4 ms·

Beware of Open Source Software Zombies

17 points by mhausenblas 11y ago

unimpressive 11y ago
This notion that software must have a community is really tiresome and reeks of entitlement. Most people who create OSS projects are doing so in their spare time as a charitable endeavor and don't want to support it forever. Open sourcing end-of-life products so they can possibly have a second life as open source software, serve as code examples to the community and be preserved as part of our collective history is undoubtedly a good thing. Spreading the idea that only supported community-driven OSS is worthy of existence is basically asking anybody who ever open sources anything to commit to involving themselves with it indefinitely. If that's the barrier to entry, I suspect you'll see a lot less open source software going forward into the future. ;)
- mhausenblas 11y ago
  Agreed for individuals. For (commercial) organizations I think it is valid. One good citizen in this respect is btw the ASF: http://attic.apache.org/ http://attic.apache.org/ … clearly flags the dead ;)
  - dalke 11y ago
    The essay puts them into the same group - "established commercial entity .. startup that pivots ... or if you as an individual trying to increase your chances to land your dream job through creating an interesting OSS portfolio". There are also people like me, the owner and sole employee of a commercial organization, and employees of a company which allows 10% time to, among other thing, work on free software or whatever else interests them. It says "next time you embark on creating an OSS project, consider the following steps to avoid a zombie". Let me tell you my story. In 2003 I wrote PyRSS2Gen, a Python-based RSS2 generator. It took 2-3 days. I wrote documentation and put it on my web site. It's now used in a lot of packages. I get a support question about once or twice a year. At this point the code might as well have been written by a stranger as I don't recall the details of how it works. I don't want to put more work into it. Which means it's a "zombie" according to this essay. Or it's a dead weight keeping me from doing other things. The essay implies that there's an obligation that I and any other distributor of free code must maintain things in perpetuity. Like unimpressive, I do not like the implicit sense of entitlement. I have also written plenty of other software which are dead ends, but it took experience under deployment to figure out why it was dead end. Am I supposed to somehow withdraw the software from the public eye once I no longer think it's useful? There's still other software which I developed for a client as free software, which I made available, but which I won't develop further unless I get funded for it. I haven't touched some of those packages for 5+ years. Do I work on it for free, and not do the the consulting work that pays my bills?
- ternaryoperator 11y ago
  Agreed. I do wish that on projects--both commercial and individual--when the owners have decided not to continue contributing, they'd put a note to that effect. A lot of times it's hard to tell if, say, there's been no dev activity for six months, whether the project is just temporarily quiescent or whether the owner has decided to move on to other things.
  - dalke 11y ago
    How does it affect your decision making? If there has been no development for 6 months, have you tried emailing them with a consulting or support offer? Because I've put things off for years sometimes due to lack of funding.
oneJob 11y ago
I suspect I'm not alone is thinking, that it's not a problem specific to OSS. The codebase at work,,, for 25% of it, basically if it worked the first time, zombie.
chris_wot 11y ago
Does Apache OpenOffice fall under this category? Seems very much a candidate - that project is moribund.
jeffreyrogers 11y ago
Some software doesn't need a community. Unless it is a framework a lot of software can be relatively stable. For example, NaCl[1], arguably the best high performance cryptographic library, hasn't been updated since 2011. [1]: http://nacl.cr.yp.to/index.html http://nacl.cr.yp.to/index.html
bshimmin 11y ago
An alternative tldr: spend your precious free time giving away your stuff for limited or (most likely) zero personal gain, and then make sure to spend even more time shepherding it afterwards, also for no gain (but possibly considerable frustration, no doubt from entitled jackasses like the author of this article). Honestly, is this guy really suggesting that when you open-source some software, you should also create "an exciting video walkthrough" and a Slack channel? Seriously?
- mhausenblas 11y ago
  Both you and others made valid points concerning individuals. I will update the post to make the scope clear. And just to clarify this here: I certainly don't feel entitled to anything concerning OSS, no need for an ad hominem ;)
  - unimpressive 11y ago
    >I certainly don't feel entitled to anything concerning OSS, no need for an ad hominem ;) No ad hominem here, it's entitlement. And even if you personally have a more nuanced understanding, what you've written will encourage people who do not. http://byfat.xxx/what-is-opensource-and-why-do-i-feel-so-guilty http://byfat.xxx/what-is-opensource-and-why-do-i-feel-so-gui... (Skip to nineteen minutes in.) https://github.com/wpeterson/emoji/issues/23 https://github.com/wpeterson/emoji/issues/23 For larger context of these, their relevant HN threads: https://news.ycombinator.com/item?id=10093332 https://news.ycombinator.com/item?id=10093332 https://news.ycombinator.com/item?id=8712035 https://news.ycombinator.com/item?id=8712035
gayprogrammer 11y ago
I don't see any problem with OSS code without a community. I've always thought of the freely available code online like a library full of books: nobody is reading 99% of them, but their availability is important for someone who wants to learn or find new ideas. What if only the most popular code repos existed?
klibertp 11y ago
> So, the question is: how can we avoid these zombies, both from a user and also from a ‘producer’ perspective? No, the question we should ask first is why being a zombie is bad (especially compared to not existing at all, ie. not being open sourced). I don't see it answered in the article. I don't see it asked in the article. It's simply assumed. There are many different kinds of zombies. Most of them are essentially immortal unless shot in the head. They may move rather slowly, but they have great strength. They may become blind, but they often get a great sense of hearing. They tend to rot if left alone but can last very, very long if preserved properly. There are instances of zombies who continue to be conscious after zombification and who continue to do what they did before (think Mr. Slant from T. Pratchett books). They are unlikely to change their ways, but sometimes it's a good thing. There are zombies who retain their emotions (see Sankarea manga). And so on. It's equally true for open source projects. Even if abandoned (for some definition of abandonment) for the same amount of time, two different projects may become two very different kinds of zombies. LiveScript, for example, is a zombie if you look at its mailing list, but is one of the most productive (for me, of course) compile-to-JS languages out there. StumpWM (a Common Lisp WM) is also nearly dead, yet it's still the best WM if you want a WM that's hackable. On the other hand, there are projects which we'd really like to see dead, yet they refuse to die or even zombify. Projects which suck a massive amount of resources into their life support systems without any visible improvements. > and if you find one in the wild: run! Or you could just take a nice pump-action shotgun (ie. ability to read and write the code) with you and have a good time. In short: this article is shallow. It reads like a generic advice written only to write something, without any deeper insight. And the author seems to know very little about zombies...
- dalke 11y ago
  Hear, hear. In the Xanth series, love rejuvenates a zombie.
daviross 11y ago
This sort of expectation ends up being a mental block for me sometimes as far as putting something I've done out there. If I put it out there, someone might use it and expect updates. How many projects could I maintain at once? Is this one really worth that much trouble?
- pnt12 11y ago
  My suggestion: just release it. If someone uses your program or gives it a new life under a fork, that's great. If no one uses it, there's no harm done. If some asshole complains about support, either kindly ask him about a paid offer, remind him of his right to fork the project or ignore him.
ternaryoperator 11y ago
This reads more like a description of what to do if you want to build an active community. We're long past the point where putting OSS code out there implies a project that the owner wants to build a community around.
CmonDev 11y ago
A lot of .NET indie games are relying on such a zombie - Farseer. A Box2d clone so full of features and complexity, nobody but the owner can support it. And the owner checked out.

