Everyone seems to forget why GNOME and GNOME 3 and Unity happened (2022)
liam-on-linux.dreamwidth.org
liam-on-linux.dreamwidth.org
I think why GNOME eventually became more widespread than KDE with most major distros adopting it instead of KDE in recent years (at some point RHEL was KDE-based, SUSE always was, we had Mandrake too) is another question.
[1] https://www.linuxtoday.com/developer/stallman-on-qt-the-gpl-...
Before then it seemed like KDE and Gnome were on roughly equal footing, but maybe that was just my perception.
(Now, we have Plasma 5, and Plasma 6 in a few days, which are even better :-))
The later versions of KDE4 didn't change much, they just pared down resource requirements, stabilized things and let the average computer specs catch up to them. Now KDE is considered the more "conservative" DE.
Was it really, or was that just a rumor from non-technical "C++ is bloated" Linux crowd?
I used to run KDE in the early 00's, and it run just fine, in very primitive CPUs of the time.
So our anecdotes are contrary to each other's.
KDE4 was released in 2008.
Hunted for workarounds only to find out that disabling the above services also disables some essential desktop features, and made me rethink whether the switch from GNOME 3 was worth it. Sure, GNOME 3 had horrible UX, but at least my PC still had resources to do other stuffs then. But I sure missed GNOME 2 at that time.
But it's totally with the resources <3
It's really hard to do the same with gnome-shell, without either forking it or writing a replacement.
Not that the latter is a bad thing, necessarily, gnome wants to provide a cohesive experience. That just doesn't fit in as much for environments trying to go against that vision (Pop_OS!, SteamOS, etc).
I'm speaking about the time when Gnome was created.
KDE 4 happened long after these issues of Qt being non-free (and it's not FUD, Qt was not free, it's a fact). Qt was released under the GPL in 2000, 8 years before KDE 4. At the time of KDE 4, nobody was (even) mentioning the non free Qt issue anymore. It was already old history.
The instability of the first releases of KDE 4 might have played a role in Gnome success, but Gnome was already the dominant desktop. It was Ubuntu's default desktop, and Ubuntu was the major desktop linux distro. I think it was still Gnome 2, with Unity about to be released.
KDE 4 was a big shiny promise and took some time to be released (because it was a huge revamp), that's why we wanted to have it.
It was GPL licenced. If you call that "non-free" that's downright Orwellian.
It was already GPL for almost a decade when the described events happened.
The flamewars over this at the time were epic. Had Qt already been GPL'd, the chances of Gnome taking off would have been far smaller - though there were other differences as well, the GPL issue was by far the most prominent.
It wasn't. That was the point here. You do not know your hoitory well enough to comment.
I'm all for using the definitions correctly, but people can make mistakes and deserve to be answered gently.
The only related thing I can think of is the argument that permissive free software license lets people build proprietary software on top of the licensed work, but the argument is not surprising and it's still true. I also don't remember slavery being mentioned. That would be a bit much. Now I can't manage to link this matter of fact with the discussion, KDE and GNOME both being GPL software.
Besides, I hate this sentence. It is very wrong, or true for any OS, whichever you prefer, and I'm seeing it more and more often. This kind of falsehoods doesn't need to be spread like this. Linux just doesn't get in my way and it has just worked for how I've been using my computer for years. To the contrary, the desktop environment and tools it lets me use are very efficient for me.
My issue is that I find the argument plausible and seducing, but do we know this is an actual reason or is it just guessing?
If you want to make an application that uses keyboard input, don't chose Gtk anymore.
There are licensing issues with using Qt, though. I'm fine with most of my stuff being licensed GPL but if you want to have any other license on your project, you'll probably have to use something like GTK.
What's wrong with using Qt with LGPL? You can't link your app statically, and you have to publish changes that you make to Qt's source code, if you do, but that's mostly it. That's what I'm using for my app[1].
Have you searched "Qt bindings" before posting this comment? Qt is in no way restricted to C++ and/or JS.
