Live data from Hacker News

Homebrew 1.4.0

brew.sh

201–210 of 219 posts

Re: Homebrew 1.4.0

#201
post #166

Earlier quoted context omitted.

Opt-in is a decent enough method. But "Occasionally re-prompt a user" is a horrible idea, in my opinion. From a user perspective, that's kind of like the software forgetting the options you set. The very first time that something re-prompted me about something I had already opted-in to or out of, I would feel intense rage that it's asking me something that it had already asked me.

You would feel intense rage ? Jesus. That sounds like intense overkill . You might be quite unlike an average person if you specifically--and without fail--recall every choice you've made forever when you're in the process of setting up tools. No, it's not at all like the software forgetting the options you've set. It's the software, in the interest of providing a user with the greatest possible control over their so…

I stand by my intense rage that I would feel if some software I was using asked me a setting question that I had already set (for no reason other than the length of time since I last answered the question). Sorry, but I'm particular about my user experience. I'm not arguing anything for or against the privacy concerns, as I do not believe the fact that this is a privacy setting means it should be treated differently than other settings.

I guess I should have gone into specifics about why I don't like that idea... I can think of lots of things that let you set a setting and say "never ask me again", but those aren't quite the same. They're usually for things that get asked _every_ time you run those programs, so until you select "never ask me again" you're not actually setting the setting, you're just giving a temporary answer for that one use. Notice how this differs from what you've suggested (asking you based on an arbitrary time period). Your suggestion is a different paradigm than the already existing ones. I can see a case for having it ask you if you want to be tracked each time you run homebrew, or each time you upgrade homebrew, but doing so based on an arbitrary time-based interval (be it days, months, or years) seems odd and needlessly different than the existing accepted UI paradigms.

Oh, and I'm not sure Facebook's privacy check-ups is the best analogy. I always thought they did the check-ups when they've specifically changed something and need to give you the option to opt-out of the change, not as a friendly "do you still want your settings this way". Maybe I'm wrong about that, but that's what it always felt like to me when I was using facebook more, since I'd see a lot of articles/blog posts about new changes, and then the next time I logged in I'd get the privacy check-up.

Re: Homebrew 1.4.0

#202
post #183

Earlier quoted context omitted.

> And no, that's not a good description, because it gives me literally zero information to go on in order to find out what the workarounds were, how many formula actually needed them, and what the problems were that necessitated this. Why do you feel that you're entitled to this information? And/or that providing this information is a good use of the maintainer's unpaid volunteer time (and that it's immediately obvio…

If you don’t want to engage with the community, then don’t. But if you do it behooves you to behave in a manner that respects your users and the community. Doing otherwise is how you lose that community. It’s not “entitlement” to expect the Homebrew maintainers to be good stewards of the project and the community.

Sure it is. It's entitlement to expect literally anything for free. You and I, personally, have not done anything to deserve Homebrew, and yet we have it (well, ok, you do; I haven't run macOS in a few years, so I haven't had need of it). It's a net good, and we, personally, did not have to do anything for it to come into being. I'm not saying you should revere and prostrate yourselves before the people who gave it to you, but you should also not expect any more from them.

You should also give them the benefit of the doubt and cut them a bit of slack when they don't behave exactly as you'd think that you would. (I say "think" because it's all well and good to take the moral high ground in theory, but unless you're actually in that person's shoes, it's hard to speak in absolutes.) I've said this elsewhere, but: perhaps you hit the edges of a long drawn-out discussion that took place out of your sight, a discussion that was tiring and draining, and he just didn't feel like rehashing for you. Or maybe he simply just got out of an argument with a friend or family member, or just had a bad day.

That may not excuse his behavior in polite society, but it makes it understandable. You had a one-off bad interaction with someone. Boo hoo. It happens. Give people the benefit of the doubt, and, unless you have very strong evidence to the contrary, assume your experience was an unfortunate outlier. Homebrew is obviously a successful, robust project, with lots of contributors, so maybe, just maybe, your negative experience isn't typical, and perhaps most people who he interacts with actually come away with a positive impression. We can't all always have our best faces on, 100% of the time. Generalizing your single interaction as the norm is arrogant, ignorant, and selfish. Apply the principle of charity, and move on. If a person/community is truly toxic, then that needs to be addressed, but you seem to be assuming that your single bad experience is the norm, without taking the time to actually do any research... and feel the need to repeatedly trash-talk here for some reason.

If you met this guy in person and said "hey, I had a bad experience interacting with you online", and he didn't talk it out and apologize, then that would speak poorly for his character. But you have one tiny insignificant data point to go on, and I think it's incredibly disheartening that you've chosen to take that as representative to such a vehement degree.

Re: Homebrew 1.4.0

#203

Earlier quoted context omitted.

Not prompted but were notified.

The notification when it first came out wasn’t really noticeable for a tool that frequently has a lot of output an average user running `brew upgrade` would be used to ignoring. Prompting would have been the better choice. I didn’t notice until the change hit HN and created a bit of a shit storm. PS: Despite my feelings of disagreement on how brew handled this, I still appreciate your work.

Homebrew literally never prompts for anything so a prompt would likely break many scripts.

