Hacker News Reader

Comment by lewispollard | Hacker News Reader

> I've been down this road a lot with FOSS software: "I don't use it because it doesn't do what I need", "sure it does!", "no, it really doesn't", "well you don't need that anyway!"...

Just recently (I was on a Fedora box temporarily) I was trying to figure out how to get Nautilus to stop starting a search whenever I typed something, and instead highlight the nearest match in the current directory, y'know, like pretty much every other file manager in existence. Eventually I found a thread where the core Gnome contributors were saying "it sounds like you don't want this feature, what you really want is a better search function". They probably should never have enabled "reactions" on their bug tracker since it was piled 100's high with negative reactions - yet they still doubled down and said it's not something they're going to support (despite having supported it in the past, and Ubuntu even bundling patches to make it happen up until recently).

It's a frustrating trend in FOSS that contributors feel like "we're doing this for FREE don't you know, so whatever we say is lore and since you're not paying, we don't have to listen to you". Yes, I understand how amazing it is that this software is truly free and created by volunteers, no, I don't understand how that gives you dictatorial power over 100s of users who don't like the changes you've made.

View on HN

In my experience Gnome is the worst major FOSS project for this. Ever since Gnome 3 the design philosophy of all of their projects seems to be "find out what the users want and do the opposite."

Case in point: https://mail.gnome.org/archives/gnome-shell-list/2011-June/m...

"Facilitating the unrestricted use of extensions and themes by end users seems contrary to the central tenets of the GNOME 3 design."

I've heard this from a couple of power users before. But maybe GNOME does not optimize for power users?

I remember following the GNOME blog for some time and finally decided I'd never go back after the article on how gEdit was being "improved" by completely hiding the menubar, the toolbar and basically any usable controls. It was written such that "focus" and "distraction-free" computing was the ultimate goal and that UI controls were somehow "distracting".... The file open/save dialog involves multiple clicks before it is usable because obviously a path edit box is WAY too distracting...

I don't ever remember cursing my Windows 3.11 machine that the scrollbar and toolbar icons were distracting in File Manager, nor do I crash my car when the speedometer changes and wish that the distraction of the speedo and rev counter was removed.

I remember using GNOME2 and loving it.

I often wonder constitutes a "power user" - someone who wants to actually get stuff done and not fight the UI to discover things??

I wonder when iOS became the ideal - blindly fumble around whitespace to see if some text is actually a button or if there is some unfathomable control hidden that you need a magic incantation to discover (swipe left on a list control is delete - my mum still doesn't know this)?

Is a "power user" someone who has learned all the different ways of interacting with a modern user interface? You used to learn: click, double-click and scrollwheel. Now every interaction with a control is different and impossible to learn or discover simply by using it. Absymal.

What is a "power user?" Sounds like the same nonsense we hear from proprietary software vendors all the time, "That feature is for enterprise users."

Gnome more or less optimizes for minimal configurability, the theory being that most users want the UI to be as clean as possible. They do not particularly care if someone would like to be able to change some annoying behavior, because they thought that behavior made sense and they feel the cost of adding a button to disable it, or even burying a configuration bit in gconf, is far too high. Of course, they do not have the resources needed to run actual user studies the way Apple, Google, and Microsoft do, so their idea of what users want is mostly based on their own ideas about what users want (since they usually dismiss feature requests from their actual users, because those are "power users").

With apologies to Joseph Heller:

"There was only one catch and that was Catch-22, which specified that a concern for one’s own UI in the face of flaws that were real and immediate was the process of a power user. Orr was an average user and could be listened to. All he had to do was write a feature request; and as soon as he did, he would no longer be an average user and would have to be ignored."

For example, GNOME doesn't support session save out of box. I think it's an incredibly helpful feature, especially for beginners - but maybe GNOME developers simply know that, deep down, all of us yearn for a clean blank slate whenever we restart our box, and showing the apps we've been using five minutes ago will only confuse our brain.

Also it doesn't even show hibernate button on the "start menu" (or whatever it's called these days) - because apparently the ability to actually power down the box while keeping all states is too confusing to have!

