LWN: Comments on "Catanzaro: GNOME Core Apps Update"

Catanzaro: GNOME Core Apps Update

nix — Tue, 04 Jul 2023 15:22:09 +0000

It's not a road to a DoS -- vte has handled infinite scrollback properly for many years. The scrollback is encrypted, compressed and written to an unlinked file in /tmp. If /tmp fills up, it's truncated. No DoS (though running out of space in /tmp for a second or two might be noticeable).

Catanzaro: GNOME Core Apps Update

Wol — Thu, 18 May 2023 13:48:47 +0000

:-)

If you're a contractor, yes. If you're a software house, it's in your own interest to eat the cost. And - true for me for pretty much my entire career - if you're in-house, you should be telling your boss "fix it now, or it'll be more expensive later". Personally, I've rarely had problems getting that message across. YMMV.

Cheers,
Wol

Catanzaro: GNOME Core Apps Update

sam.thursfield — Thu, 18 May 2023 13:41:46 +0000

Every time there's an LWN article about GNOME a few people make the effort to comment on said article how they don't like GNOME, loosely linking their comment into whatever the article is about.

For balance, I will add that it works excellently for me & have been happily using GNOME 3 everywhere since soon after it was released. I'm glad people are spending time to make sure the core apps are as good as they can be.

Catanzaro: GNOME Core Apps Update

jem — Thu, 18 May 2023 13:14:48 +0000

>If the customer asks for a solution to a problem, STEP BACK. All too often it's a special case (like MS's label problem). Solve the GENERAL problem - and by trying to identify the general problem that will help you understand the customer's problem! - and then solving the customer's problem will probably be much easier as well.

If you are working for a paying customer you often don't have time to solve the general problem. The customer wants the fix right now, and you certainly can't bill the customer for the two week work that the general solution takes to implement, no matter how much prettier and more useful it might be in the long run.

Catanzaro: GNOME Core Apps Update

Wol — Thu, 18 May 2023 11:03:07 +0000

I'm fairly sure it's genkernel. Building the initramfs etc.

I can't remember what happened, but it build grub.cfg, and messed it up. As did SUSE ... :-(

I've always stored multiple distros in a single /boot, with grub giving me a choice. Which means I need to hand-edit grub.cfg because the automated tools can't cope with several different / partitions in a single cfg file. And then my root partitions are in an lvm volume over raid and SUSE especially really fscked that up ...

Cheers,
Wol

Catanzaro: GNOME Core Apps Update

dullfire — Thu, 18 May 2023 07:30:39 +0000

out of curiosity: which Gentoo packages are you talking about? It sounds like it's kernel related, but beyond installing the kernel source packages, I have custom tooling to handle kernel builds & deployments

Catanzaro: GNOME Core Apps Update

mb — Tue, 16 May 2023 07:13:25 +0000

Often these options historically come from completely different use cases and only appear to fit your use case by accident. But the devs cannot adapt them to the new use case because of xkcd 1172.

Catanzaro: GNOME Core Apps Update

IanKelling — Tue, 16 May 2023 02:56:22 +0000

I think his overall point is good, but the acronym is bad and the disparaging of teenagers is just as bad. When I was younger, I was inclined to deep dive into fixing random bugs in free software far more than I am now and I think that is true of many young people.

Catanzaro: GNOME Core Apps Update

Wol — Mon, 15 May 2023 22:53:37 +0000

It's really not that hard to step back, and look at a truth table, isn't it?

If the customer asks for a solution to a problem, STEP BACK. All too often it's a special case (like MS's label problem). Solve the GENERAL problem - and by trying to identify the general problem that will help you understand the customer's problem! - and then solving the customer's problem will probably be much easier as well.

That's what gets me, again and again. People can't see beyond the end of their noses, and by fixing the special case they screw up the general case.

Perfect example - I've got a weird setup where if I let my distro update /boot, it basically trashes it and renders the system unbootable. Gentoo has a flag - "don't automount /boot". Which is, basically, UTTERLY USELESS.

If I don't have a \boot partition, it does nothing (of course). If I do have a \boot partition the flag defaults to false and trashes my setup. If I set the flag to true, the obvious thing to do is update the \boot directory. Except it doesn't. It just demands I set the flag to false and aborts. WHAT'S THE POINT OF HAVING IT!?

And then the workarounds I get given (a) don't actually do what I want, and (b) don't work. I was advised to set the "don't install" flag ??? but I DO want it to install! I jiust want it to install in the \boot directory, not the boot partition. Even worse, setting the "don't install" flag changes the directory search order and causes the build to crash with "modules not found"???

