Weaponized code

This morning someone asked me if I could “weaponize” his code, i.e. turn his research software into something robust enough for production. I like that term.

12 thoughts on “Weaponized code

  1. To me the term sparked cyber security implications though. I would try “industrialize” but indeed “weaponize” sounds a lot better.

  2. Really, weaponize? You want your code to inflict damage on users? The furthest I usually go is “harden” (against attacks) or “idiot-proof” (against idiocy). Weaponize implies you’re going on the offensive.

  3. TomF: Weaponized for battle against the problem the user wants to solve, strong enough to work in an unforgiving environment.

  4. I usually “user-proof” my code, but “weaponize” has a certain ring of truth to it.

  5. Any chance the person you were working with is in Biochemistry or a related field? The term ‘weaponize’ there is used in this sense and for good reason: Many biochemical effects were detected in compounds which were modified to increase the effects to damaging levels (weaponizing) especially in the 1940-1970s period. Hello nerve gases etc.

    Many of the comments here seem to suppose ‘weaponize’ means ‘idiot-proof’. In fact it’s used to mean ‘make more potent’ or ‘isolate the core, useful essense of’. Also, while the origins of the word is violent, I often hear this term from doctors and other medical students describing the modification of medicines.

  6. I agree with TomF, weaponize doesn’t seem to be the right word here. At NASA, we used the term “flight-ready” to indicate robustness, even if the software wasn’t going aboard a vehicle; we’d also call robust software “mil spec” even though it wasn’t for the military. To weaponize means to give something fangs, not enhance its robustness.

  7. What’s harder, I think, is making underlying designs and algorithms robust, even if the code is solid. I’m not saying making code solid is not important. Indeed, a lot of “released code” seems to evade quality assurance, and is “made better” by being fixed after customers have problems with it. I agree with I. J. Kennedy, that perhaps my expectations are (too) high, as I spent 20 years in the military, aerospace, and embedded controls world (e.g., safety code for nuclear power reactors). Still, even there I sometimes saw algorithms originally designed for bombers transplanted into helicopters with but a modicum of additional simulation to verify that the original assumptions were met. Scary.

Comments are closed.