I'd imagine that most things can be deconstructed, with the nanites taking notes in the progress, allowing them to reconstruct them item later, and that's the method probably used for a lot of materials. The problem is that the result is very messy, takes up a lot of memory/bandwidth to store or transfer, and is very difficult to modify or integrate into larger constructions. Early nano-engineers probably scanned an apple, and printed it out again. They then poked around the scan and grabbed a handful of cells, and tried replacing most of the contents of the apple with that one section repeated, with noise and perhaps rotation around the central axis to keep structure varied and natural-seeming. They then replaced the cells in that sample with cell macros, where instead of recording the positions of the many atoms/molecules that make up the cell, they just programatically generate the organelles in the correct quantities and appropriately randomized locations. While they're doing this, they're doing the same for the apple's skin. Now that the apple has been reduced human-understandable structures, it's easier to do something like replace the core with more apple flesh, and adjust the micro-nutrient levels to make the apple more healthy. Now assume these nano-engineers release this blueprint for free on the mesh. Now J. Random Nanohacker can take the blueprints for creating apple and mix them with other blueprints, and make herself an orange with an apple interior, or edible jewelry. For most common items, chances are that somebody's scanned it and posted the raw dump on the mesh. Many of them have probably been turned into more usable blueprints, but not all. A nanoprogramming check represents finding the usable components and describing to the replicator how to put them together, and if a proper blueprint for a component doesn't exist, taking one of the raw scans and either selecting a portion of it to copy/paste over and over, or actually take the opportunity to extract the underlying structure, publish it to the mesh and get some rep for it. ------------- As for why people don't take objects that have been nanofabricated with a restricted blueprint, I suspect that there are a set of well-known tricks for making a nanofabricated object difficult to reliably deconstruct - perhaps bits of the design would move about during the deconstruction, or will fall apart if deconstructed (or constructed) in the wrong order.

Chances are that nanofabrication blueprints are procedural generation instructions designed to tell a fabber the specific order and process by which it has to produce an object. Such instructions cannot necessarily be deduced through nanoscopic reverse-engineering, as all that a nanobot should be able to deduce is location, atomic property (whether it is ionized, what specific isotope it is) and what neighboring atoms it is linked to. The nanobots likely have to require a specific order of steps to properly produce an item that functions; simply knowing how it was placed together doesn't necessitate knowing how to put it together in the first place. To that end, having the raw structure data of an object would take up a disgusting amount of hard drive space... so much so that I don't even know if future computers would be able to handle it. Let's say that reverse-engineering an object made up of a single type of atom only needed to give you 3 64-bit positional addresses: one for X, one for Y, and one for Z on a three-dimensional plane (accurate reverse engineering would require far more data than that for every atom, and likely much larger positional addresses). For an object the size of a drop of water, the total size of the reverse-engineered "blueprint" would be 961,920,000,000,000,000,000,000 bytes. Now just imagine how much bigger the blueprints would be for larger objects, or objects made up of multiple types, isotopes and ionic properties of atoms where there is more data to be held for each and every atom. The memory requirements would become massive, even for future generations.

Totalgit wrote:

So what are Ep blueprints, why is the information they have in them to produce item x small enough to be shared among alot of people? If somebody designs a gun blueprint from scratch surely it will contain all the information to make it from the ground up using the nanites. why is the information smaller than what you would get from doing the reverse? Are people saying the nanites dont need to know what atoms go exactly where and theres a sort of fuzzy logic when making the item so theres less data on current ep blueprints? Considering making blueprints from scratch can take weeks of real time i dont see them not needing to map each exact atom or whatever, sure theres probably alot of ways you can automate alot of it via shortcuts etc but needing weeks to make one blueprint seems to suggest they are complex and hold alot of information. So i fail to see how reverse engineering a object would really require more space, am i missing something?