Python bindings for Qt are very popular and work well. Even with C++, JS and Python you end up covering most developers. But those 3 are far from the only languages you can use with Qt.
A quick look at the Qt bindings page shows bindings for other languages like Rust, Java, C#, Go and many more.
By contrast GTK bindings tend to get maintained and used by devs more frequently because they’re easier to generate and make idiomatic. I see GTK projects written in all sorts of languages but rarely see Qt ones that aren’t C++, JS, or Python.
You can quickly verify it, there is very few qt programs written in something other than c++/python. As much as I dislike C, C FFI is universal and thus gtk can actually be used from anything .
This is needlessly inflamatory. Many people simply don't like macOS style desktops.
In particular I like my menus attached to my windows because I don't run apps maximized.
When many people (like me) use that phrase, we're not referring to the UI/UX. It's instead a comment on GNOME's desire to have a controlled and universal design that is more focused on user ease and accessibility versus the componentized/customizable design of GNOME2, KDE, XFCE, etc.
My current preference is GNOME shell modded to move the top bar to the bottom, merged with a task bar, a rarely used Start menu, an often used Places menu, no Activities, virtual desktops activated by hotkeys and many other less visible things. That's much better than Windows, so thank you Microsoft, at least for me.
I personally use KDE with the taskbar on the left, as it allows for greater vertical screen real estate. Also changed the workspaces keybindings to the same ones as Gnome, which are quite better!
On a Mac the menu bar is always at the top. Because of that, the menu bar is at the logical end of mouse movement (it’s like a virtual wall, which the mouse can’t go beyond).
Because of this design, you can “slam” the mouse against that invisible barrier and it will stop where you need it (at least in the Y axis).
“Slamming” the mouse is always faster than trying to move it to an arbitrary point on the screen because there is no need to slow down the pointer on the screen. It just stops immediately.
The real correct solution is pie menus[1], which reduce both accuracy and distance requirements regardless of screen size. Blender has pie menus and it's one of the reasons its UI is so fast to use once you've learned it.
Also, the order of the window buttons for close, maximize, and minimize are incorrectly ordered, so a mistake causes you to click the opposite of what you actually want.
IMHO the bigger problem on Mac is the weird 'application centric' window system behaviour instead of the much more intuitive 'window centric' behaviour of Windows (and most Linux desktops).
For instance, "Alt-Tab" flips through applications on macOS, and it's possible for applications to be running even though they have no windows open, and in that case they just present a "naked" menu bar which I find utterly confusing. The whole thing feels like a dead-end UX experiment from the early 90's.
I'm mostly using the "3-finger-swipe" on macOS now to directly pick a window instead of tabbing though, but the main reason for that is not that it's better, it's just better than the weird macOS Alt-Tab behaviour.
PS: this sounds like I just switched from Windows to Mac, but I made that switch more than a decade ago, and it still feels "wrong" ;)
The fact that the taskbar has grouped windows by applications by default since XP/Vista hints at this scalability problem, with the taskbar quickly getting similarly unusable without the grouping.
I can never remember the (easy) shortcuts so this is helpful: https://learn.microsoft.com/da-dk/windows/powertoys/shortcut...
From PowerToys: https://github.com/microsoft/PowerToys
(Which also have the (in this context) useful FancyZones)
Back in a days when Virtual Desktops were not that handy (technically those available since win2000 if I recall correctly), I've used Blackbox for Windows and my own homemade program to quick search though all PuTTYs (written in VB6 + Winapi). Now I don't need it anymore :)
This can work really well even on MS Windows. It is unfortunately not default nor obvious to the regular user. But what comes in the box sprinkled with a PowerToy is rather good.
That is the paradigm I strongly prefer.
I have always liked the tight integration on Mac (such as printing - no driver fuss) and the UI looks really sleek. But the UX of the window/app handling does not sit well with me. Not for the lack of trying.
My first GUI was GEM Desktop which might have damaged me for life.
Edit: also if Apple cares so much about this, why do they make the close/minimize buttons so tiny?
