprap can deposit a variety of materials and do cutting through hardwoods and milling thru light surface work on aluminum, and we have a furnace which can melt and pour brass with minimal attention, and a lathe with screw cutting abilities, I think we still need a grinder/polisher and a wire pulling/insulating machine. What else is needed to have a self reproducing workshop?

Edited 1 time(s). Last edit at 03/01/2008 04:38AM by BDolge.

"Maybe we should respond by getting to work designing and building fabrication machines we can build using a RepRap. Say, a laser cutter and a large CNC router.

Would people be interested in this?"

Yup, already got plans for shoebox sized lathes, drill presses and a gear cutter for small gears that I plan to adapt into reprappable objects. I'm also thinking about how to put together a pick-and-place machine. My current project in to make a reprappable 3D scanner. Sharp has an IR distance measuring chip for US$12.50 that should make that totally possible.

There's nothing for us in a $10,000 desktop machine shop that you can only build in a factory. I follow Terry Pratchett's dicta...

"...the only tools a dwarf needed were his axe and some means of making fire. That

"I DO think we can achieve a self reproducing WORKSHOP in 5-10 years given a few bright but not technically skilled folks to do assembly."

The Gingery Family have been publishing a series of books on how to do this for the past 20-30 years.

[www.lindsaybks.com]

A lot of people have done it, too.

However, you can't take the Gingery route to a DIY industrial revolution if you are one of the 50-60% of today's bright kids who are growing up in a single mom family. With luck you have a room of your own and maybe room under your bed to store your stuff. For sure you don't have a yard either big enough or in an area where you aren't going to attract official notice when you try to run one of the easy-to-make Gingery charcoal foundries. For sure you haven't got room to build your own garage workshop to house their more or less standard sized lathes and milling machines.

That's where I think Reprap has something real to offer. Darwin 1.0 will fit inside a 15" cube. Tommelise 2.0 is being designed to be able to be folded down and slid under a child's bed. FDM won't set your house on fire and doesn't need to happen outside. That's the one part of the Gershenfeld paradigm that IS right. The workshop needs to be desk sized. We're long past the era when kids and most adults had enough living space to be able to throw their elbows out without bruising them.

I checked the link. Is there a particular order to get the books? I would guess the foundry is first, but is there a necessary second? What books have other books as prerequisites, and what are those prerequisites? Dropping $60 on all the books isn't something I can justify, especially as the difference between the package deal and the individual books is somewhat less than half the price of the books. (If I buy them individually, and only get through half the series, I'm maybe out the price of the last purchased book. If I buy them in bulk, and only get through half the series, I'm out the price of half the books, less two thirds of one.)

Space, I have.

You definitely need the lathe to make parts for the other machines.

The order goes:

Foundry
Lathe
Shaper
Mill
Drill Press
Accessories
Brake

Thank you.

You can get an idea of what is involved before you buy. This guy has built one of the foundries and worked with metal casting quite a bit. He has quite a nice little website.

[www.backyardmetalcasting.com]

Here is his lathe project.

[www.backyardmetalcasting.com]

Here's another with what I think are better pictures.

[tjanstrom.com]

I think that you will definitely get the idea that doing the Gingery route is more of a vocation than a part-time hobby. The impression that I get is that it is bloody hard work, but satisfying.

I have an artist friend who does his own foundry work. I showed him the books and his comment was that they underplayed the really bad things that can happen when you cast metals. It's apparently damned dangerous work.

Sean-

I agree that most folks don't have alot of room these days, but an HDPP hammer isn't good for much but threatening your siblings. Materials matter. A machine that can make a copy of itself is only a curiosity unless it can make other things people want. Look around you and I would bet you find very little in the way of STRUCTURAL (as opposed to cosmetic) thermoplastic. The last thermoplastic furniture I had was a set of milk crate bookcases. If RepRap is to become a real game changer rather than a geek toy (and I love geek toys) it has to be able to make things that are bigger than a breadbox and stronger than a spatula. Yes, there is a "market" for a desktop fabber, but you can't use one to make a desk. I don't think every individual will own a Self Reproducing Workshop (TM)- (although I am planning on building a garage) but every community center might; as I said I think the real changes that will come from the fab rev are social, as local communities become able to provide products customized to their needs and mores I think they will tend to contract and differentiate; I hope that an increase in wealth and communication will serve to prevent this from collapsing into xnophobia, regardless until we reach a level of hard nanotech as portrayed in say "Diamond Age" technology will remain a community endeavor.

