I mean, there is the field of critical systems engineering. The problem is that every schmuck with a computer and some venture capital is off to “disrupt” something, regardless of how much was already learned in that field by the previous generations.
Lol, I wrote contractor code for DoD. Obviously, DoD wants really good security on their code. One particularly bad project I ended up as a subcontractor on, management kept insisting that what we were coding was a prototype, and we could add in the security in the actual project. And all us coders were like, “No, you’re having us write the actual project and the security has to be designed into it from the base up, ‘adding it in later’ like you won’t admit you’re planning on doing will leave way too many places for security holes to occur. Let us stop programming this shit and design some actual security and then get back to work.” We were told “lol, no, you don’t know what you’re talking about, this is just a prototype, get back to work.”
We had little buttons printed up saying, “Don’t worry: this is just the prototype, we’ll do the real programming later.”
Of course, two years later, the “prototype phase” ends, and management comes to us and says, “Hey, okay, so we’ve decided that what you’ve been working on is what we’re actually going to ship. You need to go back and make it fit all these really-high-level-security requirements.” Which of course would mean going through all this code and essentially redesigning and rewriting over half of it from scratch. Over half the coders were gone in six weeks.
The problem isn’t with testing (which is an issue), but standardization on specific solutions. When everyone needs to use the same thing, it’s a lot more valuable to attack it.
So what we need is more alternatives that work together.
I don’t know anything about the trucking industry, so I’ll use IT instead. A lot of companies standardize on Cisco equipment, so when there’s a breach, everyone is screwed. The problem isn’t that Cisco is insecure, it’s that Cisco is ubiquitous, so one breach screws over everyone. If networking equipment was more a la carte, it’s unlikely a breach would impact all of the equipment used (e.g. a Mikrotik Router, Mikrotik Switch, Ubiquiti Access Points, etc). But bundling solutions is the name of the game for these large operations, which increases the fallout from a breach.
That’s why Windows gets so many viruses, it’s not because Windows sucks (it does), it’s because it’s such a huge target and you’ll get so much more value from attacking it than attacking a potentially easier target.
So much technology these days seems to be just rolled out without any real world testing done.
I’m in my third decade in IT, and I can assure you such testing was never a priority. We just have capable computers everywhere nowadays .
I mean, there is the field of critical systems engineering. The problem is that every schmuck with a computer and some venture capital is off to “disrupt” something, regardless of how much was already learned in that field by the previous generations.
Critical Systems Engineering? Sounds expensive. Nah fuck that, let’s write the code in VB6 and stick it on an unsecured Windows NT box.
Lol, I wrote contractor code for DoD. Obviously, DoD wants really good security on their code. One particularly bad project I ended up as a subcontractor on, management kept insisting that what we were coding was a prototype, and we could add in the security in the actual project. And all us coders were like, “No, you’re having us write the actual project and the security has to be designed into it from the base up, ‘adding it in later’ like you won’t admit you’re planning on doing will leave way too many places for security holes to occur. Let us stop programming this shit and design some actual security and then get back to work.” We were told “lol, no, you don’t know what you’re talking about, this is just a prototype, get back to work.”
We had little buttons printed up saying, “Don’t worry: this is just the prototype, we’ll do the real programming later.”
Of course, two years later, the “prototype phase” ends, and management comes to us and says, “Hey, okay, so we’ve decided that what you’ve been working on is what we’re actually going to ship. You need to go back and make it fit all these really-high-level-security requirements.” Which of course would mean going through all this code and essentially redesigning and rewriting over half of it from scratch. Over half the coders were gone in six weeks.
I still have my nifty little button, though.
The problem isn’t with testing (which is an issue), but standardization on specific solutions. When everyone needs to use the same thing, it’s a lot more valuable to attack it.
So what we need is more alternatives that work together.
I don’t know anything about the trucking industry, so I’ll use IT instead. A lot of companies standardize on Cisco equipment, so when there’s a breach, everyone is screwed. The problem isn’t that Cisco is insecure, it’s that Cisco is ubiquitous, so one breach screws over everyone. If networking equipment was more a la carte, it’s unlikely a breach would impact all of the equipment used (e.g. a Mikrotik Router, Mikrotik Switch, Ubiquiti Access Points, etc). But bundling solutions is the name of the game for these large operations, which increases the fallout from a breach.
That’s why Windows gets so many viruses, it’s not because Windows sucks (it does), it’s because it’s such a huge target and you’ll get so much more value from attacking it than attacking a potentially easier target.