Seriously, how much effort does it take to step back, and think "does what I'm doing make sense?". It appears a LOT of people can't. Even when a bug report points it out to them. (Okay, my fruity language didn't help. But I always work on the basis "if the person I'm dealing with is frustrated by things OUT OF THEIR CONTROL, then give them slack"!

My big bugbear is that communication is supposed to be two-way. If you don't listen, and try to understand, then communication is impossible. (Yes I now some people might level that accusation at me ... :-) And an important part of listending is trying to understand other peoples' frustration.

Cheers,
Wol

Catanzaro: GNOME Core Apps Update

NYKevin — Mon, 15 May 2023 20:21:53 +0000

I don't exactly hate jwz, but I really wish he would stop using a mental dysfunction as an epithet, and it makes me sad that others are now repeating that epithet. This is of course a real problem, but it should have a less obnoxious name.

Catanzaro: GNOME Core Apps Update

MrWim — Mon, 15 May 2023 20:18:39 +0000

> And I know I bang on about it, but this is also the "wizard" mentality - "that's what they asked for, let's give it to them". The *correct* approach (imho) is to ask "what is the underlying problem", and give THAT to them.

That comes with risks as well of course. Unless you have a really good understanding of the problem you risk "Even better than that!": https://www.youtube.com/watch?v=11fCIGcCa9c . Understanding other peoples' problems is **hard** and requires empathy and humility.

Catanzaro: GNOME Core Apps Update

NYKevin — Mon, 15 May 2023 20:10:49 +0000

> And now it has exhausted its potential to the point that just maintaining a simple stupid official image viewer is deemed to be hard.

But if this all started life as a GIMP-based project, can't they just fork GIMP, rip out all of the editing functionality and, uh, call it a day? That doesn't sound hard to me. Is there some property of the GIMP codebase that makes it harder than it sounds?

Catanzaro: GNOME Core Apps Update

jafd — Mon, 15 May 2023 13:17:57 +0000

I had glitchy tabs in Console and it really liked to segfault on me randomly, so that was the end of it. I mean, some advanced stuff may be lacking, but this is a terminal emulator.

Catanzaro: GNOME Core Apps Update

nim-nim — Mon, 15 May 2023 09:01:45 +0000

> This is what always happens when you relying on volunteers a lot. Volunteers want to have "fun" and you can't blame them because no one pays them.

Volunteers also want some fame which means lots of other useful code to build upon and shine. Gnome is reaching the point where its platform has been completely diluted by years of half assed unfinished fad-of-the-day rewrites and its value proposition for new developpers is approaching zero.

Remember that if all started by reusing the GIMP UI toolkit and was supposed to fix and generalize it to make it easier to write many useful FLOSS apps. And now it has exhausted its potential to the point that just maintaining a simple stupid official image viewer is deemed to be hard.

Catanzaro: GNOME Core Apps Update

marcH — Sun, 14 May 2023 17:27:21 +0000

This is exactly why my Linux terminal is by choice... a Mac. I use each OS for what it's good at: Linux for the command line and macOS for the GUI and the "enterprise" cr*p. Combining the two and using the best of each has surprisingly fewer challenges that you would expect (Emacs' TRAMP and ssh saved my life). Another big difference is of course having to _pay_ for the Mac: not a coincidence.

A lot of "enterprise" users have no choice but to follow the same approach except Windows replaces macOS.

Nothing will change until some company starts making and _selling_ a coherent and polished "workstation" experience based on Linux. It may or may not be free as in free speech but it cannot be free as in free beer otherwise it will keep falling into the CADT trap https://www.jwz.org/doc/cadt.html.

Linux-based products have been successfully marketed for mobile (Android) and "consumer" (ChromeOS) but nothing has really been done for "workstations" yet. Maybe the market is just too small. Or maybe the constant GNOME churn is bearable. Or maybe using a Mac or Windows as a Linux terminal is good enough. Or rather: a deadly combination of all the above will keep making sure the year of the Linux desktop never comes.

Catanzaro: GNOME Core Apps Update

marcH — Sun, 14 May 2023 15:54:29 +0000

Because alternatives have even fewer volunteers and resources?

Catanzaro: GNOME Core Apps Update

marcH — Sun, 14 May 2023 15:46:25 +0000

> Things would look a lot better if instead of rewriting everything all the time in the latest fad* they would focus resources on maintain things

This is what always happens when you relying on volunteers a lot. Volunteers want to have "fun" and you can't blame them because no one pays them.

Note you can of course be a professional during the day and a volunteer at night, these are not mutually exclusive.

Sometimes corporations also fund "former volunteers" without the former having much power over the later - because there's no one else available to do the work.