I think daily users graduate to ⌘W and ⌘M (which are all-app standards) fairly quickly, while power users graduate to the window manager they prefer.
> …I'm still slower at this specific thing on Mac OS because I always first have to make sure the right window is focused…
You might find this useful: https://hazeover.com/
Quoting this as an anecdote - showing that to some people this doesn’t matter. In my case I guess it’s because I don’t use these menus that much.
I’m in this camp. The only windows that get maximized are things like IDEs with tons of panes which are unusable otherwise.
One perk of the menubar being this way that I’ve come to appreciate is that programs can’t try to remove it or sweep it into a hamburger menu in some misguided quest for minimalism. It’s gonna be there no matter what, which is a welcome bit of consistency.
I wish Linux had better support for this option. KDE lets you set up a global menubar but sadly it’s not supported by any programs under Wayland and only works with Qt apps under X11, with the GTK plugin required having been abandoned last I checked. There should really be an XDG standard for programs to expose their menus with, not just for customizability but as as an accessibility affordance.
Sure, but those same applications just leave a bare and non-functional menu bar at the top, which is almost more offensive.
Windows enforces the same kinda constraints on MenuBars:
https://learn.microsoft.com/en-us/windows/win32/api/winuser/...
The only difference is the Windows' version disappears if you choose not to use it. And the same with KDE and GTK.
Oh trust me they can. There's some really terrible electron-based Mac apps out there. They just have a token menubar with a few options like copy/paste and the rest is the usual hamburger crap.
https://news.ycombinator.com/item?id=32258142
https://news.ycombinator.com/item?id=32259325 (lead designer of Unity)
https://news.ycombinator.com/item?id=32259074 (lead of Ubuntu Desktop product)
https://news.ycombinator.com/item?id=32258650
https://news.ycombinator.com/item?id=32258035 (founder of GNOME)
This article really shouldn't be allowed to be posted to Hacker News anymore. Anything with this many errors, and is essentially a mickey mouse investigation, isn't worth discussing again.
This is why I like the fragmentation of Linux - really what's the problem? Those who want uniformity can go to windows and enjoy having everything dictated to them. I did #$%#$ off to XFCE and I am using it now very happily. Later MATE and Cinnamon came out to save the rest of us "luddites" and I realised then that I wasn't alone in my chagrin which was quite cheering.
Nobody explained the hit on GNOME2 by talking about "threats from Microsoft" and I can't understand how copying the Mac made Linux desktop safer legally since Apple isn't known for refraining from legal action either.
I always thought GNOME AND Windows copied the Motif style of menus in windows which I didn't like because I was weaned on GEM on the Atari and on DOS. I don't know though.
Me too! And this mindset persists today, every time I needed something I googled it and found it was already declared "wontfix" by the devs because it's something us lowly users should not need nor want.
I tentatively call bull, although it's possible that the most public facing narratives weren't the most urgent reasons I suppose.
I'm using KDE for a desktop, funnily enough customised heavily to look like Windows 9x.
I usually run something lighter like WindowMaker or IceWM though, but need Wayland for Waydroid atm, so KDE gives me a pretty nice Wayland experience out of the box.
Did I mention I run Arch?
GNOME3 shifted because of design philosophies in a good chunk of the core developers and a desire to have a more cohesive/user friendly experience a la Mac OS.
Sure, its contributors are from all over, but
* It was started by a German living in Germany while a student at a German university
* Its current-day non-profit backing org[1] - which owns the KDE trademark and represents it in legal matters - is headquartered in Berlin.
[1]: https://ev.kde.org
Even at the time, he wasn't a sole contributor. He posted to a Usenet board and gathered interest of a multinational group before commencing. Today, it's even harder to make that argument considering the minority of KWin, Plasma, etc commits come from Germans (or even Western Europeans).
> * Its current-day non-profit backing org[1] - which owns the KDE trademark and represents it in legal matters - is headquartered in Berlin.