Yeah seriously. The Gnome project is go-to example of non-ad-driven user-hostile FOSS software development.

I have no idea why it has become so common on here to dog on GNOME for being "user-hostile" because they targeted a different kind of user than the average HN engineer/startup person. You can use one of the "GNOME 2 revival" type desktop environments if that's your fancy.

Because the user they were hostile to were the users that had been until then the majority of Gnome's users. It may be a good business move to ditch your existing userbase in favour of a potential larger one but the userbase you've moved on from have no obligation to be happy about that

I don't expect the community to be happy, I would say I more expect a community of engineers to be realistic. The ship has unfortunately sailed long ago and it's not coming back.

That's their justification, but put a non-technical user in front of a gnome system and you will find nothing but confusion. It doesn't operate like any other OS GUI, nor is it intuitive or discoverable, nor does it have feature parity. They must be designing gnome for someone, but I don't know who.

I don't know, last time I tried stock Ubuntu I found it to be pretty similar to current macOS with the launchpad and mission control. In any case, the answer to those problems you're describing would be more and more breaking changes, not less.

Ubuntu ships with a patched version of GNOME, default experience is completely different.

> because they targeted a different kind of user than the average HN engineer/startup person.

It is not, sometimes GNOME devs even say outright that the project is for their peers and those who share their vision. They do their testing on other GNOME aficionados. In the days of Unity, Canonical did some of their user testing on regular people, those days are long gone.

They produce bad software with a bad interface based on a bad theory of how people use things and when people critique their bad software they give bad reasons, argued badly, with a bad attitude and in bad faith to boot.

It truly needs its own fractal of bad design essay save it would have to a farce in 7 parts.

To say that they targeted a different kind of user is insufficiently descriptive. 3.0 was the beginning of ticking off the users they actually had in order to target users they wish they could get with a total package that such users could never encounter in a million years because joe average wasn't installing Linux in 2011 regardless of how simple you made the interface.

9 years later the Linux user base is virtually unchanged and gnome 3 is no longer horrifying as 9 years of labor have rendered it modestly acceptable after the worst ideas were put off or mitigated.

If you spent 300k building a hovel where the roof fell in quarterly and you after much consternation and a decade of work figured out how to keep the roof mostly in place most of the time I would pretty much expect such a project to be the go to horror story among home builders in the neighborhood.

This isn't meanness its just honesty.

Agreed. And yet it's somehow ascended to being the default environment for Fedora, Ubuntu, Pop!_OS...

It gives them dictatorial power because it's their project.

I don't use Gnome because I don't like the direction they've taken it in. But they have absolutely every right to do with it as they please. And I have every right to insist it's totally not fit for my purposes and refuse to use it.

The same way I have no (current) interest in using Wayland because it'd be too much hassle to try to replicate my workflow with it. But they're free to ignore me the way I ignore them.

> dictatorial power over 100s of users

GNOME has tens of millions of users.

I do understand your sentiment. I don't like GNOME myself. But there are alternatives. If you don't like it, try using something else, contributing patches with configurable behavior or maintaining a fork.

FLOSS maintainers are a rare breed. We as a community need to find ways to not burn them out. Even if you have millions of happy users, just a couple of really toxic ones can make it feel not worth it.

> GNOME has tens of millions of users.

I was referring to the hundreds of users who "reacted" to the bugzilla thread on this issue, but I'm sure it effects many many more.

> I do understand your sentiment. I don't like GNOME myself. But there are alternatives. If you don't like it, try using something else, contributing patches with configurable behavior or maintaining a fork.

Generally I don't use GNOME, on this occasion I was working on a laptop supplied by an employer. The frustrating thing isn't that they don't support a feature that I want, it's that they strip away features that existed before, and completely ignore and talk down to users who tell them that the feature is very useful and much wanted.

Yes, obviously there's no way to tell what proportion of users think the same way.

> they strip away features that existed before, and completely ignore and talk down to users who tell them that the feature is very useful and much wanted.

I'm under the impression that GNOME strives for similar UX to macOS. This behavior sounds just like Apple to me, so maybe it's just the correct way to do it.

