SHOT IN THE DARK: An Arms Race in the Digital Age

Not since The Graduate had “plastics” been so important. In the summer of 2013, Cody Wilson — then, a law student at University of Texas — decided to leave academia in pursuit of darker ideas. These concepts had been brewing in suburban basements, academic labs, and comic conventions for nearly a decade, and Wilson was determined to bring them to fruition. In that moment, societal conversations regarding manufacturing rights, design, individual liberty, and safety were transformed. 

 
Cody Wilson had become the first person to successfully fire a 3D printed handgun, The Liberator, in a test on a firing range outside of Austin. That moment was the culmination of an initiative Cody, amongst several other members of the fledgling consumer 3D printing community, had been working on for at least several years (1). This initiative didn’t only produce a physical object, however, it also brought to light a set of legal and technical challenges we’re still grappling with today.
 
It had long been possible, and in many cases legal, to hack together a combination of custom and off-the-shelf parts from places like hardware stores in order to build custom firearms(2-4). 3D printing, and the digitization of manufacturing at large, just made it easier to produce homemade firearms, with the printer handling most of the complex assembly that is so crucial to get right in order for the gun to function properly.
 
The Liberator test inspired further scrutiny towards existing laws regulating DIY firearms, as well as new industry and governmental policies geared specifically towards curbing the distribution of 3D printed guns. Aimed towards regulation in the physical space, however, these policies and proposed laws fail to take into consideration the fact that these homemade weapons begin as files on a computer and, as such, require a level of surveillance and enforcement in the digital space that is impossible to achieve. 
 
Instead of trying to direct resources towards regulations that would be at best ineffective and at worst damaging to the much more common innocuous uses of digital manufacturing technologies, we should instead focus our efforts on developing forensic technologies that are only just emerging in the digital manufacturing space, but which are poised to have a significant impact on the production and distribution of guns produced via 3D printing and other digital manufacturing technologies. 
 
 
 
 
To understand how we got to that fateful “shot heard ‘round the world” in 2013, however, we have to start the story about 10 years earlier, with the earliest manifestations of what would come to be known in the industry as “consumer 3D printing” and an ambitious mission to make self-replicating machines.
 
Although 3D printing — known more formally as “additive manufacturing” — can trace its origins back to the late 1980s, a combination of restrictive patents and high costs for mechatronic components and control systems restricted the technology to niche industrial applications (5). However, by the early 2000s, the expiration of those original patents, in combination with reduced costs around core hardware components, created an opportunity for people with enough patience and technical acumen to construct their own DIY 3D printers at home. While these original “desktop 3D printers” looked more like Erector Sets than futuristic pieces of manufacturing equipment, they were comparable to a subset of the machines being produced by contemporary industrial 3D printing companies (6-9). 
 
Though still a relatively new technology, additive manufacturing is also an accessible manufacturing process to replicate. Additive manufacturing, at its core, is all about the layering of one material on top of another in a controlled manner, typically at specific locations in 3D cartesian space. This is in contrast to what is widely considered “traditional manufacturing” — which is milling, cutting, and other subtractive operations that remove material from some initial block at specific locations in 3D space. 
 
The first 10 years of 3D printing research and development were largely dedicated to finding all the best ways to layer materials on top of each other to create a physical object (5, 10-13). Since the late 1990s, there have been three primary ways to facilitate additive manufacturing: by fusing fine powders, solidifying photoactive polymer resins, and bonding melted plastics.
 
Resin-based 3D printing involves the selective irradiation of a vat of liquid photopolymer to produce a 2D cross-section, or “slice”, of the desired 3D model. Once that layer has reacted to form a solid polymer, a stage moves in the vertical direction in order to expose a new layer of liquid resin and the next 2D slice of the model is exposed. The polymer resin’s natural affinity for itself enables the subsequent exposed resin layer to not only solidify in-plane, but also to the layer before it, enabling the inter-layer adhesion that makes all 3D printing possible.
 
Contrasted with resin-based 3D printing, the bonding of melted plastics, most commonly known as Fused Deposition Modeling, or FDM, uses the physical transformation of melting to allow flow spool of plastic filament to flow through a small nozzle which is moved around in three dimensions. For example writing out your name with a hot glue gun is the most basic form of FDM printing. It’s also worth noting that FDM is the printing technique Cody Wilson used to make the world’s first 3D printed gun.
 