Sure, KDE EV is German. And both the Linux Foundation and FSF are headquartered in the US...yet you would probably take umbrage with GNU/Linux being called "American" for that reason. You have to incorporate somewhere and it's probably going to be somewhere in North America or the EU.
> Red Hat didn't "create" GNOME
I didn't say they did.
> SuSe didn't "save" KDE
I never claimed they did.
(And it was SuSE, now just SUSE. When attempting to rebut, attention to detail is paramount.)
> Microsoft's legal threats were not what motivated the shift
[[Citation needed]]
> considering it came out 5 years later.
How long do you think this stuff takes?
> GNOME3 shifted because of design philosophies in a good chunk of the core developers
[[Citation needed]]
Or at least wanted. I'd like to know. I've met with the team, at their invitation and expense, and asked them personally, and I still didn't get a straight, coherent answer.
> a desire to have a more cohesive/user friendly experience a la Mac OS.
Well that failed then, didn't it?
You sure as hell heavily implied it then:
>> so Red Hat refused to bundle it or support it, and wrote its own environment instead.
Unless by "wrote its own environment" you mean "packaged GNOME instead of KDE".
> (And it was SuSE, now just SUSE. When attempting to rebut, attention to detail is paramount.)
Got me. Except I'm not the "journalist".
> How long do you think this stuff takes?
Considering development on it didn't begin until late 2008, about two years after the supposed "drastic need to shift"? 2.5-3yrs, give or take.
> Well that failed then, didn't it?
And so did the Zune. Yet it was still designed as an iPod-killer.
> [[Citation needed]]
You're making the claims in your half-researched/personally-biased article.
You cite them, ideally in the "article" itself.
You made that claim, it's on you to prove your claim. Instead, you claim something and pretend it is true unless someone provides proof about a claim that just isn't true.
> > GNOME3 shifted because of design philosophies in a good chunk of the core developers > > [[Citation needed]]
A release team member mentions that this is true. What more do you need? I was also a release team member for 10+ years. I've never heard of anything like what you claimed. I spoke to loads of people. Further, you're confusing GNOME 3 with gnome-shell. Also, GNOME 3 took ages to come out while you're pretending it was something quickly developed.
Lastly, GNOME classic. As shipped with RHEL 7. GNOME 3 with a more classical interface. Meaning, pretty close to GNOME 3.
Either way, it doesn't really matter. RHEL 6 was released in 2010 with a traditional menu for starting applications, long after the alleged cutoff date that, according to the article, should have made this impossible.
But I also thought it had to do with VFAT(?) being the default file system on Flash Drives and a patent M/S had.
IIRC, RHEL and friends refused to sign because that M/S patent was working it way through the courts.
But as others said, that article does not seem to reflect any kind of reality :)
Balmer made a bunch of really vague public threats where he refused to be drawn on which patents Linux (not Gnome, or KDE, or Gnome or KDE using distros specifically) he implied were infringing, most of which never materialised into any court cases or patent deals. Did they enter any other deals than the Novell deal, which involved MS paying Novell 11x what Novell paid MS?
GNOME was started on 15 August 1997 by Miguel de Icaza and Federico Mena as a free software project to develop a desktop environment and applications for it. It was founded in part because the K Desktop Environment, which was growing in popularity, relied on the Qt widget toolkit which used a proprietary software license until version 2.0 (June 1999). In place of Qt, GTK (GNOME Toolkit, at that time called GIMP Toolkit) was chosen as the base of GNOME. GTK is licensed under the GNU Lesser General Public License (LGPL), a free software license that allows software linking to it to use a much wider set of licenses, including proprietary software licenses. GNOME itself is licensed under the LGPL for its libraries and the GNU General Public License (GPL) for its applications.
I remember back in the day when he started Gnome with Federico, he got a place in a Time magazine 's list of influential people. I being a kid from the same country, wrote an email telling him I wanted to help building Gnome, and he replied to mee! To me he always appeared a pragmatic person.