17 points by mhausenblas 11y ago

This notion that software must have a community is really tiresome and reeks of entitlement. Most people who create OSS projects are doing so in their spare time as a charitable endeavor and don't want to support it forever. Open sourcing end-of-life products so they can possibly have a second life as open source software, serve as code examples to the community and be preserved as part of our collective history is undoubtedly a good thing. Spreading the idea that only supported community-driven OSS is worthy of existence is basically asking anybody who ever open sources anything to commit to involving themselves with it indefinitely. If that's the barrier to entry, I suspect you'll see a lot less open source software going forward into the future. ;)

mhausenblas 11y ago
Agreed for individuals. For (commercial) organizations I think it is valid. One good citizen in this respect is btw the ASF: http://attic.apache.org/ http://attic.apache.org/ … clearly flags the dead ;)
- dalke 11y ago
  The essay puts them into the same group - "established commercial entity .. startup that pivots ... or if you as an individual trying to increase your chances to land your dream job through creating an interesting OSS portfolio". There are also people like me, the owner and sole employee of a commercial organization, and employees of a company which allows 10% time to, among other thing, work on free software or whatever else interests them. It says "next time you embark on creating an OSS project, consider the following steps to avoid a zombie". Let me tell you my story. In 2003 I wrote PyRSS2Gen, a Python-based RSS2 generator. It took 2-3 days. I wrote documentation and put it on my web site. It's now used in a lot of packages. I get a support question about once or twice a year. At this point the code might as well have been written by a stranger as I don't recall the details of how it works. I don't want to put more work into it. Which means it's a "zombie" according to this essay. Or it's a dead weight keeping me from doing other things. The essay implies that there's an obligation that I and any other distributor of free code must maintain things in perpetuity. Like unimpressive, I do not like the implicit sense of entitlement. I have also written plenty of other software which are dead ends, but it took experience under deployment to figure out why it was dead end. Am I supposed to somehow withdraw the software from the public eye once I no longer think it's useful? There's still other software which I developed for a client as free software, which I made available, but which I won't develop further unless I get funded for it. I haven't touched some of those packages for 5+ years. Do I work on it for free, and not do the the consulting work that pays my bills?
ternaryoperator 11y ago
Agreed. I do wish that on projects--both commercial and individual--when the owners have decided not to continue contributing, they'd put a note to that effect. A lot of times it's hard to tell if, say, there's been no dev activity for six months, whether the project is just temporarily quiescent or whether the owner has decided to move on to other things.
- dalke 11y ago
  How does it affect your decision making? If there has been no development for 6 months, have you tried emailing them with a consulting or support offer? Because I've put things off for years sometimes due to lack of funding.