A metaphor-- The RepRap as washing machine- lots of people have one, some bigger than others, some older or more capable, some in the basement, some in the closet. If you can't afford one or don't have room you can go a couple blocks to the fabb-o-mat and rent the use of one for a few hours whenever you need it. Or maybe your church or community center has one that it bought with grant money so local kids can always have access. Of course if you don't want to spend your time on that sort of thing or need something special done (dry cleaning or new shock absorbers) there are folks who will do that for you for a consideration. This sort of mixed economy will evolve over time as the machines become smaller and more capable until eventually we achieve a sort of nanogoo which listens to you talk and then oozes out the answers to all your desires- by that time I hope to be off world or somewhere where things are still handmade for people who like and can afford such anachronisms. Because I like making things. So there it is.

Edited 1 time(s). Last edit at 03/02/2008 12:11AM by BDolge.

Forrest,
Yes, the Gingery lathe is a lot of work, that's why we want to automate the process ;-).

Yes foundry work can be dangerous, so can traveling at 80 miles per hour. We have safeguards and procedures that protect us. I've done castings with bronze, zinc, aluminum, and iron. The basic rule is stay out of the way of the hot stuff. That and be able to put out any fire you start pretty much sum it up. You plan your work and work your plan. Pay attention.

Anyway, I'm not advocating everyone join me around my charcoal furnace, but if someone wanted to help me design a controlled furnace with inlets for fuel and metal, an exhaust vent and outlets for molten metal and slag, with microcontrollers to handle the safety interlocks, then we might have a worthwhile project. After all if I told you I planned to get from home to work by detonating a gallon of gasoline....

"I agree that most folks don't have alot of room these days, but an HDPP hammer isn't good for much but threatening your siblings."

You need to read a bit more deeply into exactly what can be done with FDM technology now that we are reaching the end of the Stratasys patent. For example,

"Most of the existing research efforts in FDM have been primarily directed toward the development of new materials or techniques for material deposition. Agarwala et al. (1996) demonstrated the possibility of applying fused deposition ceramics (FDC) and fused deposition metal (FDMet) to a variety of ceramic and metal particle systems. Gray, Baird, and Bohn (1998) investigated the use of short thermotropic liquid crystalline polymer fiber reinforcements using existing FDM equipment. Yardimci et al. (1996) developed three heat transfer models to describe the temperature profile of the deposited roads for FDM and FDC. Research efforts were also made to characterize FDM part quality. The mechanical properties of filament made of the polyamide material used in the FDM process were provided by Comb, Priedeman, and Turley (1994). Van Weeren et al. (1995) described the defects present in parts manufactured by fused deposition. Dutta and his associates (Kulkarni and Dutta 1999) investigated the relationship between deposition strategies and the resulting stiffness of FDM parts thus produced. Matas (1999) characterized the mesostructural features of unidirectional extruded material as a function of FDM process variables."

[findarticles.com]

One of the most interesting possibilities that has been explored in the past ten years is the use of FDM to create precise objects made of silicon nitride either as a paste or entrained in polymer filament. These green objects are then fired in a high temperature microwave furnace. (the lab version of this furnace is a modified kitchen microwave oven)

Also...

[www.springerlink.com]

FDM also gets used for making molds for high strength ceramic objects made of such ceramic materials as silicon nitride. Here is a paper showing how microturbine impellers can be made.

[www-rpl.stanford.edu]

If you want to wait around for intelligent goo to grant your wishes, be my guest. In the meantime, however, there are some damned interesting things going on, in my opinion anyway.

Guys - I have a Gingery Foundry and am working on a Gingery Lathe at this time. Here is what I have done so far:

[www1.teeter.com]

As you can see it can be done very easily. Just some work in the shop. And yes working with hot metal has some danger in it. Working with 1100 degree F metal has some danger in it but if you follow some simple rules you will not be injured. But just think about this I can wake up in the morning with an idea of something that I want to make. And can have that item in the afternoon. No calling vendor to find out that what I want is not a standard item and will cost 10 times more than I want to spend and will take 6 weeks to make.

The secondary reason for making the tools that you use is so that you can make any tool that you need to create something special. (ie custom horizontal milling machine to make a foozbart-clamp-support). Or maybe you want to make a turbine and generator to supply emergency power for your home. Also the electronics and stepper motors can be applied to the machines that you just built and you have a cnc lathe/ cnc milling machine.

So until I can walk up to the machine and say "Tea, Earl Gray, Hot" and get a hot cup of Tea I will not be satisfied that we have finished what we started.