I've used both macOS and GNOME as daily drivers and got frustrated with both. I'd certainly get frustrated if an employer forced me to use unsuitable equipment (and some have in the past). But that's not the fault of GNOME developers.

> I don't understand how that gives you dictatorial power over 100s of users who don't like the changes you've made.

Those hundreds of user are:

1. A vocal minority, most likely. People happy about it don't go on bug trackers to support devs.

2. Not making the software, not part of the decision making process. Don't like the end result ? Fork it if you have the ability. Give constructive criticism if you want and you think the team is receptive to it (or even cares about it. Sometimes, I just don't want your genius idea).

So, yes. They're doing this for free. Don't like it? Fork it, or don't be an ass. If you can't do either of those, the third option is to fuck off.

Why are these people not part of the decison making process?

They are (predominantly) not doing this for free. They are employed by RedHat.

Then they should pay for a RHEL subscription. That includes personal support.

RHEL 8 is on Wayland by default. If you complain about any issues whatsoever, the personal support tells you to use X.Org.

> I don't understand how that gives you dictatorial power over 100s of users who don't like the changes you've made.

It doesn't - you are completely free to roll back those changes, or make different changes of your own.

Hot take: Maybe, if you're building something (or especially changing something) as a SME that many people use, and you fuck it up in a way most people don't like, you're still a jerk.

Maybe.

I think the discussion should happen.

Because I don't think "I wrote it, if you don't like it, you write it" is a 100% defense to hordes of upset users.

When you have enough users, any change will make hordes of them unhappy, just because 0.01% is already a horde.

Hot take: As a software engineer, or absolutely anything else, don't owe these people anything. I'm building software in the direction that I want to bring it. If they're unhappy, they're free to leave. But to have them come on public forums and call you a jerk, that's just comical.

There's no defense needed; nobody is on trial. These hordes of users can start paying for developers if they want certain work done.

You got me! The word defense is, in fact, only used in courts of law.

Imagine if you asked what makes Steven Spielberg dictatorial power over all our screens! Not even paying him money gives you a veto on his casting choices. What you are experiencing is the expectation of continuing to receive a benefit, useful software, interpreting no longer receiving the same quality software as a loss, and intuiting that you have a rights to avoid that loss. You don't because the benefit was unearned in the first place.

This doesn't mean you don't have a right to discuss or critique and discuss and if appropriately fork or write you own.

> I don't understand how that gives you dictatorial power over 100s of users who don't like the changes you've made.

The option always remains to change it yourself or stop using it. It's not as if it's bound to the hardware (Apple) or shipped with the machine (Windows). It's not a SaaS that may vapourise or change underneath you.

There's no way to force consensus in situations like this. The best you can do is force out those maintainers you don't like with a harassment campaign.

"The option always remains to change it yourself or stop using it."

If I changed every bit of software I dislike, I would have no time left for my actual job. Yeah, I get it, these volunteers found the time to work on this one project and therefore get more of a say about what changes will be made, but you know what? They also have to use other software, and they would not have time to work on something like Gnome if they were busy making changes to Wayland or the Kernel or whatever else.

Worse, in my experience, Gnome developers will turn down patches that do not align with their ideas about how things should work. I have had it happen to me, I have seen it happen to others -- even when we find time to change something, we cannot commit to maintaining our own private fork, and dealing with merge after merge or trying to keep up with a project that has as much activity as Gnome or Wayland. Obviously some patches need to be rejected, but in the case of Gnome I have found that rejection often comes down to their notorious "less configurability is better" attitude.

As for stopping using it, how does that work with critical packages like systemd? At some point you wind up having to ditch an entire distribution, with all its infrastructure and maintainers, just to get one use case to work. In some cases you are left with a choice between a project that has the features you need but has received no updates (including security updates) in years, or a project that does not have the features you need but is actively maintained. More often than not people will just give up on their use case, which in some cases means giving up on open source software entirely since that use case was the reason they were using open source in the first place. Some might think that is fine, users have the freedom to choose what software to run, but the fact is that fewer users means fewer contributors (including many people who make only the small but very important contribution of reporting bugs) and a less robust open source ecosystem.