He started off with GNOME, got interested in .net and Mono (originally as a means to integrate them into GNOME/GTK) and shifted focused on those. He built a company around that (with others), which then got bought by Microsoft due to their obvious interest in .net (and probably internal talks about the future direction of .net Core). Since then until 2022, his work was mostly on .net and its open ecosystem.
So it's neither contradictory nor counter to his roots, but is humorous when you say "the guy who created GNOME works for Microsoft". Despite the fact that he probably did some major work on bridging the two worlds together and leading to modern MS actively incubating and contributing to Open Source projects.
I'll focus on one thing: the genesis of GNOME 3.
Firstly, it felt like the right time. Many of us were looking around for ways to break out of the Windows 95 paradigm. New form-factors were both inspiring and challenging.
One of my clearest memories of early ideas that became GNOME 3 was an ongoing conversation about a "chromeless desktop". How could we get as much "chrome" (user interface elements) off the screen as possible, to make every pixel available for what the user was doing? Some of us started from "what if there were nothing on the screen at all?" and added back only what was necessary. These ideas dovetailed with the form-factor challenges posed by early tablets (including the OLPC), beefier PDA devices (Nokia 770), and netbooks (tiny, cheap laptops, like the Eee PC).
You can see how that led to GNOME 3's minimalist top bar (though we had tried even less "chrome"!) and modal approach between the app experience and the control experience.
> It's difficult to elaborate on "nope", unless someone wanted to do a point-by-point takedown
I would love that, and I think given that I tried hard to give references for my assertions, countering it merits the same.
> which it doesn't deserve.
I am saddened to hear that. I tried to be professional and impersonal about this, and such scorn is hurtful.
> GNOME 3's minimalist top bar
Which is one of my personal primary objections to the environment: this colossal waste of precious vertical pixels, which is squandered to no point.
Yes, the patent threats were real but nobody has proof that they affected the direction of desktop Linux.
And MS making a $440m payment to Novell/SuSE for Linux support to get Novell to make a $40m payment to MS for "patents", which seems a rather lopsided agreement to announce as a combined deal if the patent threats had been anything but MS trying to scare people.
Like linking to your own text at another place? https://lobste.rs/s/kz0jpg/everyone_seems_forget_why_gnome_g...
You've asserted that "everyone seems to forget" why something happened when a whole bunch of people who were there while it happened, and made it happen do not agree with your version of events. People tend to react badly to that sort of thing - especially when things land on HN and threaten to become accepted fact by a lot of folks who weren't there.
Microsoft rattled its patent sabers in the direction of Linux, true enough. But those patent allegations were believed to cover [1] the kernel, OpenOffice.org, the "Linux GUI", and an assortment of "other" things.
I won't claim to have been "there" myself - but I was "there"-adjacent: either working for one of the vendors (Novell), or writing about GNOME, doing an (admittedly small) amount of volunteering for GNOME around marketing/PR, attending GUADECs, and so forth.
My memories are, at this point, admittedly hazy -- but I cannot recall a single conversation or suggestion that the direction of GNOME 3 or Unity were prompted by, inspired by, caused by, or otherwise motivated in any way by Microsoft patents. Not officially in public, not behind closed doors while working at Novell, not in the hallways of SUSE's office in Germany, not during my tenure at Red Hat, not over a beer with any of the GNOME developers, nor any people I talked with between the early whispers of GNOME 3 planning to today.
It seems deeply odd that, had GNOME 3 been a "oh shit, Redmond's gonna sue us" action, nobody in all that time would've let spill. Odder still that those vendors wouldn't have simply pivoted to KDE, CDE, Xfce, or any of the other available environments. Nor did Red Hat jettison FVWM-95 IIRC from any and all repos they'd have been liable for -- given that Red Hat has no love for allowing risky things into RHEL or Fedora, I find it odd that they'd have puttered along with GNOME 2 under threat of patent suits until GNOME 3 was ready.
If you can cite anyone who was involved in GNOME development who says differently, who says "yes, we had to change desktop design due to patents," I'd be curious to hear that story. But when folks like Jeff chime in to say "nope," I'd put the burden on you for proof.