Powder-based additive manufacturing offers by far the most diverse selection of materials, although different powder printing systems are required for almost every material option. In powder-based additive manufacturing, successive layers of powder are bound together through either the inkjet deposition of liquid adhesives, or the melting action of a high-powered laser. Compared to the systems that control resin and FDM 3D printers, powder-based printers are much more complex, much more difficult to operate, and almost exclusively used in research and heavy industrial settings, as opposed to residential, commercial, and light industrial settings.
 
Of these three approaches, FDM quickly became the most popular among early 3D printing tinkerers and hobbyists. This was due to the ease-of-use of its building material — solid plastic filament, rather than chemically irritating liquid resin or super-fine particles that could quickly become hazardous aerosols — in addition to the comparatively easy machine-construction process. As such, in the early 2000s an initiative known as the RepRap Project was launched to bring 3D printing to the masses by leveraging FDM technology (14). 
 
The mission of RepRap was simple, if ambitious: to make a machine that could make more of itself. Considering the vast majority of 3D printers can only work with plastics, this vision has yet to be 100 percent realized. The first few years after the RepRap project was launched, only around 40 percent of the components needed to build a functional FDM 3D printer could be printed by an existing FDM printer. The rest of the materials had to be supplied from external sources. Nevertheless, for the first time in history, it was becoming extremely easy to procure these parts. Online resources — cobbled together over time by a niche and dedicated community — coupled with cheap equipment made it feasible to find, purchase, and connect all of the essential non-printed components. Suddenly, anyone with a free weekend and enough patience could assemble their own 3D printer. Once they had done that, it would be significantly easier to make a second one, since they’d be able to print all of the plastic components themselves.
 
As this process evolved, people began to incorporate their own “mutations” and “adaptations,” which increased the percentage of the final printer construct that could be made on an existing printer. Ever-more advanced and creative constructions began generating a feedback loop — not only in engineering design but in sentiment. The sentiment permeating conversations in academic communities soon entered general society: someday in the near future, printers would be able to completely replicate not only themselves, but a myriad selection of other consumer goods (15-16).

It had long been legal to make your own firearm at home . What hadn’t been possible was the ability to easily make a firearm at home. Suddenly, the existing laws around DIY weapons didn’t seem sufficient.

However, academic communities failed to consider that the selection of consumer goods to be replicated would be anything but innocent or innocuous. Little did the academy know, the DIY-spirit driving the amateur engineer communities fostered an environment that directly resulted in Cody Wilson’s frightening project: a fully printable firearm. 
 
Tracking the development of the RepRap project, the push to produce printed guns didn’t immediately yield a 100 percent printed gun; however, the development of a fully 3D printed gun progressed much more quickly than the manufacturing project that kickstarted the consumer 3D printing movement itself. By 2013, virtually the entirety of the Liberator could be made on an FDM 3D printer (1). Soon after, designs for all-plastic 3D printed firearms began appearing on various forums and databases — chief among them Wilson’s own database and DIY firearms company Defense Distributed (2).
 
This online activity quickly attracted the attention of the U.S. State Department, which forced these gun blueprints — which took the form of digital file formats akin to PDFs, TXTs, and MP3s — off of major databases (1). However, the Napster and Limewire era made it evident that it was impossible to completely remove data from the internet, and files continued to emerge on various websites and shadowy links for years (2, 4, 17). By that time, though, 3D printed guns had largely moved out of the public view. This conversation was replaced by a broader debate surrounding the future of 3D printing itself, as the consumer sector began to question the technology’s seemingly inevitable ubiquity (18).
Even though the five years that followed the unveiling of the Liberator saw the collapse of the RepRap-germinated idea that everyone would have a 3D printer in their homes, the debate around 3D printed guns raged on, albeit out of the public eye. In response to the forced removal of their blueprints from the internet, Defense Distributed sued the State Department, citing a breach of First Amendment rights (19-20). Digital files, like the ones that comprised the Liberator plans, had traditionally been protected from censorship under the First Amendment. Yet these new forms of digital manufacturing created a unique conundrum: for the first time, a digital file could easily be translated to a physical object. 
 