All I see here is that you're not going to get everything that you want. No-one is. We are all using software that's imperfectly adapted to our personal use cases. Unfortunately, unless we're paying, no-one is under any obligation to change it for us, and we'll have to make do as best we can.

The funny thing is that a lot of users switch to open source software because the proprietary software they had been paying for was not supporting their obscure use case. You know, it is not something worth the investment because only a handful of users care or it works against the interests of some business partner or whatever.

Yeah, I get it, these volunteers found the time to work on this one project and therefore get more of a say about what changes will be made, but you know what?

They don't get "more of a say", they get ALL the say. This is the way of Free/Open Source.

As for stopping using it, how does that work with critical packages like systemd?

It works the exact same way. If you don't like it, don't use it. Simple as that.

Nothing wrong with that attitude if you don't mind large numbers of users, many of whom are very technically competent and able to make meaningful contributions, giving up on the open source ecosystem. Most users do not have time to figure out how to stop using systemd when that is what their distro uses, let alone how to do so while staying on top of updates for whatever they are using instead. Most users who stop using systemd are going to switch to OS X or Windows, and they will contribute less, maybe not at all, to open source projects. The result will be an open source ecosystem that is less able to keep pace with proprietary software and has an even harder time staying relevant.

> Yes, I understand how amazing it is that this software is truly free and created by volunteers, no, I don't understand how that gives you dictatorial power over 100s of users who don't like the changes you've made.

To be honest I don’t see how you can understand the first bit but not the second. Yes, they can make changes that you don’t like without consulting you.

They can; but they shouldn't expect a life free of complaints.

Commercial software, like most things in life, is responsive to incentives. The people who buy things get the most say in the product.

For open source where the end users aren't buying anything, and the developers are either creating product for their own gratification, or for some more nebulous corporate goal, the incentives are rather indirect. And that makes a big difference. Software, under those conditions, will frankly be worse for the majority of users.

>but they shouldn't expect a life free of complaints.

This seems to be the dominant culture, but it's actually pretty toxic that people feel entitled to whine about software created by people who owe them absolutely nothing.

Constructive feedback is useful, and appropriate if the creator indicates that they want it.

Whining and ranting is a waste of energy. It makes people feel bad and leads to no beneficial change.

I don't think Drew DeVault wrote this because he's received lots of useful constructive feedback.

"people who owe them absolutely nothing"

What makes you think people owe nothing to each other in the open source ecosystem? I maintain a package for Fedora, I have submitted a few patches to the Linux Kernel, and I regularly report bugs. Do the Wayland developers not benefit from any of the above?

Even my mother, who is not very technical, has managed to identify bugs here and there which I report on her behalf. Do the Gnome developers not benefit from that?

We cannot all be contributors to every package we use, but we are all part of an ecosystem and we all have the right to ask for changes by others in that ecosystem. You call it whining, but you know what? When developers break use cases, users have the right to complain about it. If the Wayland developers do not want to hear those complaints, they should not be involved in developing software that millions of people depend on, nor should they have tried to replace software that millions of people depend on.

Giving a project feedback does not impose any obligation on the maintainers of that project to help you – even if they find the feedback useful. In general, it's rude (and ineffective) to try to unilaterally impose non-trivial obligations on people without getting their prior agreement.

Actual contributions are another matter. If you contribute to a project, then sure, you should get a say proportional to the size of your contribution. And I wonder how many of the people who complain about Wayland have made sizeable contributions?

An important point here is that it's up to the maintainers to evaluate the value of your contributions from their point of view. If you don't respect this, you're again trying to impose obligations unilaterally. That is, offering help that may not really have been wanted in the first place and then expecting something in return.

>If the Wayland developers do not want to hear those complaints, they should not be involved in developing software that millions of people depend on, nor should they have tried to replace software that millions of people depend on.

This seems way off base to me. I hope this attitude is not pervasive because it seems like it's a huge discouragement to people who want to write useful software. (If you succeed to any degree, millions of people are automatically entitled to shout at you whenever they experience some inconvenience.)

