4 ms·

They Write the Right Stuff

25 points by prakash 19y ago

Husafan 19y ago
This is one of the coolest things I have ever read. Thanks to OP.
mynameishere 19y ago
It's been posted around a thousand times, and the standard criticism still holds: The space program represents a unique, unreproducible environment in which quality is a benefit that outweighs every possible cost.
- mattmaroon 19y ago
  Exactly. I imagine writing bug free code is a lot like site uptime, where each 9 added after the decimal place doubles the cost. There are few software companies (or maybe none) where debugging to that extent would make fiscal sense.
- yters 19y ago
  It's crazy the amount of effort they put into the process, and yet the slightest computer bug blows up a whole shuttle. I forget which shuttle it was, but an arithmetic bug was replicated in both failsafe systems, and it was exactly this bug that caused the shuttle to crash. Nowhere is Murphy's law more true.
  - hollerith 19y ago
    There have been two losses of shuttles: 1986 with the failure of the O-ring in the solid-rocket booster and 1996 with the foam that fell off during lift-off that damaged a tile on the leading-edge of a wing which caused the shuttle to burn up on re-entry. Which loss are you talking about when you say, <i>it was exactly this bug that caused the shuttle to crash</i>?
    - aston 19y ago
      Not shuttle. Mars probe. And it was a bad unit conversion.
      - yters 19y ago
        Yeah, that's the one.
- omnipath 19y ago
  No, it's not a unique situation. Or rather, it's only unique in one way: if failure occurs, it occurs loudly, spectacularly, and in front of millions of people. The article itself mentioned the Denver airport, where billions of dollars still didn't prevent software errors from happening, delaying a massive project. And let's not forget the cascading failures that happened just last year because of bad weather and poor planning/routing software programs. Every time this article is posted, and it's been around since at least 2004, the one thing that strikes me profoundly is how great communications play an integral part in the entire affair. It's getting to the point where we as software programmers and engineers can't improve our software and craft until we learn the importance of communication, between ourselves and our clients. Yeah, we talk about it, but we don't really implement it. It's like saying we understand the need to watch what we eat to be healthy, and yet still eating a McDonalds value meal every meal, only with Diet Soda instead regular. We need to improve our all parts of our craft to grow. Technically is alright. The people part, the interaction part, the expectation part, goal-setting, testing, planning needs to improve tremedously. We know the problem, and we must provide the solution. If a client doesn't know what they want, ask more information. Educate the client about how the development process should work. Find out more about the industry. See what was wrong/troublesome about the old way they did their business. If information isn't making it's way from the managers to the coders/planners, find the problem and eliminate it. This is not new. I am not smart, and I know these problems. Surely, we can do better. Yes, the projects will probably end up taking longer. Yes, development will probably end up costing more. But it will be better. And the clients will be happier. p.s. YC.news needs a preview option instead of just a reply option.
Hexstream 19y ago
"The answer is, yes, the process does stifle creativity. [...] It's the process that offers a template for any creative enterprise that's looking for a method to produce consistent - and consistently improving -- quality." Come again?... "In the modern software environment, 80% of the cost of the software is spent after the software is written the first time -- they don't get it right the first time, so they spend time flogging it." I'd say most sofware cannot be gotten right the first time...
noonespecial 19y ago
Cool story, but don't run your startup like that. If you do, you will deliver perfect software, years too late, that nobody wants.
- Husafan 19y ago
  Not to say that every project needs to be planned like NASA, but I would like to know how many start-ups have calculated their acceptable rate of failure (in their software) prior to launch.
johnrob 19y ago
If normal software buyers were as averse to defects as NASA, all software development would probably be like this. However, normal customers tend to prefer more features and overlook defects (relatively speaking).

25 points by prakash 19y ago

This is one of the coolest things I have ever read. Thanks to OP.

It's been posted around a thousand times, and the standard criticism still holds: The space program represents a unique, unreproducible environment in which quality is a benefit that outweighs every possible cost.

mattmaroon 19y ago
Exactly. I imagine writing bug free code is a lot like site uptime, where each 9 added after the decimal place doubles the cost. There are few software companies (or maybe none) where debugging to that extent would make fiscal sense.
yters 19y ago
It's crazy the amount of effort they put into the process, and yet the slightest computer bug blows up a whole shuttle. I forget which shuttle it was, but an arithmetic bug was replicated in both failsafe systems, and it was exactly this bug that caused the shuttle to crash. Nowhere is Murphy's law more true.
- hollerith 19y ago
  There have been two losses of shuttles: 1986 with the failure of the O-ring in the solid-rocket booster and 1996 with the foam that fell off during lift-off that damaged a tile on the leading-edge of a wing which caused the shuttle to burn up on re-entry. Which loss are you talking about when you say, <i>it was exactly this bug that caused the shuttle to crash</i>?
  - aston 19y ago
    Not shuttle. Mars probe. And it was a bad unit conversion.
    - yters 19y ago
      Yeah, that's the one.
omnipath 19y ago
No, it's not a unique situation. Or rather, it's only unique in one way: if failure occurs, it occurs loudly, spectacularly, and in front of millions of people. The article itself mentioned the Denver airport, where billions of dollars still didn't prevent software errors from happening, delaying a massive project. And let's not forget the cascading failures that happened just last year because of bad weather and poor planning/routing software programs. Every time this article is posted, and it's been around since at least 2004, the one thing that strikes me profoundly is how great communications play an integral part in the entire affair. It's getting to the point where we as software programmers and engineers can't improve our software and craft until we learn the importance of communication, between ourselves and our clients. Yeah, we talk about it, but we don't really implement it. It's like saying we understand the need to watch what we eat to be healthy, and yet still eating a McDonalds value meal every meal, only with Diet Soda instead regular. We need to improve our all parts of our craft to grow. Technically is alright. The people part, the interaction part, the expectation part, goal-setting, testing, planning needs to improve tremedously. We know the problem, and we must provide the solution. If a client doesn't know what they want, ask more information. Educate the client about how the development process should work. Find out more about the industry. See what was wrong/troublesome about the old way they did their business. If information isn't making it's way from the managers to the coders/planners, find the problem and eliminate it. This is not new. I am not smart, and I know these problems. Surely, we can do better. Yes, the projects will probably end up taking longer. Yes, development will probably end up costing more. But it will be better. And the clients will be happier. p.s. YC.news needs a preview option instead of just a reply option.

Hexstream 19y ago

"The answer is, yes, the process does stifle creativity. [...] It's the process that offers a template for any creative enterprise that's looking for a method to produce consistent - and consistently improving -- quality." Come again?... "In the modern software environment, 80% of the cost of the software is spent after the software is written the first time -- they don't get it right the first time, so they spend time flogging it." I'd say most sofware cannot be gotten right the first time...

noonespecial 19y ago

Cool story, but don't run your startup like that. If you do, you will deliver perfect software, years too late, that nobody wants.

Husafan 19y ago
Not to say that every project needs to be planned like NASA, but I would like to know how many start-ups have calculated their acceptable rate of failure (in their software) prior to launch.

johnrob 19y ago

If normal software buyers were as averse to defects as NASA, all software development would probably be like this. However, normal customers tend to prefer more features and overlook defects (relatively speaking).