By 2017, Defense Distributed v. United States Department of State had reached the 5th Circuit of the U.S. Appeals Court. The Appeals Court upheld the right to re-publish and distribute the 3D model files (21). Ultimately, the federal government settled with the company, and the files were once again made available online (20, 22-23). Just like five years before, a media firestorm swept the nation, and by the end of summer, legal actions in several states had largely removed the files from the internet. Although jurisprudence surrounding 3D printed guns had expanded by this time, there were still methods for Wilson and Defense Distributed to get digital blueprints  to customers in some states (24-25). However, the 2017 lawsuit proved to be significant, as it brought the lackluster regulations surrounding DIY weaponry to the forefront of the public consciousness. 
 
The digital manufacturing industry did not stand still in the years following Wilson’s breakout moment. In the five years between the original posting of the Liberator and its return to Defense Distributed’s website, both additive and subtractive manufacturing advanced dramatically. Not only did additive manufacturing become more reliable and precise, but subtractive manufacturing began to transition from all-industrial ecosystems into home ecosystems. This created an even more momentous shift in the manufacturing sector, as many consumer milling machines were capable of working with metal, which is what conventional firearms are made from. During this time, Defense Distributed released a low-cost milling machine designed specifically to work with firearm precursor metals and partially-completed gun stocks (26). Although the company was fighting a lawsuit determining the legality of digital blueprint distribution, these other actions were completely legal, taking advantage of laws that could never have anticipated the rise of personal manufacturing technologies precipitated by manufacturing advances in the early 2000s.
 
It had long been legal to make your own firearm at home (27-30). What hadn’t been possible was the ability to easily make a firearm at home. Suddenly, the existing laws around DIY weapons didn’t seem sufficient, and 3D printers were the least of peoples’ worries.
 
In 2018, if one wanted to make their own AR-15-style rifle, the best option would not be downloading a model for 3D printing. The layer-wise deposition fundamental to all 3D printing often results in slightly weaker bonds between layers than within layers. As a result, there’s a significant chance such a printed firearm would at best fail to fire due to loss of pressure through tiny gaps between layers, and at worst, explode in the user’s hands (31-32).
 
Traditional gun manufacturing utilizes subtractive milling and drilling operations.  With subtractive manufacturing, the object is constructed from a stock material that is already perfectly bonded to itself in all directions. The machining operations merely shave material until the desired final form is achieved. As it became financially and infrastructurally feasible to own milling machines capable of performing these operations in homes, it also became more feasible to produce a more robust, metal firearm in homes.
 
Still, owning a personal milling machine doesn’t necessarily translate to knowing how to manufacture a gun. Existing legal loopholes, however, have allowed for a scenario where you don’t need machining knowledge in order to construct your own semi-automatic rifle. In addition to the milling machine purchasable from Defense Distributed, unfinished stocks of an AR-15 can also be purchased, and subsequently can be machined to the final form with a purchased milling machine, using digital machining instructions provided to you by Defense Distributed as a guide (33). None of that is illegal.
 
Even though a homemade, metal-machined gun stock will show up on a metal detector while a printed all-plastic one will not, both weapons are just as untraceable, just as unregistered, and just as easy to make without the awareness of any authorities (34-35).
 
So, what are we to do? In the five years following the launch of Defense Distributed, the status quo looked very much like the protocol that was adopted following the Napster era of music — the Digital Millennium Copyright Act, or DMCA. Under the DMCA, the law’s attitude towards all content improperly shared online was “notice and takedown” — if one were to come across something illegal on a website, that person is tasked with notifying the maintainers of the website, who are then tasked with removal of the offending media (36-37). 
 
This puts the legal responsibility upon website managers. Unless these people could remain ever-vigilant, constantly scanning the content posted to their pages and immediately removing offending content when it appeared, files like the plans to printable or millable guns could make their way around the internet and end up downloaded during the lag time between the “notice” and the “takedown.” Should the manager be held responsible if they can’t instantaneously remove blueprints to the Liberator from their website and someone subsequently downloads the model, prints it, and uses it to lethal effect? Policing the internet is a monumental task, and “notice and takedown” methods are likely not the most effective means to that end.
 