"I wonder how many of the people who complain about Wayland have made sizeable contributions?"

How many people asked for their distro to switch to Wayland?

You seem to think that these critical packages that millions of people depend on can be treated like hobbyist projects. Yes, when your project is a hobby, you can tell people that their complaints are not your problem and that they should just fork the code if they need something different. However, distros should not make hobbyist projects core components that users are unable to avoid, nor should anyone push their hobbyist project as a replacement for a core component.

The fact is that the Wayland developers set out to replace X11. In what world are they not opening themselves to criticism from people whose use-case was supported by Xorg but not supported by Wayland? By now a decision made by the Wayland maintainers will affect the vast majority of desktop Linux users; why should they be immune from complaints? When people depend on your software you have to deal with their complaints, that is just a fact.

> This seems to be the dominant culture, but it's actually pretty toxic that people feel entitled to whine about software created by people who owe them absolutely nothing.

I don't think that's quite right. It's certainly bad manners to be rude, but the underlying complaints are one of the better - if not best - means by which the people who make the software can get feedback to ensure they're building the right thing.

Would you invest in a startup whose founders constantly complained about how their potential users whined about their product?

Aside from auxiliary revenue sources like complements and consulting, open source software has value in large proportion to how many people use it. Reputation and goodwill, if nothing else, bragging rights, job opportunities, and so on. You don't get these things from writing software than isn't used.

And you can't really expect everyone to provide constructive feedback. Often, the more emotional the rant, the more pain the user is feeling. Not all humans are dispassionate rational agents. Not every user is a professional product manager able to outline constructive incremental improvements to satisfy requirements. And it's silly to expect that. Product managers are paid well in industry to do this. Doing it for free for open source is just as much a donation.

I'll go further. If you're not willing to tolerate more and more rants as your software gets more popular, to the point that you end up not contributing or quitting, it might be better if you didn't start. Abandoned and decaying software hurts too, and it especially hurts the people who took a bet on it.

(I learned my lesson young. I created a few bits and pieces of open source, and quickly found the the effort of integrating the contributions of third parties more work than it was worth. The only way I'd do open source again is as a developer outreach / organization halo PR strategy for the benefit of a company I'm being well compensated to work for.)

I think we’re talking largely at cross purposes. It might be a good idea for maintainers of open source projects to listen to all feedback. However, nothing obliges them to do so, and you’re not entitled to be angry with them if they don’t.

If someone relies on a piece of open source software on the assumption that it will be maintained in perpetua for free, then they’re simply foolish, and they have only themselves to blame for any resulting harm.

> Not all humans are dispassionate rational agents

You seem to be forgetting that this applies just as much to the people who maintain open source projects. The endless negativity inevitably has an effect.

>> To be honest I don’t see how you can understand the first bit but not the second. Yes, they can make changes that you don’t like without consulting you.

People in charge of large FOSS projects like Gnome should not pretend it's just some hobby project they can do whatever they want with. Those projects are essentially public infrastructure and should be treated as such. It's not their personal plaything.

Yes, people who do the work have the most say, but sometimes they really use that to do stupid things with the code. I don't know how to correct that though. Commercial software is no better about this.

>People in charge of large FOSS projects like Gnome should not pretend it's just some hobby project they can do whatever they want with. Those projects are essentially public infrastructure and should be treated as such. It's not their personal plaything.

They're not pretending. I'm afraid you're laboring under a misconception. Nothing obliges the GNOME team to accede to the wishes of people who contribute nothing to the project.

"Nothing obliges the GNOME team to accede to the wishes of people who contribute nothing to the project"

Nothing, until users flee and they stop receiving bug reports, patches, donations, and so forth. Moreover, because of what Gnome is, if they drive users away it will also negatively impact other projects, because for a lot of those users leaving Gnome means ditching open source software entirely and switching to e.g. OS X.

If the users aren't willing to fork, then having them run away to osx (where, by the way, everything is also imposed on you) doesn't seem that great of a loss.

Part of the joy of open source are the freedoms it gives you, including the freedom to fork a project that doesn't meet your needs and make it so that it does. This works as it should pretty often - look at the gnome2 forks, the systemd-less distros etc. There's plenty of choice out there, and plenty of opportunity to get involved.