Re: Homebrew 1.4.0

#204
post #15

Earlier quoted context omitted.

If it is clearly documented that should be it. Maybe a blurb on the main page about it would be nice but not necessary.

I feel that if you're collecting data about me and sending it to a third party, there should at least be a clear notice during installation. It should really be opt-in, but at the very least it should be as obvious as they would make it if it were opt-in.

There is a clear notice during installation.

Re: Homebrew 1.4.0

#205
post #34

Earlier quoted context omitted.

Out of curiosity, would you be more comfortable if brew webcalled a different service? one not affiliated with Google?

Well different, i would prefer it would make only calls to services that are required for what the application does, so sites' project i guess is okay, but certainly not third parties. And then certainly not any big data third parties. I'm quite sure, there are download stats for the brew application itself to monitor usage and so on. Though i had the option to block it, using little snitch (great app, not affiliated…

We do not have the resources to be able to host something like this ourselves or we would have done so.

Re: Homebrew 1.4.0

#206

> Homebrew requires the CLT on all but the latest version of macOS (to avoid copious workarounds in formulae) Why? What are these workarounds? The linked PR didn't explain, and as someone who's still got one machine on 10.12 I'd really rather not have to keep CLT installed as well. Edit: Wow, the maintainer on that PR is a real jerk. He refused to answer a legitimate question (and locked the PR at the same time) beca…

> Wow, the maintainer on that PR is a real jerk. He refused to answer a legitimate question (and locked the PR at the same time) because he thought the original (since edited) complaint of "The error message for this patch is terrible" was "rude".

The commenting user in this case immediately opened another issue where they were rude to another maintainer too, had to be warned about their behaviour and (unlike most people who are warned) did not apologise. This issue was not locked.

My work on Homebrew over the last 9 years has been almost entirely in my free time. That's time I'm not spending with my wife, friends, dog or (now) new baby. As my Twitter and GitHub bios point out: "I block rude people". I'm not interested in giving my free time to people who cannot be polite.

The GitHub UI indicates whether someone is a first-time contributor, contributor or maintainer. You get cut proportionally more slack for having a bad day and being rude if you've contributed to Homebrew already. I try to be over-the-top positive for new contributors and I will immediately and genuinely gratefully accept apologies that are given.

> MikeMcQuaid, if you ever read this, this sort of behavior on your part is a great argument for ditching HomeBrew.

If you'd like to ditch Homebrew: please go ahead. There's things that MacPorts do better than we do and the MacPorts maintainers I know are good people. If you use Homebrew: you are the one who benefits, we do not. I want to build something that's useful but primarily maintaining Homebrew needs to be an enjoyable experience for me and others for it to be a sustainable project.

Re: Homebrew 1.4.0

#207
post #97

Earlier quoted context omitted.

MacPorts user here. Uses case are the same, but: - MacPorts does not track you by default with Google analytics. - MacPorts lives by default in a self-contained directory tree. - MacPorts tends to be more stable than Homebrew, but that’s just my personal experience. - I didn’t like the slogans Homebrew used for years, implying MacPorts wasn’t good quality.

Does MacPorts have anything like Homebrew Cask? I know it's technically a plugin/addon, but it owes its existence to the Homebrew tooling and ecosystem it's on top of. The ability to very easily install full GUI apps, or closed-source ones, is absolutely indispensable for automatically configuring Macs in many large environments, especially those (like a former one of mine) that heavily used Puppet for provisioning.…

The MacPorts system is flexible enough that you can use it to install any kind of software.

Honestly I don't know if that's popular, when using Homebrew and MacPorts I was always installing open source stuff...

Re: Homebrew 1.4.0

#208
post #187

Awesome! At some point I would like to get into Homebrew development (I'm already familiar with Ruby).

We'd love that! Here's a good starting point and feel free to contact me directly if you'd like other pointers.

Re: Homebrew 1.4.0

#209
post #18

Earlier quoted context omitted.

Any opt-out features should be disclosed to the user on install. That seems fairly straight forward to me.

Which, as someone pointed out in a separate comment thread here, Homebrew is already doing. I'm not sure what else you want from them, they seem to be following best practices here.

You are right. If I put more thought into my comment I would have urged them to prompt the user. But I rushed, and the post was invalid and senseless.

Re: Homebrew 1.4.0

#210

> Homebrew requires the CLT on all but the latest version of macOS (to avoid copious workarounds in formulae) Why? What are these workarounds? The linked PR didn't explain, and as someone who's still got one machine on 10.12 I'd really rather not have to keep CLT installed as well. Edit: Wow, the maintainer on that PR is a real jerk. He refused to answer a legitimate question (and locked the PR at the same time) beca…

> Wow, the maintainer on that PR is a real jerk. He refused to answer a legitimate question (and locked the PR at the same time) because he thought the original (since edited) complaint of "The error message for this patch is terrible" was "rude". The commenting user in this case immediately opened another issue where they were rude to another maintainer too, had to be warned about their behaviour and (unlike most pe…

I just read zenspider's other issue. It's not rude. You and the other maintainer are.

If I was on the fence before, I'm not now. Time to ditch Homebrew and stop recommending it to friends. You are poor stewards of the project and of your community.

Post reply on HN