Considering the major shortcomings of the DMCA protocol, others have proposed that the manufacturers of these fabrication machines themselves serve as the gatekeepers between weapons files and the machines that make them. At face value, the plan seems sound enough — just program the printers and milling machines to not fabricate weapons if such a file is uploaded. These plans for firearms are just code anyway, so use more code to keep them from getting made. However, altering some part of the code encrypting the plans for the firearm, which may not even impact the structure or function of the gun itself, could once again render the model 3D printable (38). 
 
There are also many more complex and harder-to-detect ways one could modify the code and/or form of a digitally manufactured firearm without reducing its ability to function as a firearm (39). In the digital arms race that would ensue between manufacturers and bad actors, it would only take one slip through the cracks to result in a bespoke firearm making its way into the world. Policing the internet is very, very hard, but algorithmically predicting every way someone could manufacture a weapon is more difficult and likely impossible, not to mention such an endeavor could quickly become the single largest drain on resources in the manufacturing industry.
 
Perhaps even more alarming than the technical infeasibility of blocking completely homemade firearms from being produced is the fact that finishing a pre-made gun in one’s own home is both legal and an easier way of obtaining an effective weapon. One of the few good things to come out of the renewed debate around homemade firearms was the impetus to produce new legislation limiting people’s capacity to make homemade weapons.
 
A prime example of this new legislation is H.R. 7115, which was introduced in the 115th Congress. The bill, titled the “3D Firearms Prohibition Act,” seeks to prevent the distribution of firearms parts kits (40). Unfortunately, even this new proposal requires a logistically ineffective implementation strategy. This bill  “require[s] homemade firearms to have serial numbers…” But who is going to make check in with everyone who has a 3D printer, or mill, or laser cutter in their home to make sure their firearms are imprinted with a serial number? Or will there be tighter controls on files that can be downloaded? How do you stop someone from making a design of their own on a computer at home without tracking everything they do on that computer? What if they’re designing the model offline? How can you possibly even identify every design variant that will enable ballistic action? Ultimately, although the bill would make it illegal to sell and distribute parts kits, it only prevents actual companies, like Defense Distributed, from participating, rather than deterring individuals. At the individual level, the same technical problems occur as before, except now even more kinds of manifestations of weapons would have to be tracked.
 
All of these technical issues around the regulation and control of 3D printed guns certainly makes it feel as though we’re on the verge of an impending catastrophe. There’s still ample time, however, to effectively address these issues. For the time being it’s definitely still easier to get a standard gun illegally than it is to print or machine your own, parts kit or no kit, so it’s unlikely that any active shooter would be using a 3D printed weapon. That being said, the risks of homemade firearms shouldn’t go ignored. Although it’s clear that trying to police and regulate DIY weapons out of existence is likely to be ineffective, there are emerging technologies and approaches that are worth dedicating resources to. These technologies, while not rendering home weapon production impossible, would establish parity between homemade guns produced illegally and standard guns procured illegally.

…policies and proposed laws fail to take into consideration the fact that these homemade weapons begin as files on a computer and, as such, require a level of surveillance and enforcement in the digital space that is impossible to achieve.

At the 2018 Association for Computing Machinery Conference on Computer and Communications Security, a team of researchers from SUNY Buffalo, Rutgers, and Northeastern University presented a paper on Prin Tracker: a new software system for the identification of homemade, 3D printed firearms. The core concept underlying PrinTracker is that “3D printers possess unique fingerprints, which arise from hardware imperfections during the manufacturing process, causing discrepancies in the line formation of printed physical objects” (39).
 
The PrinTracker software works by using both 3D scanners as well as conventional 2D scanners to capture the surface texture of the printed weapon. This surface texture captures the “fingerprint” of the printer that produced it. Ultimately, the printer’s “fingerprint” is processed and given quantitative metrics that describe the geometry of the surface texture (39). By clustering the processed texture data and comparing it with known printer models, the make and model of the printer that produced the weapon in question can be determined, a process equivalent to determining the make and model of an unregistered conventional firearm from forensic data such as the patterning on a fired bullet casing (41-45).
 