If you're unwilling to exercise those freedoms, then you're just whining about the stuff you're getting for free.

You are right that with OS X you have no particular say over the features. On the other hand, you do not have to deal with poorly supported hardware, and if something goes wrong there are people whose job is to help you. Most users find those things to be advantages, but many will choose open source software because of the various other freedoms they can get. There are also people who choose open source software because they have unusual requirements that proprietary systems either cannot meet or are too difficult/expensive/onerous to deal with.

Losing users is always a bad thing for the open source ecosystem. Lost users means fewer bug reports, fewer patches, fewer donations, and less support from paid developers working for various corporations (easy to forget but there are people whose job is to maintain open source projects like the Linux kernel). Fewer users means less incentive for hardware companies to provide any support, even just minimal technical documentation, needed for open source developers to write drivers (let alone contribute to that).

Forks require a significant commitment of time and effort, and really should be considered a last resort. The time commitment is much greater for prominent projects that receive a lot of activity like Gnome or Linux. Most users, even those with the necessary technical skills, do not have enough spare time to maintain a fork when things break.

Your dismissive attitude is fine for a obscure hobbyist project with a dozen users. It makes no sense for desktop Linux distros that now have many millions of users.

I understand where this thinking comes from but I still think it's kind of a simplistic attitude. People depend on the technology they use. When you convince someone to rely on something you've produced (and, crucially, continue working on), I think you do actually create a duty to be responsive to them. They depend on your work. Are we really so limited in our ethical thinking that we require an exchange of money to create obligation?

I understand that people can be aggressive, rude, and downright hostile. I'm not suggesting that such behavior be rewarded. Nor am I suggesting that every open source project has to do everything any user says. I do, however, object to the notion that there is no obligation whatsoever between the producer of a technology and the users of that technology, regardless of whether money has changed hands.

If you read the fine print there is, quite literally, no obligation at all. From the MIT license:

    THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
    IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
    FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.

You will find that most other FOSS licenses contain a similar clause.

That's a legal disclaimer designed to head off a specific commercial legal liability. If you go looking to use law as a foundation for right and wrong, you're going to be sorely disappointed.

What I quoted isn't a law, this is something that authors have to deliberately put in their licenses/contracts for the specific work in question. If you wish to have other obligations with your open source software that you believe are less wrong, it would be wise to get those in writing.

> it would be wise to get those in writing.

I agree with that, projects should clearly state their goals and priorities upfront. But relying on legalese from MIT license is not the solution.

It's not clear what else it is that you'd want. The MIT license seems to get used because it's short and to the point in stating priorities: Here's the software, do what you want with it, but don't take credit and don't come after us if it doesn't work for your use case or if it causes your CPU to explode. If your issue is with the "legalese," it wouldn't be any less meaningful if you stated it differently.

You somehow think that someone who’s done you a favor (by providing you with useful software for free) owes you more than someone who’s done nothing for you at all. That’s exactly backwards. You are obliged to them!

I struggle to express how strongly disagree. This is like saying that someone who gives away a 3D printer on freecycle is then obliged to help you fix any problems with it.

>> Nothing obliges the GNOME team to accede to the wishes of people who contribute nothing to the project.

I said nothing of the sort. They have an obligation to the project and its value as a piece of infrastructure used by many (including noncontributors).

But by saying that they have an obligation to noncontributors you do seem to be saying something at least in the neighborhood of what I suggested. I can’t understand where you think this obligation comes from. I used GNOME for a while. I never once imagined that the people who worked on it owed me a damn thing.

> People in charge of large FOSS projects like Gnome should not pretend it's just some hobby project they can do whatever they want with. Those projects are essentially public infrastructure and should be treated as such. It's not their personal plaything.

Pretty sure they are aware of this, even Gnome. Gnome's Dev Team is often citing user studies etc. and I believe they are genuine in trying to produce software for the public, rather then themselves.

...Though I believe they don't grog how disruptive changes (any change) can be.

Reply on news.ycombinator.com