5 ms·

Software needs to be maintained to be useful. Proofs don't. So for many, it is not the same.

by 99_00 5y ago

atq2119 5y ago
Try reading a 19th century or earlier math paper and then tell us again that proofs don't need maintenance...
- LordDragonfang 5y ago
  Considering most software written in the 20th century no longer works in a modern environment, I feel like having to go several lifetimes back kind of proves the point.
kkjjkgjjgg 5y ago
People are not obligated to maintain their open source software, though. In fact I would expect most open source software to be unmaintained.
scotty79 5y ago
You don't need to maintain anything, you can just drop it and go do something else if you prefer. Maybe someone will pick it up. Maybe noone. If you maintain something you chose to maintain it. You have no obligation to continue maintaining it.
- rndhouse 5y ago
  Most of the value is in the maintenance. In responsiveness to changing demand. I bet all the valuable FOSS software you can think of today is maintained.
  - scotty79 5y ago
    Sure, but you could argue if it's useful because it's maintained or is it maintained because it's useful? Both. But to what degree?
    - imachine1980_ 5y ago
      i think is usefully because is maintained, most of the time when you search for project you search for trust and being continues maintained is key proxy for that, ins't the only one openbox is trustworthy(the last update is some years ago) but is trustworthy in part because you have a long record of doesn't having problems.
- 99_00 5y ago
  There is social pressure to maintain software. There is no social pressure to maintain a mathematical proof. So the analogy fails. The choice people make doesn't change the expectations of others.
kortilla 5y ago
No it doesn’t, it’s open source, not an obligation.
jasonkester 5y ago
Some software does. Most probably wouldn’t if the author could be convinced to stop when it was done. Personally, I’ve shipped several open source libraries that are still in use in lots of places despite me not having added a commit in over 10 years. They do the thing, don’t do anything else, don’t depend on other stuff, and thus don’t have a lot of surface to attack. Imagine if the log4j guys had built their thing to spit out logfiles, then stopped before adding the bit where it can execute arbitrary code. I’m sure there was a version early on that did that. And I bet that version would still work perfectly well today. It was finished. It just needed somebody to not unfinish it.
- usrusr 5y ago
  An interesting perspective. A late youth hobby of mine was being active in a coding community that wasn't open source, it was "lost source": based around a piece of freeware that was literally lost to a local storage hardware malfunction. But it had an API and a decade of stability (as in lack of API change, not lack of crashing unfortunately) and that caused an amazing amount of creativity, cooperation and so. Perhaps this is an inherent weakness of open source: it's always easier to accept one more feature PR than committing to an API that remains sufficiently stable for that feature to happen in an isolated module. I'm not saying that its always better to strive for the API instead of inclusion, it's a weakness as much as it is a strength, but chances are many situations where separation would have been better will be missed. It's a cultural bias.
- jjav 5y ago
  > Most probably wouldn’t if the author could be convinced to stop when it was done. It's nice to view software as something that can be done. I cringe when people look at activity on commits and think that if there isn't continuous activity, the project is dead. Maybe, but it might also be complete. My most successful open source project in terms of adoption (it's packaged in nearly every distro) is one that is done. I wrote it about 10 years ago and other than one small bugfix it hasn't seen any activity since. It's useful, it's complete, nothing more to do.
chongli 5y ago
Software needs to be maintained to be useful Maybe that’s the problem we need to be addressing. More and more these days I find myself frustrated using modern software when old unmaintained software had already solved the problem. We consider it a novelty when a writer like George RR Martin still uses WordStar on an old DOS machine [1]. Maybe it shouldn’t be a novelty. Maybe it should be the normal thing to do. It’s utterly at odds with a consumer culture though. People don’t seem to want well-made things that last a lifetime. They want shiny disposable crap. [1] https://youtu.be/X5REM-3nWHg https://youtu.be/X5REM-3nWHg

Software needs to be maintained to be useful. Proofs don't. So for many, it is not the same.

by 99_00 5y ago

Software needs to be maintained to be useful. Proofs don't. So for many, it is not the same.

atq2119 5y ago

Try reading a 19th century or earlier math paper and then tell us again that proofs don't need maintenance...

LordDragonfang 5y ago
Considering most software written in the 20th century no longer works in a modern environment, I feel like having to go several lifetimes back kind of proves the point.

kkjjkgjjgg 5y ago

People are not obligated to maintain their open source software, though. In fact I would expect most open source software to be unmaintained.

scotty79 5y ago

You don't need to maintain anything, you can just drop it and go do something else if you prefer. Maybe someone will pick it up. Maybe noone. If you maintain something you chose to maintain it. You have no obligation to continue maintaining it.

rndhouse 5y ago
Most of the value is in the maintenance. In responsiveness to changing demand. I bet all the valuable FOSS software you can think of today is maintained.
- scotty79 5y ago
  Sure, but you could argue if it's useful because it's maintained or is it maintained because it's useful? Both. But to what degree?
  - imachine1980_ 5y ago
    i think is usefully because is maintained, most of the time when you search for project you search for trust and being continues maintained is key proxy for that, ins't the only one openbox is trustworthy(the last update is some years ago) but is trustworthy in part because you have a long record of doesn't having problems.
99_00 5y ago
There is social pressure to maintain software. There is no social pressure to maintain a mathematical proof. So the analogy fails. The choice people make doesn't change the expectations of others.

kortilla 5y ago

No it doesn’t, it’s open source, not an obligation.

jasonkester 5y ago

Some software does. Most probably wouldn’t if the author could be convinced to stop when it was done. Personally, I’ve shipped several open source libraries that are still in use in lots of places despite me not having added a commit in over 10 years. They do the thing, don’t do anything else, don’t depend on other stuff, and thus don’t have a lot of surface to attack. Imagine if the log4j guys had built their thing to spit out logfiles, then stopped before adding the bit where it can execute arbitrary code. I’m sure there was a version early on that did that. And I bet that version would still work perfectly well today. It was finished. It just needed somebody to not unfinish it.

usrusr 5y ago
An interesting perspective. A late youth hobby of mine was being active in a coding community that wasn't open source, it was "lost source": based around a piece of freeware that was literally lost to a local storage hardware malfunction. But it had an API and a decade of stability (as in lack of API change, not lack of crashing unfortunately) and that caused an amazing amount of creativity, cooperation and so. Perhaps this is an inherent weakness of open source: it's always easier to accept one more feature PR than committing to an API that remains sufficiently stable for that feature to happen in an isolated module. I'm not saying that its always better to strive for the API instead of inclusion, it's a weakness as much as it is a strength, but chances are many situations where separation would have been better will be missed. It's a cultural bias.
jjav 5y ago
> Most probably wouldn’t if the author could be convinced to stop when it was done. It's nice to view software as something that can be done. I cringe when people look at activity on commits and think that if there isn't continuous activity, the project is dead. Maybe, but it might also be complete. My most successful open source project in terms of adoption (it's packaged in nearly every distro) is one that is done. I wrote it about 10 years ago and other than one small bugfix it hasn't seen any activity since. It's useful, it's complete, nothing more to do.

chongli 5y ago

Software needs to be maintained to be useful Maybe that’s the problem we need to be addressing. More and more these days I find myself frustrated using modern software when old unmaintained software had already solved the problem. We consider it a novelty when a writer like George RR Martin still uses WordStar on an old DOS machine [1]. Maybe it shouldn’t be a novelty. Maybe it should be the normal thing to do. It’s utterly at odds with a consumer culture though. People don’t seem to want well-made things that last a lifetime. They want shiny disposable crap. [1] https://youtu.be/X5REM-3nWHg https://youtu.be/X5REM-3nWHg