This isn’t as outlandish at it may sound. Small variations in models produced by mechatronic processes are inevitable and can trace a product’s origin to a specific manufacturing device (46-49). Anyone intimately acquainted with manufacturing can tell you by what process — e.g., printing, milling, bending, injection molding — a particular item was produced. Those people who are experts in their field and method of manufacture can even tell you which particular factory produced it. Every 3D printer deposits each layer in a slightly different way (50). Those differences are enough to confidently determine which printer fabricated the firearm in question. The same is true for other methods of manufacture, and these are just the things we can spot with our own senses. Algorithmic based methods such as PrinTracker are even more sensitive, and can hold up against tampering that would attempt to remove these inherently embedded “fingerprints” (39, 51-53).
 
These digital forensic approaches, especially those focusing on 3D printing and other forms of digital manufacturing, are still very much in their infancy. Nevertheless, the results they have demonstrated so far are extremely promising and offer a solution that would allow digitally manufactured firearms to be traced and investigated in a way comparable to the forensic techniques already employed to trace illegally-procured firearms. Such an approach would not force an undue burden upon manufacturers, websites, or other internet-based services to be the gatekeepers and regulators of a problem they simply can’t contain. No one holds Smith & Wesson, a prominent gun manufacturer, responsible when a shooting occurs with a conventional firearm. The manufacturers that produce 3D printers and milling machines, and the people that maintain file-sharing repositories — which are vast sources of useful and beneficial content — should not be held responsible when crimes are committed with weapons manufactured on their devices. It could be argued that there should be less burden placed upon these parties. A manufacturer of 3D printers produces a device that is intended to be used for more manufacturing — such a product could be used for beneficial purposes just as easily as it could be used for malicious ones.
 
In a perfect world, violence, including gun violence, would be nonexistent and this discussion would be irrelevant. But this world is nowhere near perfect. While we should undoubtedly direct energy to work towards shaping a safer, kinder world, we must do so by also addressing the problems still present today. On the whole, 3D printed, or otherwise digitally manufactured firearms are not a major public safety concern, nor will they be for some time to come. Nevertheless, these manufacturing technologies do present increasingly societally relevant problems lacking adequate solutions, and while the efforts that have been made so far are commendable, their structure and implementation bely a lack of understanding of the technologies that are enabling these issues in the first place. The promise of digital forensic solutions, such as PrinTracker, is that they leverage the existing manufacturing infrastructure to provide crucial data that allows public safety authorities to approach the problem in a manner comparable in structure in function to the approaches they already employ with conventional firearms. Most importantly, these digital forensic techniques acknowledge that, in a world where individuals have access to advanced manufacturing technology in their own homes, it will be impossible to prevent the fabrication of weapons before it happens. We should not dedicate our resources to goals that are impossible to achieve; by those metrics we will never succeed. Rather, we should direct our energies towards the further development of technologies that provide us with the best chance of successfully investigating malicious acts when they do occur. We can’t stop people from 3D printing guns, but we can make it in their best interest to choose not to.
 