When a corporation really wants to focus a project and finds the resources to do so, it just forks. Apple for instance has frequently done that.

Catanzaro: GNOME Core Apps Update

swilmet — Sun, 14 May 2023 10:11:45 +0000

Yes there was a C# wave, with for example Banshee (music player). There is still MonoDevelop / Xamarin Studio (IDE).

Catanzaro: GNOME Core Apps Update

Vipketsh — Sun, 14 May 2023 08:46:02 +0000

> I just wish they had the resources [...]

This sounds great and all because it is always true to at least some extent, but somewhere in there is also an expectation for projects to use the resources they have available *efficiently*.

From the outside the GNOME project appears to love wasting as much resources as it can making "they have limited resources therefore my experience will suck" a little hard to swallow. Things would look a lot better if instead of rewriting everything all the time in the latest fad* they would focus resources on maintain things. Perhaps they could put off making up new "Human Interface Guidelines" all the time to justify sweeping UI changes and, instead, maintain what they have. Perhaps it would be a little more resource efficient if, instead of breaking API of core libraries (e.g. GTK) forcing porting efforts all over the place, they would put in some effort to keep the API stable.

*: Initially GNOME was written in C. Sometime they made up Vala and decided to (re)write things in that. Then they decided to (re)write some things in javascript. Now it seems like they want to (re)write things in Rust. In the early days there was also some implication that they want to (re)write things in C#, though not sure if anything ever was in C#. Sorry, but if you have resources for this many rewrites then, I think, you should have resources to maintain decent apps.

Catanzaro: GNOME Core Apps Update

raven667 — Sun, 14 May 2023 03:05:29 +0000

Reading other comments and thinking about it more, if GNOME was a product team and not community driven then I think they should pick a specific kind of user to cater to, like creative professionals such as audio/video artists and sysadmins/software developers, and make the core experience about that. Those users have work to do, and need an appliance with a browser, terminal, text editor, screenshot and image viewer, audio and video player, file organizer, support for video conferencing and screen sharing, maybe file sync with Google Drive or MS OneDrive and run a handful of professional apps such as vscode, audacity, blender or autocad.

I would think about moving the apps which are not supporting that use case out into independent projects and flatpaks, so those who volunteer on them and use them can still do so. There could be a SIG for mobile phone/tablet (Maps, Calls) or home user desktop (Photos, Music) or whatever for contributors who are more interested in their niche.

I look at a project like vscode for inspiration, out of the box it is just a plain text editor, but you can learn the keybindings, command pallette and settings as much or as little as you want, extending it through plugins to rival emacs in functionality, without breaking the simplicity of the core experience of editing text files. I like a desktop the same way, where the core experience is very simple, advanced features exist and you can learn them as much or as little as you want, and you can extend it with simple plugins without breaking the core simplicity. I tend to find many "powerful" tools to have complex and surprising behavior that makes them less useful, so I really appreciate the GNOME focus on simplicity, I just wish they had the resources to keep more than just the core polished and high quality.

Catanzaro: GNOME Core Apps Update

dilinger — Sat, 13 May 2023 17:30:30 +0000

Both Loupe and Snapshot are written in Rust. I don't know how featureful or useful the new applications are, but that's a trend that I'm happy to see.

I say that as I sit here watching gnome-shell slowly leak memory* over time, regularly restarting numerous Vala-based apps like Geary and Dino because of similar memory leaks and other issues, and not even bothering to use stuff like Totem because it's built using a bunch of C code that really wishes it had first-class objects and is more than happy to dangerously cast types everywhere (spitting out tons of warnings/errors in the process) to approximate such a thing.

If gnome can standardize on a memory-safe language, starting with its apps, that would be a very good thing. I like C, I've worked on plenty of hardware drivers using it, but it's really not appropriate for a desktop environment.

* lol, embedded javascript interpreter

Catanzaro: GNOME Core Apps Update

ballombe — Sat, 13 May 2023 16:34:40 +0000

> But then I wonder why is GNOME the default desktop almost everywhere ?

This is a mystery to me. GNOME change so much that it cannot be for backward compatibility.

Catanzaro: GNOME Core Apps Update

5fdb1f — Sat, 13 May 2023 09:31:28 +0000

So only MIME type associations were removed, but totem is still able to play audio files if opened manually. It makes more sense now.

Catanzaro: GNOME Core Apps Update

Wol — Sat, 13 May 2023 06:22:24 +0000

> > If you want to serve a certain group and to keep all members of said group happy needs, say, 100 options you need those 100 options in the UI.