---

Bob Teeter
"What Box?"

Forrest-
I like the FDM ceramics, do you know where I can get some more in depth info? Still a ceramic hammer is lighter than a metal one (thus less efficient ) and prone to shatter rather than bend. Every material has it's best uses and there is no single way or substance that will replace everything else, especially when you start paying attention to people's aesthetic wants as well as their technical ones. Wealth is not having a lot of stuff you don't want, it's having enough of what you really want.

Also, I'm not sure you're aware of it , but that microwave kiln can also be used to melt metal for casting and SiN4 can be sintered over charcoal. I would certainly advocate using ceramics where they are appropriate, but they are not inherently safer or "better" than metal.

"Every material has it's best uses and there is no single way or substance that will replace everything else, especially when you start paying attention to people's aesthetic wants as well as their technical ones."

LOL! I'm sure that that argument was used when weaker but much, much cheaper wrought iron began to compete with bronze in the manufacture of weapons in the 12th century BCE. Farmers couldn't really afford bronze tools. They could, however, afford iron.

Silicon nitride is the same sort of revolutionary material. When you think, for a moment, about what it is made of, viz, silicon (sand) and nitrogen (air) and compare that with what it is competing with you can see why it is going to be a very important structural and tool-making material.

Making plastic hammers is hard, but making plastic nails is an already existing technology. You fire them out of an air-driven nailgun, which could presumably be made of plastic. They have half the shear strength of stainless steel nails, but are also half the price.

Well I happen to think that bronze is still prettier than iron and jewelry sales say most people agree with me.

Iron tools probably won out because they were more durable than bronze, not cheaper. Iron shows up in the archaeological record in the same way you would expect any expensive item to, first in jewelry and decoration then in weapons, then in tools and everyday items. Iron was initially expensive because, while the ores are more common than those of tin and copper they are harder to extract. Also iron required special techniques and high temperatures to work, whereas bronze could be efficiently recycled with tools available on the farm. Bronze remained in use for quite some time after the introduction of iron, the usage simply changed.

Speaking of which, my reading tells me that sinterable SiN4 has been around for 30 years now, but is still considered a niche material, when does the revolution start? I'm not really knocking ceramics, I think they are a great technology and an advancing one but like any technology they need to be understood for what they are.

Ah well, there you are then.

This is an interesting debate.

On Nanotec, for me this is still appears to be at the "Everything is possible" and the "Panacea" stage of a new technology.

There is an awful lot of hype, extrapolation and religion involved with it. As yet there is insufficient hard knowledge to actually displace the religion with hard engineering and scientific fact.

Yes it has huge potential, quit what that potential is depends on the Physical Constraints and Limits that still have to be determined, challenged and eventualy acepted as fact.

Interestingly enough the world is replete with Nano Technology that we dimsiss as biology.

This is self fabricating, repairing (to a degree) and reproducing.

For example a solar powered manufactory that constructs building materials for us from little more than trace elements found readily in the environment, Carbon and Water is called a Tree and dismissed as nano technology because we did'nt make it in a lab.

Enzymes are molecular cutters and sometimes joiners (Yes they are not usually consumed by their action upon the mollecules that their nanostructure (shape) determines thy will act upon) that are manufactured by the million in automated manufactories called cells, to pre specified designs stored on short lengths of twisted tape called DNA.

I actively use nanotechnology and yeast enzyme manufactories to turn fruit and vegetable sugars into wine.

We are all nano machinery, cellular automata and highly complex.

But there are a bunch of things that this nanotech which is arguably incredibly diverse and complex cannot do.

There will be very good reasons for this that we have yet to get a handle on, let alone understand the nanotec soup in which we live and are components of.

On FDM using polymers, there is no reason why given mastery of the subject with the materials at hand FDM cannot be utilised using more durable materials ie metals and particularly alloys. the same techniques and knowledge set will underpin it but with differing emphasis depending on the final material.

Oh and lastly if you want to do some form of investment or lost something casting using FDM technology with wax is going to be cheaper and less chemicaly messy than foam.

Either way FDM is a very valuable tool that we need to do more research on.


Thoughts for what they are worth.

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I agree with much of the above

Especially the bits about replicating workshops vs replicating machines (my take on this is that it would be better to have several repraps with different heads, rather than a complex, multiple head or interchangeable head do-it-all device). Constructing a lathe tough enough for metalwork would be the trickiest yet most useful bit at this point, though easier than making a mill. Would I be right in thinking that the only bits that can't be fabricated using FDM on darwin are latheable?