Works Cited
 
  1. Holpuch, A., & London, E. M. C. A. in. (2013, May 10). State Department orders firm to remove 3D-printed guns web blueprints. The Guardian. 
  2. 3D Printing Community Updates Liberator with Rifle, Pepperbox and Glock-Powered “Shuty-9.” (2013, July 1). Retrieved January 19, 2019, from https://www.guns.com/news/2013/07/01/3d-printing-community-updates-liberator-with-rifle-pepperbox-and-glock-powered-shuty-9
  3. After 3D Printable Gun Bans, The Pirate Bay Steps In to Distribute Plans. (2013, May 11). Retrieved January 19, 2019, from https://www.technobuffalo.com/2013/05/11/pirate-bay-distributes-3d-printed-
    gun-plans/
  4. Kolawole, E. (2013, May 9). Plans for 3D-printed gun downloaded 100,000 times; State Department in contact with Defense Distributed. The Washington
    Post.
  5. Backeris, P., & Borrello, J. (2017). Rapid Prototyping Technologies. In Rapid Prototyping in Cardiac Disease - 3D Printing the Heart (1st ed.). Springer International Publishing AG.
  6. Jerez-Mesa, R., Travieso-Rodriguez, J. A., Corbella, X., Busqué, R., & Gomez-Gras, G. (2016). Finite element analysis of the thermal behavior of a RepRap 3D printer liquefier. Mechatronics, 36, 119–126.
  7. Kentzer, J., Koch, B., Thiim, M., Jones, R. W., & Villumsen, E. (2011). An open source hardwarebased mechatronics project: The replicating rapid 3-D printer. In 2011 4th International Conference on Mechatronics (ICOM) (pp. 1–8).
  8. Pearce, J. M., Wijnen, B., & Anzalone, G. C. (2015). Multi-material additive and subtractive prosumer digital fabrication with a free and open-source convertible delta RepRap 3-D printer. Rapid Prototyping Journal, 21(5), 506–519.
  9. Tymrak, B. M., Kreiger, M., & Pearce, J. M. (2014). Mechanical properties of components fabricated with open-source 3-D printers under realistic environmental conditions. Materials & Design, 58, 242–246.
  10. Crump, S. S. (1992). US5121329 A. 
  11. Deckard, C. R., Beaman, J. J., & Darrah, J. F. (1992). US5155324 A.
  12. Hull, C. W. (1986). US4575330 A.
  13. Hull, C. W., Spence, S. T., Albert, D. J., Smalley, D. R., Harlow, R. A., Steinbaugh, P., … Remba, D. Z. (1991). US5059359 A.
  14. 3D Printing and Humanity’s First Imperfect Replicator. (2014). 3D Printing and Additive Manufacturing, 1(1), 4–5.
  15. Harrod, H. (2012, May 3). Make your own: the 3D printing revolution.
  16. The disruptive future of printing. (2010, April 30).
  17. Brian Krassenstein. (2015, June 18). Users can Download the Liberator Gun on 3DSha.re for Free. 3DPrintBoard.com. Thread. Retrieved January 18, 2019, from https://3dprintboard.com/showthread.php?13769-Users-can-Download-the-LiberatorGun-on-3DSha-re-for-Free
  18. Stevenson, K. (2017, October 27). 3D Printing: The Story So Far. Retrieved January 19, 2019, from https://www.fabbaloo.com/blog/2017/10/27/3d-printingthe-story-so-far
  19. Defense Distributed v. United States Department of State. (2017). Harvard Law Review, 130(6), 1644–1751. 
  20. Winick, E. (2018, July 10). The files you need to make your own gun can now be legally shared online. MIT Technology Review.
  21. Greenberg, A. (2018, July 10). A Landmark Legal Shift Opens Pandora’s Box for DIY Guns. Wired. 
  22. Israel, S. (2018, July 19). The Age of the Downloadable Gun Begins. The New York Times. 
  23. Sherfinski, D. (2018, July 22). Gun company wins legal fight to post 3D printable gun plans online. The Washington Times.
  24. Montgomery, D., & Hsu, T. (2018, December 6). Blocked From Posting Printable Gun Plans, Activist Will Mail Them Instead. The New York Times. 
  25. Shear, M. D., Hsu, T., & Johnson, K. (2018, December 6). Judge Blocks Attempt to Post Blueprints for 3-D Guns. The New York Times.
  26. Murphy, M. (2018, July 31). 3D printed guns are here, but that’s not what we should be worried about [News].
  27. 18 U.S. Code § 922 - Unlawful acts. (n.d.). U.S. Code.
  28. 26 U.S. Code § 5822 - Making. (n.d.). U.S. Code.
  29. Bureau of Alcohol, Tobacco, Firearms and Explosives. (2017). Does an individual need a license to make a firearm for personal use? (FAQ).
  30. Zezima, K. (2018, August 1). Despite ruling on 3-D-printed guns, it remains legal to make your own gun at home. The Washington Post.