Likely because they use procedural generation: the blueprints for an object likely contain the instructions on how to build the object, rather than the piece-by-piece information of the object itself. The former is far smaller, and far easier to work with than the latter. For instance, if I wanted to make a monomolecular line of carbon atoms 1,000,000,000,000,000,000,000,000,000,000 atoms long, the procedural blueprint might look like this: [code]PLACE ATOM(C) AS X(1) LOOP Y = 2 TO 1000000000000000000000000000000 BIND ATOM(C) TO X(Y-1) AS X(Y) END LOOP FINISH[/code] On the other hand, the raw material data might look like this: [code]C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-C-...[/code] This would continue to repeat 1,000,000,000,000,000,000,000,000,000,000 times. The former is instructions, the latter is the raw product of output in digital form. You can even see this in current file formats: any procedural generation data takes up far less space than raw digital data.

Totalgit wrote:

Hmm, ok makes sense but barring the lack of storage space it would be possible for somebody/thing to look at the raw data and make the a procedural model from it? Or even somehow compress the data, maybe even compressing it as its actually scanning something. For example its scanning all those C-C-C-C-C's and after every 10 the scanning machine would just store it as multple blocks of 10xC- until the whole object is scanned, the act of scanning might take along time since as you say there is alot of data to process but i dont see how it couldnt compress the data as it went along too? Im only asking because im wondering about types of corporate espionage esp reverse engineering the latest secret stuff a hypercorp could try just stealing the blueprints but what if its only able to get ahold of a working physical object, it would no doubt try and reverse engineer it.

Sure you could, but looking at raw data would probably take a very... very long time. Even if you were to compress the data, you aren't going to be looking at decently small file sizes: my example was just a simple line of carbon atoms, and most objects aren't going to be anything near that simple. Blueprints will still be significantly simpler than the raw data that they will produce in the process of execution. No matter the case, creating blueprints from raw data would still require you to program the blueprints; creating blueprints simply by deconstructing an object isn't really possible. Even if you had a very advanced AI to create the blueprints from the raw data, it still sums to that AI doing a programming test using the raw data as work notes. As someone previously mentioned, there may also be certain factors inhibiting deconstruction and recording of the raw data. For instance, if the object is fragile, perhaps made of glass, and it cracks during the deconstruction process, the entire blueprint will be ruined. The same could occur if, say, the object has an unusual geometry and collapses due to a certain part being deconstructed. In fact, this could be used as a copyright protection scheme; something that is vacuum-sealed might contain easily-corroded components which are designed to destroy themselves if the object is deconstructed and the internals are exposed to open air or even mere light. This is not to say that reverse-engineering isn't a possibility, however. Nanotechnology is likely even used in the process... but it probably isn't as simple as plugging the object into a nanofabricator and pressing "decompile".

For scanning natural things from the environment (vs. designer items with protection), professional scanners would probably have an AI to help them compress the results. Writing new algorithms to reduce raw scan data into a more programmatic structure would be an active field of research, both for its own sake, and as a side-effect of attempts to reverse engineer protected devices.

Various roadblocks to copying protected designs:

• Parts need to be assembled in the correct order, or with appropriate scaffolding set up.
• Parts that assist construction (or prevent the device from operating during construction) are removed by the user before use, or by the machine at the end. E.g. battery pull tab.
• Use of chemicals that react with scaffolding commonly used by reverse engineers.
• Fractal or similar volumes. Make pieces of the design three-dimensional cross-sections of a complex mathematical function. It's easy to generate if you know the function and the appropriate parameters, but could be impossible to reverse engineer, especially if additional functions are applied to it after. For bonus points, make the way that the volume interacts with the rest of the object depend on the function used, so a replacement function can't be used instead.
• Portions that look like the above, but are generated randomly instead. Most sequences are incompressible.
• Integrated smarts that verify the integrity of the rest of the design, as well as being necessary for its normal functioning.
• AGIs that hunt down large-scale counterfeiting operations and send Reapers after them.

In the inner system where most fabs are locked, the fabs probably only print appropriately signed blueprints, meaning (a) it's easy to track a counterfeit design back to its designer, and (b) the price of a license is high enough it's not worth it for just a fancy pair of jeans.

Decivre wrote:

This is not to say that reverse-engineering isn't a possibility, however. Nanotechnology is likely even used in the process... but it probably isn't as simple as plugging the object into a nanofabricator and pressing "decompile".

I am inclined to agree with you on this point. However, I get the feeling that one of the shadier uses for nanotech which would spur the technology forward would be developing nanoswarms that assist in the process of reverse engineering... Just a random thought. Or plot point, should anyone wish to make use of it.