5 ms·

Thank you for linking this article. > Nobody wants to look stupid or get told to RTFM, and so, they choose silence. That's a great insight. I wonder how many

Thank you for linking this article. > Nobody wants to look stupid or get told to RTFM, and so, they choose silence. That's a great insight. I wonder how many users are lost to "setup / entry point problems" in projects. These problems may not even be "bugs" but issues such as difficulties in configuration, too many steps, something not entirely straight forward. One thing that comes to mind is analytics to see conversion funnels, that may work for websites but not so much for tools or libraries. It's a great motivation for testing, and perhaps user studies. I would be willing to pay a few bucks for someone to try my project and tell me what issues they ran into. Unfiltered.

pjc50 7y ago
Analytics inside libraries is the sort of thing that if HN finds out about it you will get death threats. Even if you want to help not everyone wants you looking over their shoulder.
- toxicFork 7y ago
  Yes, that is unfortunately the case... I guess that is why some companies have open beta versions where in the eula you agree if you use this version you will be analysed. See Microsoft Windows insiders edition, Jetbrains, etc.
- sneak 7y ago
  They’re fine, just as long as they are opt-in.
  - acdha 7y ago
    That’s your opinion but it’s not universally shared and the people who take more extreme positions also, in my experience, tend to be the most vocal about any perceived abuse. All you need are one or two of them to spoil you on open source for awhile.
    - sneak 7y ago
      It is not an extreme position to think that software spying on its users without their consent is rude and inappropriate. It's fine to collect telemetry, just as long as the user agrees to permit it. Otherwise, your software is malware: it benefits you, at the expense of the user's privacy.
      - acdha 7y ago
        The problem is how you define those terms: “spying” has a strong value judgement and opinions vary widely, as does the threshold for what people believe constitutes informed consent.
        sneak 7y ago
        “They didn’t say no” isn’t consent.
        acdha 7y ago
        I think you should stop swinging at that straw man long enough to read what I actually wrote.
        unionpivo 7y ago
        that's a feature not a bug. If you go down and try pin those terms, you will get "language layers" that will try to go as close to the line as possible and when they cross it argue the definitions. Be as open and forthcoming as possible.
        acdha 7y ago
        I think you meant “language lawyer” or, given the context, just “lawyer” but it’s not really what I was talking about: my point was just that projects can define what they collect, what they do with it, and how to control it, and they will still get angry rants from strangers on the internet. Analytics is a really useful tool but it is also a lightning rod for criticism and conspiracy theories.
ryukafalz 7y ago
I've been that person before. If I install something and it crashes or segfaults before I can even use it, the impression I get is "I guess this isn't ready for use right now" - more often than not I just move on at that point. It's different of course if it's a large well-known project, but there are enough smaller projects that really aren't ready that it often doesn't feel worth it to report things like that.

pjc50 7y ago

Analytics inside libraries is the sort of thing that if HN finds out about it you will get death threats. Even if you want to help not everyone wants you looking over their shoulder.

toxicFork 7y ago
Yes, that is unfortunately the case... I guess that is why some companies have open beta versions where in the eula you agree if you use this version you will be analysed. See Microsoft Windows insiders edition, Jetbrains, etc.
sneak 7y ago
They’re fine, just as long as they are opt-in.
- acdha 7y ago
  That’s your opinion but it’s not universally shared and the people who take more extreme positions also, in my experience, tend to be the most vocal about any perceived abuse. All you need are one or two of them to spoil you on open source for awhile.
  - sneak 7y ago
    It is not an extreme position to think that software spying on its users without their consent is rude and inappropriate. It's fine to collect telemetry, just as long as the user agrees to permit it. Otherwise, your software is malware: it benefits you, at the expense of the user's privacy.
    - acdha 7y ago
      The problem is how you define those terms: “spying” has a strong value judgement and opinions vary widely, as does the threshold for what people believe constitutes informed consent.
      - sneak 7y ago
        “They didn’t say no” isn’t consent.
        acdha 7y ago
        I think you should stop swinging at that straw man long enough to read what I actually wrote.
      - unionpivo 7y ago
        that's a feature not a bug. If you go down and try pin those terms, you will get "language layers" that will try to go as close to the line as possible and when they cross it argue the definitions. Be as open and forthcoming as possible.
        acdha 7y ago
        I think you meant “language lawyer” or, given the context, just “lawyer” but it’s not really what I was talking about: my point was just that projects can define what they collect, what they do with it, and how to control it, and they will still get angry rants from strangers on the internet. Analytics is a really useful tool but it is also a lightning rod for criticism and conspiracy theories.

ryukafalz 7y ago

I've been that person before. If I install something and it crashes or segfaults before I can even use it, the impression I get is "I guess this isn't ready for use right now" - more often than not I just move on at that point. It's different of course if it's a large well-known project, but there are enough smaller projects that really aren't ready that it often doesn't feel worth it to report things like that.