Hacker News Reader

Comment by oshanz | Hacker News Reader

GNOME has no thumbnails in the file picker

https://jayfax.neocities.org/mediocrity/gnome-has-no-thumbna...

This is an issue for me; I don't enjoy having to work around it by finding the file in a separate file browser window.

Also an issue for me, it does nothing for the already abysmal photography/image editing workflow. I try to use Qt programs where I can or xdg-desktop-portal where it's supported but some programs like Gimp still don't support it.

This is due to a missing package. Some distributions are not well structured and removing one package may lead to these errors.

Clearly this isn't an issue that people care about enough to do anything about, otherwise someone would have gone and implemented it.

“Stop being poor” kind of an argument

Not at all? This is FOSS. If this was a critical issue that blocked people from using the software, or aggravated a developer enough, then someone would go fix it or be paid to fix it.

Developers don't have an obligation to implement every feature request. Clearly this is a nothingburger

> Clearly this is a nothingburger.

Hrm, I don't think so. It's probably more a matter of people that are accustomed to Gnome having accepted and/or habituated to the fact that it doesn't display thumbnails in the file manager. The same folks may have devised clever work arounds (opening the filesystem in a browser for example) or otherwise solved the problem in a way that it's not an impediment.

The problem with this line of thinking is that it ignores a huge cohort of potential adopters that would be stopped dead in their tracks at this issue. I think about scenarios where I might introduce an older family member to a Linux desktop for all the benefits it would bring (low cost, stable, secure, etc), and then how it would feel having to explain that they can't easily preview a photo when navigating the filesystem (like I was making excuses for a platform with a gaping hole).

In any system, it's heard to measure losses from things you don't have. A business might get that feedback from prospects that don't close, etc ("you don't have widget X, so we won't sign"), but it's harder to measure those feedback loops in the FOSS word. In short, I think not having this is a big deal, and folks that won't admit to that probably aren't considering a bunch of adoption that can't/won't happen until the issue is fixed.

> This is FOSS. If this was a critical issue that blocked people from using the software, or aggravated a developer enough

Ah, there it is. It hasn't aggravated a developer enough, so it is a non-issue. And people wonder why the Linux Desktop is unpopular.

That's the way it is when you depend on volunteers, they work on things when they feel like it. A solution would be to pay for a company-supported distro. Some were mentioned in other comments here.

> a company-supported distro

Oh, you mean like the one from Red Hat?

Yeah that's one of them. If you don't like the RHEL desktop you can choose another vendor.

For me, it's less trouble to just use other tools that suck less.

I'm annoyed but not annoyed enough to learn proper Glib C for so I can fix it. I can probably make it work if I tried, but my attempts would never get accepted back into the code base because of the terrible mess I'd turn the code into.

This issue is one of the things that a normal, non-technical user would run into if they'd ever try Linux. If the file picker can't even get feature parity with Windows XP's, you're not attracting a lot of growth.

I'm fine with waiting for someone to eventually fix it in Gnome 8 or 9, but this is a real usability issue that indicates an entire area of the framework can use some work.

I loathe this line of thinking.

Yes, "It's FOSS, go fork it" and all that malarkey.

Painting with broad strokes: MacOS and others make money on things being polished, consistent and understandable. They don't get to use their hacker blinders and say "Who needs a GUI for that" or act like Firefox and rearrange the UI every six months.

Thus, often times, the usability of OSes with a financial incentive for broad accessibility will be the most polished.

Ubuntu for a period wanted to break into the desktop OS market. They focused on polish and went so far as to create their own desktop environment (and display server)! They didn't fully succeed, but the point stands: There are many people outside the hacker community who are not going to write their own DE, who nevertheless hold the valid (and often, IMO, correct) opinion that Linux UIs blow more often than Windows/Mac.

PS: This isn't an argument about rights and obligations; I'm not saying randomGnomeDev123 has some moral obligation to do as randomUser345 asks. Just don't confuse "lack of obligation" with "being right".

And this feature is in PopOS 20.04 which iirc is a GNOME desktop, so someone did care enough about this and created a polished desktop environment with it.

Ah yes, System76, that other company that is trying to make money on hardware and creating a competitive advantage with a polished UX. Thank you for providing my case with evidence.

The people most likely to be impacted will be artists, photographers, and similar.

This is a useless viewpoint when the grief is held by mostly non-devs.

I wish things worked worked like that. That users could actually tell you exactly what would make their lives easier. Even more: that they would go ahead and implement those things themselves!

Good software design encompasses listening to and talking with users, and actually discovering what they need, where their pain points are, etc. They won't tell you explicitly.

Apart from being a software developer, i'm also a user of course. I use GNOME every day, and even though i'm affected by this, i haven't found the time, patience and dedication to actually go and try to fix it. I understand that the GNOME developers might not have the time either. But it's not like this issue does not exist, or that it's not important.