oneJob 11y ago

I suspect I'm not alone is thinking, that it's not a problem specific to OSS. The codebase at work,,, for 25% of it, basically if it worked the first time, zombie.

chris_wot 11y ago

Does Apache OpenOffice fall under this category? Seems very much a candidate - that project is moribund.

jeffreyrogers 11y ago

Some software doesn't need a community. Unless it is a framework a lot of software can be relatively stable. For example, NaCl[1], arguably the best high performance cryptographic library, hasn't been updated since 2011. [1]: http://nacl.cr.yp.to/index.html http://nacl.cr.yp.to/index.html

bshimmin 11y ago

An alternative tldr: spend your precious free time giving away your stuff for limited or (most likely) zero personal gain, and then make sure to spend even more time shepherding it afterwards, also for no gain (but possibly considerable frustration, no doubt from entitled jackasses like the author of this article). Honestly, is this guy really suggesting that when you open-source some software, you should also create "an exciting video walkthrough" and a Slack channel? Seriously?

mhausenblas 11y ago
Both you and others made valid points concerning individuals. I will update the post to make the scope clear. And just to clarify this here: I certainly don't feel entitled to anything concerning OSS, no need for an ad hominem ;)
- unimpressive 11y ago
  >I certainly don't feel entitled to anything concerning OSS, no need for an ad hominem ;) No ad hominem here, it's entitlement. And even if you personally have a more nuanced understanding, what you've written will encourage people who do not. http://byfat.xxx/what-is-opensource-and-why-do-i-feel-so-guilty http://byfat.xxx/what-is-opensource-and-why-do-i-feel-so-gui... (Skip to nineteen minutes in.) https://github.com/wpeterson/emoji/issues/23 https://github.com/wpeterson/emoji/issues/23 For larger context of these, their relevant HN threads: https://news.ycombinator.com/item?id=10093332 https://news.ycombinator.com/item?id=10093332 https://news.ycombinator.com/item?id=8712035 https://news.ycombinator.com/item?id=8712035

gayprogrammer 11y ago

I don't see any problem with OSS code without a community. I've always thought of the freely available code online like a library full of books: nobody is reading 99% of them, but their availability is important for someone who wants to learn or find new ideas. What if only the most popular code repos existed?

klibertp 11y ago

> So, the question is: how can we avoid these zombies, both from a user and also from a ‘producer’ perspective? No, the question we should ask first is why being a zombie is bad (especially compared to not existing at all, ie. not being open sourced). I don't see it answered in the article. I don't see it asked in the article. It's simply assumed. There are many different kinds of zombies. Most of them are essentially immortal unless shot in the head. They may move rather slowly, but they have great strength. They may become blind, but they often get a great sense of hearing. They tend to rot if left alone but can last very, very long if preserved properly. There are instances of zombies who continue to be conscious after zombification and who continue to do what they did before (think Mr. Slant from T. Pratchett books). They are unlikely to change their ways, but sometimes it's a good thing. There are zombies who retain their emotions (see Sankarea manga). And so on. It's equally true for open source projects. Even if abandoned (for some definition of abandonment) for the same amount of time, two different projects may become two very different kinds of zombies. LiveScript, for example, is a zombie if you look at its mailing list, but is one of the most productive (for me, of course) compile-to-JS languages out there. StumpWM (a Common Lisp WM) is also nearly dead, yet it's still the best WM if you want a WM that's hackable. On the other hand, there are projects which we'd really like to see dead, yet they refuse to die or even zombify. Projects which suck a massive amount of resources into their life support systems without any visible improvements. > and if you find one in the wild: run! Or you could just take a nice pump-action shotgun (ie. ability to read and write the code) with you and have a good time. In short: this article is shallow. It reads like a generic advice written only to write something, without any deeper insight. And the author seems to know very little about zombies...

dalke 11y ago
Hear, hear. In the Xanth series, love rejuvenates a zombie.

daviross 11y ago

This sort of expectation ends up being a mental block for me sometimes as far as putting something I've done out there. If I put it out there, someone might use it and expect updates. How many projects could I maintain at once? Is this one really worth that much trouble?

pnt12 11y ago
My suggestion: just release it. If someone uses your program or gives it a new life under a fork, that's great. If no one uses it, there's no harm done. If some asshole complains about support, either kindly ask him about a paid offer, remind him of his right to fork the project or ignore him.

ternaryoperator 11y ago

This reads more like a description of what to do if you want to build an active community. We're long past the point where putting OSS code out there implies a project that the owner wants to build a community around.

CmonDev 11y ago

A lot of .NET indie games are relying on such a zombie - Farseer. A Box2d clone so full of features and complexity, nobody but the owner can support it. And the owner checked out.