> Sure, and that's how you end up with MS Word where every possible feature any big customer needed over the last 40 years has been implemented and is in the UI. I'm sure that has sold a lot but I'm not sure it's made the UI better

> > I think, the mentality should be all about how to organise those options into logical hierarchies.

> And that explains the ribbon menu concept, organize the options by frequency of use and logical hierarchy

And I know I bang on about it, but this is also the "wizard" mentality - "that's what they asked for, let's give it to them". The *correct* approach (imho) is to ask "what is the underlying problem", and give THAT to them.

Which is why WordPerfect came up with the idea of physical and logical pages, and solved the problem of booklets and labels (and probably several other things) with ONE feature, whereas Word has a bunch of - inflexible - one trick ponies. Ever tried to delete a *single* label from a sheet in Word? If you can edit the underlying database table, it's easy. If there isn't an underlying database of some sort, you're stuffed ...

Cheers,
Wol

Catanzaro: GNOME Core Apps Update

raven667 — Sat, 13 May 2023 03:30:53 +0000

> If you want to serve a certain group and to keep all members of said group happy needs, say, 100 options you need those 100 options in the UI.

Sure, and that's how you end up with MS Word where every possible feature any big customer needed over the last 40 years has been implemented and is in the UI. I'm sure that has sold a lot but I'm not sure it's made the UI better

> I think, the mentality should be all about how to organise those options into logical hierarchies.

And that explains the ribbon menu concept, organize the options by frequency of use and logical hierarchy

In any event with GNOME I think part of the conversation is that is glossed over is that its tech stack is probably the most widely deployed of the traditional desktops (meaning not ChromeOS) but they have around what, 100 active full time 40hr/week contributors, if even that much and are trying to maintain the whole suite, so the number of people familiar with and contributing to any given part is probably small. I would guess that Windows or MacOS have 10 people full time for every 1 GNOME has covering roughly the same number of features or code So they are trying to track or lead the state of the art but without the literally billions of dollars worth of man-hours that a peer environment like MacOS or Windows has.

I think they literally just cannot implement every obviously necessary feature, and features they do have fall out for lack of maintenance, so instead they work on simplifying the core and prioritizing only the most critical, broadest, paid or fun work.

Other desktops do exist, and some are great, but I think they have even fewer resources and correspondingly smaller scope, or rely on the GNOME tech ecosystem underneath. They might be better in some ways but so far not in enough
ways for distros to consider switching the default experience, although that could happen if distro maintainers believe the benefits outweigh the cost.

Catanzaro: GNOME Core Apps Update

Wol — Fri, 12 May 2023 20:24:23 +0000

> "don't bother with GNOME if you don't want this app installed"

Let's just hope they don't get the Android disease then ...

Cheers,
Wol

Catanzaro: GNOME Core Apps Update

st — Fri, 12 May 2023 18:20:40 +0000

My first instinct is that default installed apps should be simple in functionality and feature discovery, with advanced or specialist users free to choose more powerful apps themselves if their needs require it. The problem with that, though, is that apps that the project does not regard as core to its intended user experience will never receive first class citizen treatment. If this Console thing lacks functionality that advanced or specialist users want in their graphical terminal emulator, but then let's say at some point it benefits from some integration thing that GNOME devs can't be bothered to also backport into Terminal because Terminal isn't a core app and volunteers shouldn't be told what to work on etc .. well anyway you see where I'm going with that.

Catanzaro: GNOME Core Apps Update

paulj — Fri, 12 May 2023 15:42:19 +0000

Well, either way, the app that was fulfilling the function of playing any audio or "music" file a user clicked on had that function stripped, and the app that can play the Music/ collection can't play specific audio files.

Anyway, I'm a happy MATE user. Precisely cause of stuff like that post GNOME-2 (the UI changes in GNOME 2.0 had objective HCI studies behind them, thanks to Sun - GNOME 3 went in another direction).

Catanzaro: GNOME Core Apps Update

rahulsundaram — Fri, 12 May 2023 15:20:53 +0000

> The core app for music dropped support for audio files, is correct - the core app was totem/Videos - not Music:

Yes but the OP's wording is - We don't know what to do about our music application (Music) because in version 3.38 we "removed support for opening audio files".

This reads to me as Music dropping audio support in 3.38.

Catanzaro: GNOME Core Apps Update

mcatanzaro — Fri, 12 May 2023 14:49:26 +0000

The merge request (including a very short explanation) is: https://gitlab.gnome.org/GNOME/totem/-/merge_requests/158

Catanzaro: GNOME Core Apps Update

walex — Fri, 12 May 2023 14:34:10 +0000

My guesses:

* The browser argument can be nowadays applied to *anything*, it can be the runtime environment for whatever, given WASM etc.; see all the apps that use Electron, and not everybody is happy with that.