The notion of casting is certainly an interesting one, though I don't live anywhere I could possibly set up my own forge/foundry/charcoal oven etc etc. Creating wax positives seems to be a much better idea than making thermoplastic ones to me (ease of finding the damn stuff for a start. And you could presumably use beeswax or tallow). Has anyone tried such a thing? I guess you'd need a pellet-fed extruder, as I don't see wax filament being even slightly robust.

Andy, check out:

[parts.mit.edu]
All the parts you need to do up an algae that makes thermoplastic, maybe.

Regarding building tools, I'm more interested in using a RepRap to make the machines which are difficult or expensive to source commercially. i.e. CNC routers, laser cutters, and so on. Drill presses, not so much.

I seem to remember someone (I think it was Dead Genome) suggesting that you could actually melt glycerin soap just like wax.

Great for the left overs from biodiesel.

I keep some bees and have some beeswax to hand, Beeswax is very malleable and probably a little too soft at ambient temperatures for the kind of dimensional stability you want for investment casting etc.

However there have been some very good books published by ITDG (Intermediate Technology Development Group) one of which I recall reading was all about doing investment casting in the african bush using whatever was at hand.

They made up wax with good enough properties by blending Beeswax with other things. They also made up a refractory slurry from what was available too.

Well worth the read if this is your area of interest.

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

don't think it was me... but I'll happily take the credit if no-one else wants it

---

my 2-cents worth..

replacing existing products exactly is not what this technology is about IMHO. It is more about finding analogues that can be built with this technology that achieve the same (or better) end result, hence my comment earlier about plastic nails instead of plastic hammers.

On the subject of making multiple machines, this is ok for the hobbyist with lots of space but will never compete with mass manufacture. We have always had the technology to make objects on a multiple machine/process basis if we have the time and space, the revolution that was generated by people like Adam Smith and later Henry Ford was in the realisation of splitting up the manufacturing process into seperate components, each with it's own specialist (specialised people in the case of Smith and specialised machines in a line in the case of Ford) and thus being able to manufacture a lot of identical products more efficiently but at the cost of personalisation - You can have any colour as long as it's black.

On the other hand, the true revolution inherent in rapid prototyping comes from the fact that if each machine can be a generalist then you actually get greater economic efficiency than you do in a production line, as long as each item produced has to be reasonably different.

Think about a warehouse fitted with a production line that is built with the latest digitally controlled tools so that you can vary each product coming off the line. Each item on the line will, due to the differences, take up more or less time on each tool so you will have to have holding areas between tools to stack the quicker items, or else have downtime on your machinery. These holding areas will take up space. Whatever you do, you end up with the manufacturing equivalent of either traffic jams or car parks. These cost either time or space (and therefore money - the majority of the cost of producing a new car is the time it takes the paint to dry). This is why it is uneconomic to manufacture very customised products on a mass production basis for the same price as identical ones.

Now picture the same warehouse space filled with repraps capable of colour printing in multiple materials. Each reprap can in this model carry out all the processes required to produce a given product. Once a product is complete, it is removed for immediate delivery/pickup and that reprap then starts printing something else. No traffic jams or car parks. This enables uniquely customised products to be produced for roughly the same cost as is currently achieved by mass manufacture. Also, repraps have the added benefit of being able to produce more repraps should your demand increase.

The most difficult trick by far is getting repraps to the stage that inkjet printers are at currently, namely making them cheap, simple and robust enough so that the everyday consumer has one on their desk at home, cutting out product delivery. This is the point at which the majority of people will stop being the purely passive consumers that the economics of the production line has generated and start thinking more creatively/personally about the objects which surround them yet again.

---

I do love an early morning rant, tis good for the soul..

I agree with you, with one reservation.
Manufactured goods are faster to assemble.
Consider the length of time it takes a printer, that is a printing press, to produce one book. Compare that to the length of time it takes a desktop printer, even a fast laser printer, to produce all the pages of one book.

In your factory, having a number of repraps is still potentially slower than having a line of discrete machines, each having downtime while waiting on another machine in the line to complete some customized piece.

However, there is a point where the uniqueness of the product justify printing on the slower, but much more flexible, machine. I'd never consider sending this thread to a printer, but I might run off a quick copy on my inkjet.

Your basic premise stands, but you didn't mention the efficiencies inherent in specialized machines.

Your printed book example is a good one.

