8 ms·

Well, what if that's not how it really works, even for military applications? You know, a lot of the military software is made by private corporations. It's don

Well, what if that's not how it really works, even for military applications? You know, a lot of the military software is made by private corporations. It's done under government supervision, but it's not made by a government "Department of Software". Consider, too, that the government doesn't generally build roads -- that would imply every state or the federal government would have some civilian equivalent to the Army Corps of Engineers. Generally, the government contracts out to specialized road building firms. The road building companies are better at building roads, better at fighting with the road building machine companies, and better at getting good raw materials than the government, while the government is more focused on the road being right in a bunch of ways -- being in the right place, having the right safety features (so the government doesn't have to deal with too many accidents), dovetailing with other roads made by other road companies -- and that is how things actually work out. There are, I think, pretty good reasons for that. What do you think? This structure also shows up with non-profits. For example, a foundation wants a museum. They work with various specialized, for-profit companies to build the building, put in the art preservation system, put in the security system, &c. However, even though it would seem everything could happen with just companies, the truth is, the composition of their efforts that makes a museum would not make sense if it were animated solely by those companies pursuing profits -- it is only the beneficial purpose of the museum that brings their efforts together in that way. This points to a consistent, repeatable composition of organizations as a solution; and it's not something so novel. The government may be able to manage something like Postgres; but I suspect it is more like a museum: we need a certain kind of stakeholder in the governance of the organization that is simultaneously broader and more narrow than the government. For infrastructure software, social networks and similar things, we want, on the one hand, an international assortment of stakeholders -- and we also want stakeholders who have some kind of real connection to software.

dragontamer 2y ago
We can change that though. I think everyone agrees that Ghidra is pretty good for example. Ironically yes this is DoD project but surely it's possible to figure out how that team is setup and build out other software teams.
- solidsnack9000 2y ago
  I'm not sure that I understand how we can change it from your reply, though; or what we could change. Some things seem to be invariants here: people need to get paid (or they will do something else); corporations are a good way to organize people to get paid to do something, whereas government is not; corporations are not organized for the public benefit and have to be balanced out by some other institution for public goods, like roads and many others, where aligning incentives through the market is too difficult. What do you disagree with, in all that? Now from the example you cite, and so many others, one direction we might take is that actually DoD projects are what works and so we just need to do more of them. This actually worked okay for the early history of the internet and we got quite a few decentralized, trustworthy communication mechanisms as a result -- the core internet protocols, email, the web. That seems to be tied to a certain era, though; so that's why I'm looking for alternatives. One option would be to do something like a museum: a non-profit that takes money from the public (sometimes as tickets and sometimes as grants) and then pays people (directly or through contractors) to accomplish something in the public good.
  - dragontamer 2y ago
    > That seems to be tied to a certain era, though; https://github.com/NationalSecurityAgency/ghidra https://github.com/NationalSecurityAgency/ghidra I'm talking about modern software, Ghidra. > What do you disagree with, in all that? It seems like a balance has been achieved somewhere in the US Government. Maybe Ghidra was an accident but if it could happen once, we can analyze the team behind it, the circumstance and create new government organizations that follow it. And it can then serve as the basis of a model for other public-good software. A lot of video game reverse-engineerers now use Ghidra for video game analysis. Its a really good tool. I can imagine that the security / infosec / bug-hunting communities would also like a tool like Ghidra. So it sounds good for a net-positive effect for the public good. Its just kind of unexpected, that's all. But its still US Government software that has a good niche. ------------ I've seen some other analysis like Argonne National Labs GREET (https://greet.anl.gov/ https://greet.anl.gov/), which helps analyze different supply-chain models for how much greenhouse gasses exist. And thanks to GREET, there are large scale analysis into what kind of EVs contribute to more greenhouse-gas emissions. A subject that requires deep analysis of the various chemistries that go into different EVs: NCA vs LiFePo4 vs NCM etc. etc.. As well as hypothetical future chemistries like Silicon or Sodium batteries. So we are now better able to focus on better solutions moving forward. GREET sucks as a UI but the features are clearly good. -------- Whenever I think of US Government software, these are the examples that come up. I think its... fine. The Government should make more tools for the benefit of society. GREET isn't as well written as Ghidra but its still software that matches the profile of public good that you and I are talking about.

dragontamer 2y ago

We can change that though. I think everyone agrees that Ghidra is pretty good for example. Ironically yes this is DoD project but surely it's possible to figure out how that team is setup and build out other software teams.

solidsnack9000 2y ago
I'm not sure that I understand how we can change it from your reply, though; or what we could change. Some things seem to be invariants here: people need to get paid (or they will do something else); corporations are a good way to organize people to get paid to do something, whereas government is not; corporations are not organized for the public benefit and have to be balanced out by some other institution for public goods, like roads and many others, where aligning incentives through the market is too difficult. What do you disagree with, in all that? Now from the example you cite, and so many others, one direction we might take is that actually DoD projects are what works and so we just need to do more of them. This actually worked okay for the early history of the internet and we got quite a few decentralized, trustworthy communication mechanisms as a result -- the core internet protocols, email, the web. That seems to be tied to a certain era, though; so that's why I'm looking for alternatives. One option would be to do something like a museum: a non-profit that takes money from the public (sometimes as tickets and sometimes as grants) and then pays people (directly or through contractors) to accomplish something in the public good.
- dragontamer 2y ago
  > That seems to be tied to a certain era, though; https://github.com/NationalSecurityAgency/ghidra https://github.com/NationalSecurityAgency/ghidra I'm talking about modern software, Ghidra. > What do you disagree with, in all that? It seems like a balance has been achieved somewhere in the US Government. Maybe Ghidra was an accident but if it could happen once, we can analyze the team behind it, the circumstance and create new government organizations that follow it. And it can then serve as the basis of a model for other public-good software. A lot of video game reverse-engineerers now use Ghidra for video game analysis. Its a really good tool. I can imagine that the security / infosec / bug-hunting communities would also like a tool like Ghidra. So it sounds good for a net-positive effect for the public good. Its just kind of unexpected, that's all. But its still US Government software that has a good niche. ------------ I've seen some other analysis like Argonne National Labs GREET (https://greet.anl.gov/ https://greet.anl.gov/), which helps analyze different supply-chain models for how much greenhouse gasses exist. And thanks to GREET, there are large scale analysis into what kind of EVs contribute to more greenhouse-gas emissions. A subject that requires deep analysis of the various chemistries that go into different EVs: NCA vs LiFePo4 vs NCM etc. etc.. As well as hypothetical future chemistries like Silicon or Sodium batteries. So we are now better able to focus on better solutions moving forward. GREET sucks as a UI but the features are clearly good. -------- Whenever I think of US Government software, these are the examples that come up. I think its... fine. The Government should make more tools for the benefit of society. GREET isn't as well written as Ghidra but its still software that matches the profile of public good that you and I are talking about.