* Developers are volunteers and they decide themselves what kind of app they want to develop. Perhaps they want to maintain the possibility that Maps will have an offline version too eventually.

* GNOME exists in cellphone, tablet and embedded versions too, some with rather modest hardware capacity.

Catanzaro: GNOME Core Apps Update

farnz — Fri, 12 May 2023 13:54:24 +0000

I think part of the problem is that GNOME's Core Apps are a mix of "don't bother with GNOME if you don't want this app installed" and "the GNOME developers commit to keeping this working as best they can for GNOME users". Everyone agrees that the former category is "core apps", and covers things like the file manager; but we don't (AFAICT) have a good name for the latter category, of apps which are developed in exactly the same way as the required (core) apps, but which aren't a vital part of the desktop environment.

Catanzaro: GNOME Core Apps Update

paulj — Fri, 12 May 2023 13:43:32 +0000

The core app for music dropped support for audio files, is correct - the core app was totem/Videos - not Music:

"Previously, our default music player was Videos, which was really weird, and now we have none"

Catanzaro: GNOME Core Apps Update

eduperez — Fri, 12 May 2023 13:36:09 +0000

IMHO "core apps" should be those apps that every user _must_ install in order to use the desktop, the message should be "do no bother installing desktop X, unless you also install app Y"; and "Maps" would surely not be on my list.

Catanzaro: GNOME Core Apps Update

rahulsundaram — Fri, 12 May 2023 13:29:54 +0000

> - We don't know what to do about our music application (Music) because in version 3.38 we "removed support for opening audio files".

You are confusing things here. Totem dropped audio support apparently, not Music.

> - Of course this list of glaring success encourage us for future app replacements : "Loupe is a new image viewer app developed by Chris and Sophie to replace Image Viewer (eog)

Among all this mess, Loupe is only one that seems decent.

Catanzaro: GNOME Core Apps Update

intelfx — Fri, 12 May 2023 12:51:54 +0000

That’s hilarious and tragic at the same time.

Catanzaro: GNOME Core Apps Update

docontra — Fri, 12 May 2023 12:10:36 +0000

But audio containers (esp. those in common use) are by and large video containers without the video ( https://developer.mozilla.org/en-US/docs/Web/Media/Formats/Containers ). And even then, interpreting a container is the job of an external library (gstreamer/ffmpeg/whachamacallit).

Now, I agree that listening to an audio file from a video player is not exactly the very best UX ever, but it beats the "Listening to local audio files in my web browser" UX and handily beats the "I have no (GUI) program to listen to an audio file at all" UX

Catanzaro: GNOME Core Apps Update

Vipketsh — Fri, 12 May 2023 10:56:53 +0000

> What you don't want is to clutter the menus

I think that attitude is very wrong and is actually contradictory. If you want to serve a certain group and to keep all members of said group happy needs, say, 100 options you need those 100 options in the UI. No-one is going to be happy about not having the thing work the way that would be comfortable because "there is no clutter". Instead, I think, the mentality should be all about how to organise those options into logical hierarchies.

Perhaps GNOME and others on the "simplify" train need to understand that there is exactly one vendor (Apple) who can change anything anywhere and have their users cry out in joy about all the "cool new things". Unfortunately for GNOME their users are not like this.

Maybe I just got the wrong end of the stick and the only group that GNOME wishes to keep happy are its own developers. But then I wonder why is GNOME the default desktop almost everywhere ?

Catanzaro: GNOME Core Apps Update

zdzichu — Fri, 12 May 2023 09:05:21 +0000

Or what's the point of application like Maps? It has no offline mode, so it is not better than ordinary Web Browser. A web browser is even better, as it gives access to different map providers.
I see the developers put a lot of work into Maps (I read the maps blogposts on GNOME Planet for almost a decade), but there was no rationale given why it was selected to be a core app.

Catanzaro: GNOME Core Apps Update

rsidd — Fri, 12 May 2023 08:57:36 +0000

Is there a point of designating certain apps "core apps"? I would imagine most users have their preferred apps, maybe vlc for media, firefox for web, whatever. GNOME terminal may be nice but so is xfce4-terminal, and some may prefer foot or alacritty or whatever, all of them are usable in any desktop environment. The article makes it sound like console developers lost interest because Fedora Workstation rejected it. That suggests the motivation for developing the app is wrong. Same with GNOME Web (née Epiphany), what's the point of it? If the point is to have an alternative lightweight browser, good, that's an itch worth scratching. If the point is only to have a gnome "core app", or worse, if it's because the GNOME committee said a "core app" for browsing is needed, that's not very motivating.