Of course, specialised printing machines are much faster and more efficient, in the first instance, than a laser printer. They're also much more expensive, though they can make the book more cheaply. The problem is that by the very virtue of being faster they're perforce distant from the customers of their product.

The key is the length of the production run you have to do before you achieve those savings. If you can produce 10,000 of a specific book you can achieve considerable savings. The problem is that you then have 10,000 specific books all in one place when your customers may be all over the world.

I can give you a very specific case that happened to me a few days ago. I was faced with a short time line on an assignment and I needed to generate some intraclass correlations for a set of data that I had. My existing stat pack made vague noises about how to do intraclass correlations, but gave no specific instructions. Another stat pack had instructions, but I hadn't used it in 25 years and then only on a mainframe.

I discovered that the method and how to use it was published in a book, "Statistical Methods for Research Workers" that had been in print from 1925 till 1990. I needed a chapter of it, the bit on intraclass correlations. It was in none of the libraries and buying a copy off Amazon was going to cost me roughly $160. I had the money in my project account, but it was still going to cost me 3-5 days for the order to be processed and the book to be delivered.

I was able to find a website in Canada whose creator had lovingly put the original 1925 edition, which is in the public domain, on-line. Mathematical conventions have changed a bit in the past 80-odd years but I was able to cross-check the author's original formulae with those in the Wikipedia entry for the method. The professor who'd web-published the book had scanned the original forumlae, so I knew that there wouldn't be typos in them save for the author's I never trust Wikipedia entries as a sole source in consulting work. It's just not a good idea.

Anyway, I pulled down the relevant chapter as an electronic document and, because I needed a hard copy I printed it out for my contract presentation.

The chapter ran to about 25 pages. My printer, which cost me about $120, prints 36 pages/minute at about $0.05/page for ink and 0.025/page for paper.

Thus the chapter I needed was in my hand in less than a minute after I'd found it for a price, if I throw in wear and tear on the printer and sundries like electricity, I got it within an hour, search included, for about $0.25.

That's the difference between buying something in a box and having your Reprap machine make it for you.

The big hassle with factories that make a lot of something cheap is that you have to ship the item to the customer and the item is always less dense and less efficient to ship than its raw materials. If you just ship the customer the raw materials, he can print pretty much exactly what he needs.

For example, what if you needed a spur gear with 27 teeth to make a special power transmission. Try finding one of those in a catalog.

I was trying to include the efficiencies inherent in specialised machines. Sorry if I wasn't clear.
My point was for products that are different enough that each of the specialised machines has to be individually programmable anyway.

With a thousand identical books being printed on a press, then of course it will be quicker than inkjet or laser printing. But if you required every one of those thousand books printed to have different content, be of different length, have a different paper size, be bound differently, etc. then a fixed metal press will not do this. So you will have to use digital printers, variable paper cutters, a programmable binding mechanism, etc.

In this scenario having a line arrangement is going to be less efficient than having a stack of generic machines operating independently, assuming that you can cram all the processes within a reasonable volume and keep the time per process within the same range as the machines on the line.

Printing a thousand identical books will always be quicker than printing a thousand different ones, however printing a thousand different ones will run into the problems of traffic jams and car parks if you just use a programmable version of Henry Ford's production line unless the differences are very small.

In truth the ideal economic solution is somewhere in between. For instance, we are not currently looking for the reprap to do the chemical process of making raw materials on the fly, instead we feed in filament or chipped stock that is produced by a factory at a far greater efficiency than we can achieve. Although I do not consider it outside the bounds of possibility that someone on the fine line between genius and insanity will design an elegant and practical reprappable extruder with a waste processor built into the top of it at some point. (I think that someone already mentioned an idea like this that would be fed on old plastic bags)

All I am saying is that it makes a lot of sense to try and get the reprap to do as many processes as it can until those processes start impacting on the time and space of each other. Where they do impact on each other external specialised processes are needed. This grey area of what is practical to include within a single unit is one place that I hope there will be a large amount of continuous argument, experimentation and research.

Another point is that if we have a large array of seperate machines that are required to make a given product then even if those machines are reprappable, people will still go to the shops rather than go to the effort of printing off a load of machines before they get their product. The shop they go to may be a reprap specialist however, which would be a similar model to the business of custom printers before the DTP revolution took off (my parents ran such a business in their shed printing leaflets and newsletters for people when I was 4 years old). This seems a reasonable halfway point on the road to everyone having one of these things on their desk. Indeed the signs are that this business model is already starting and I think it is something we should encourage amoung the reprap community. However it is a business with a shelf life, if the eventual aims of this project are successful.