And if i, and many people who see this as an issue and could potentially fix it, don't do it, then how can we expect that non-technical users would?

I have tried to teach my mother how to copy files between devices countless times in Windows. It's not an easy task for someone who hasn't grown using a computer. I'd love to advocate for more people to use Linux, but i won't, because each little thing like not being able to see the image whumbnails when picking files can be a road-blocker for non-technical people, and add up quickly.

Or alternative: Gnome devs are OK with missing basic functionality and a broken experience for 2 decades on, otherwise one of them would have gone and implemented it.

There are several submitted patches that add this functionality on the bug report but Gnome developers have not merged them. https://bugzilla.gnome.org/show_bug.cgi?id=141154&

Maybe they were just trying to get a stable 40.0 and will add in 40.1 ?

If you look at the report time, this has been a known issue since 2004. This suggests the Devs have been ignoring or kicking the can down the road for over 15 years.

i dont even remember the last time i used a file picker to select a picture on a desktop......

i couldn't care less about this "issue", which supposedly shows that desktop linux is a "joke". he's not really convincing.

(it would be a nice to have feature, yes. but his reaction is way overblown)

> This is why Free desktop operating systems are a joke and haven’t been popularly adopted [...] GtkFileChooser remains broken

This is a dead meme. Boring and annoying specially to the Gnome developers. I have used Gnome for many years now. Never have this been an issue. If you look at the issue tracker[1], they are open to pull requests to add the thumbnail, we just need one person who cares about this feature!

There are many other issues with Gnome that actually is their decision and it is a problem. This specific one is a GTK issue. But for example their decision on window decoration in wayland is just wrong and makes the whole environment look weird[2].

[1] https://gitlab.gnome.org/GNOME/gtk/-/issues/233

[2] https://gitlab.gnome.org/GNOME/mutter/-/issues/217

That seems excessively dismissive; it's hardly a dead meme while the issue remains open, boring only to the GNOME developers who have decided that little things like what users want is irrelevant, annoying only because it's been a decade with zero attempt at a solution. It's nice that you never use this feature, but unhelpful to others who do. And yes, it's terribly generous of them to be willing to consider merging a fix if someone else does the work.

But, you know, that's how it works: the problem doesn't get fixed unless someone both a) cares, and b) can do something about it. Why would a developer who isn't getting paid do work they don't care about? If you're interested, maybe you can finally be the person who cares and does the work.

Giving something away for free doesn't mean that no one can criticize it. It's a severe shortcoming, it's been there for over a decade, and no obligated to fix it but that doesn't mean anybody is obligated to stop complaining about it either.

It may be time to look elsewhere if the complaints are unanswered for more than a decade; they're probably being directed towards the wrong people. That's an obligation you may want to hold yourself to, if you value your own time and don't want to spend another decade repeating yourself to those who won't listen.

It kind of depends: I don't use GNOME, but AFAIK both Chrom(ium) and Firefox use GTK and so use the broken file dialog regardless. "Go use something else" is a valid suggestion, but only when you can use something else.

Entirely true, which is why I never said they couldn't complain — just that it won't be fixed until there's a developer who cares enough.

> we just need one person who cares about this feature

Uh no. There have been at least 3 separate attempts at implementing this. One of them actually lives as a fork on github, and it works. None of them have been merged. The gtk devs don't care.

Various patches have existed for at least a decade but every time they're proposed some issue is found and they're not merged so the goal posts shift from 'just write your own patch' to 'just become a Gnome maintainer' with no guarantees that you'll get the assistance you need to implement the feature or whether it'll even be merged, and fair enough, the Gnome team really don't owe anybody anything, but in that case getting annoyed that the community keeps referencing a 17 year old issue is just a natural part of the exchange.

Maintainer refuses to merge bad work, shame on them?

If it was really bad work, then of course not.

The problem is that poor quality isn't the reason why they refuse. While not related to the file picker, I personally offered to port an important part of GTK3 to GTK4 (status icon support) and one of the maintainers told me flat out that he would not merge my work because "it was no longer in line with the direction GNOME is heading".

I suspect the real reason behind their refusal is purely subjective, unfortunately.

That's not subjective. When they merge code like that and then you disappear, they're now stuck budgeting time to maintain that API for years. They can't be expected to do that work when no maintainers care enough and no GNOME apps are even going to use that API (Status icons were removed from the GNOME HIG years ago)

If you don't want to go through the trouble of becoming a committed long-term GTK maintainer yourself just to get this in, why not maintain that particular feature as a separate library?

Why are people so dismissive about this? It's a significant usability problem among many, many others that gnome has.

It should be annoying that the issue keeps coming up.

It's not a 'meme' that gnome lacks an extremely basic feature that is incredibly common in other GUIs.

Reply on news.ycombinator.com