  31. Page, L. (2013, May 10). “Liberator”: Proof that you CAN’T make a working gun in a 3D printer. The Register.
  32. Richt, J. (2013, May 5). “Liberator” 3D-printed handgun fails after single shot in Finnish test. Yle Uutiset.
  33. Ghost Gunner. (n.d.). [E-Commerce Site]. Retrieved from https://ghostgunner.net
  34. 3D Printing and the Future (or Demise) of Intellectual Property. (2014). 3D Printing and Additive Manufacturing, 1(1), 34–43.
  35. Carrots, Not Sticks: Rethinking Enforcement of Intellectual Property Rights for 3D-Printed Manufacturing. (2014). 3D Printing and Additive Manufacturing, 1(1), 44–51.
  36. Cobia, J. (2008). The Digital Millennium Copyright Act Takedown Notice Procedure: Misuses, Abuses, and Shortcomings of the Process. Minnesota Journal of Law, Science & Technology, 10, 387.
  37. Murtagh, M. P. (2009). The FCC, the DMCA, and Why Takedown Notices Are Not Enough. Hastings Law Journal, 61, 233.
  38. Firearms: identifying information, Pub. L. No. AB857, § Chapter 60 (2016).
  39. Li, Z., Rathore, A. S., Song, C., Wei, S., Wang, Y., & Xu, W. (2018). PrinTracker: Fingerprinting 3D Printers Using Commodity Scanners. In Proceedings of the 2018 ACM SIGSAC Conference on Computer and Communications Security (pp. 1306–1323). New York, NY, USA: ACM.
  40. Pallone, F. H.R.7115 - 115th Congress (2017-2018): 3D Firearms Prohibitions Act, Pub. L. No. H.R.7115 (2018).
  41. Barnes, F. C., & Helson, R. A. (1974). An Empirical Study of Gunpowder Residue Patterns. Journal of Forensic Science, 19(3), 448–462.
  42. Chang, K. H., Jayaprakash, P. T., Yew, C. H., & Abdullah, A. F. L. (2013). Gunshot residue analysis and its evidential values: a review. Australian Journal of Forensic Sciences, 45(1), 3–23.
  43. Nichols, R. G. (1997). Firearm and Toolmark Identification Criteria: A Review of the Literature. Journal of Forensic Science, 42(3), 466–474.
  44. Smith, C. L. (1997). Fireball: a forensic ballistics imaging system. In Proceedings IEEE 31st Annual 1997 International Carnahan Conference on Security Technology (pp. 64–70).
  45. Smith, C. L., & Cross, J. M. (1995). Optical imaging techniques for ballistics specimens to identify firearms. In Proceedings The Institute of Electrical and
    Electronics Engineers. 29th Annual 1995 International Carnahan Conference on Security Technology (pp.275–289).
  46. Buchanan, J. D. R., Cowburn, R. P., Jausovec, A.-V., Petit, D., Seem, P., Xiong, G., … Bryan, M. T. (2005). Forgery: ‘Fingerprinting’ documents and packaging. Nature, 436(7050), 475.
  47. Clarkson, W., Weyrich, T., Finkelstein, A., Heninger, N., Halderman, J. A., & Felten, E. W. (2009). Fingerprinting Blank Paper Using Commodity Scanners. In 2009 30th IEEE Symposium on Security and Privacy (pp. 301–314).
  48. Holzmond, O., & Li, X. (2017). In situ real time defect detection of 3D printed parts. Additive Manufacturing, 17(Supplement C), 135–142. 
  49. Verbauwhede, I., & Maes, R. (2011). Physically Unclonable Functions: Manufacturing Variability As an Unclonable Device Identifier. In Proceedings of the 21st Edition of the Great Lakes Symposium on Great Lakes Symposium on VLSI (pp. 455–460). New York, NY, USA: ACM.
  50. Muhs, D., Wittel, H., Becker, M., Jannasch, D., & Voßiek, J. (2003). Roloff/Matek Maschinenelemente: Normung, Berechnung, Gestaltung - Lehrbuch und Tabellenbuch (16th ed.). Vieweg+Teubner Verlag. 
  51. Gassend, B., Clarke, D., van Dijk, M., & Devadas, S. (2002). Silicon Physical Random Functions. In Proceedings of the 9th ACM Conference on Computer and Communications Security (pp. 148–160). New York, NY, USA: ACM. 
  52. Guajardo, J., Kumar, S. S., Schrijen, G.-J., & Tuyls, P. (2007). FPGA Intrinsic PUFs and Their Use for IP Protection. In P. Paillier & I. Verbauwhede (Eds.), Cryptographic Hardware and Embedded Systems - CHES 2007 (pp. 63–80). Springer Berlin Heidelberg.
  53. Lee, J. W., Lim, D., Gassend, B., Suh, G. E., Dijk, M. van, & Devadas, S. (2004). A technique to build a secret key in integrated circuits for identification and
    authentication applications. In 2004 Symposium on VLSI Circuits. Digest of Technical Papers (IEEE Cat. No.04CH37525) (pp. 176–179).