What I recall was that desktop folks were trying to make the desktop enticing enough to move people from Windows. They'd heard time and again how Linux was too difficult to use and went in a direction they thought was more user-friendly.
You also assert that Red Hat wrote GNOME in reaction to Qt not being GPL. Some Red Hatters were involved in the early creation of GNOME but that's not really correct either. Red Hat provided sponsorship early on, but claiming "Red Hat wrote GNOME" is over-simplification to the point of falsification. The real history is much more complicated and more interesting.
[1] https://arstechnica.com/tech-policy/2007/05/microsoft-235-pa...
A bit of a shame how that basically good idea turned into "let's just fill the screen with pointless whitespace instead".
Are you talking about the fat window titlebars? Those are too big, imo, but I have removed them from foot, Firefox, and Brave. I don’t use anything else often enough for it to bother me.
For instance look at the settings app screenshots here, everything is surrounded by at least twice as much empty space then actually needed for no obvious reason (except maybe to prevent fat-finger-syndrome on touch displays, but I use a mouse cursor, thank you very much):
https://apps.gnome.org/Settings/
Look how few settings actually fit on one page before scrolling is necessary (and since there's no scrollbar it's not even clear in some cases that more settings are hiding below). It looks less a settings application but more like an example application for UI design mistakes.
I have the same beef with modern macOS and Windows UIs though (but GNOME definitely takes the crown of most weird desktop UI).
It's some sort of modern UI designer brain virus to waste valuable screen real estate with empty pixels (but not wasting space doesn't mean to fall into the other extreme to fill everything with clutter, UI design is a fine art of balancing functionality with aesthetics - currently the pendulum has been swinging way too much into the aesthetics direction, ignoring that UIs should be functional first, and pretty only second.
> SUSE, Red Hat, Debian, Ubuntu, even Sun Solaris used GNOME 2. Everyone liked GNOME 2.
The default DE on SUSE was KDE. If you wanted GNOME you had to request it explicitly at install time. But more importantly, GNOME 2 was enormously controversial. It split the Linux and GNOME communities, it was the systemd flamewars of its day. The developers had to constantly justify themselves and received endless flames and hate mail about it. In fact it went OK because, just like with systemd, it turned out that there was a silent majority who did like the new direction of GNOME and become enthusiastic adopters, but the idea that everyone liked it is just absurd.
Also, GNOME 2 didn't have a Win95 style interface. GNOME 1 was Windows 95 inspired, GNOME 2 was clearly a reaction to macOS although it managed to establish a unique art style and personality as well.
The drivers of the new direction for GNOME 2 were Havoc Pennington and Calum Benson at Sun (who did a usability study on GNOME 1). Pennington spent a lot of time explaining to Linux hackers that more options and preferences wasn't always better. For example in this essay:
https://ometer.com/preferences.html
Both GNOME1 and KDE were basically direct mappings of the Linux CLI experience to widgets; it was common to have checkboxes in apps with labels like "Use Xrender", no further explanation provided. Pennington revolutionized the Linux DE space by arguing that the GUI should reflect what tasks people wanted to do, should try to automatically configure itself and that adding settings had a cost as well as a benefit. Some people saw this new direction as undermining the reason they liked Linux in the first place, as something endlessly tinkerable and tweakable for technical people. They didn't particularly want Linux to be approachable by non-hackers.
Depends on the time frame and which SUSE you refer to, to be clear. When Novell bought Ximian it made GNOME the default for SLED and KDE continued to be the default for openSUSE.
Thank you, Gnome designers and devs! (Also, I love PaperWM, and am looking forward to the tiling project y’all are brainstorming about.)
I know there were a few early implementations of pie menu widgets for Gnome many years ago, but as far as I know the window manager and desktop never adopted them.
Was it because of NIH? Was it because Gnome only aspired to imitating other mainstream desktops and not innovating? Was it because the Gnome developers were tricked by Alias's FUD about their illegitimate patents? Or some other reason?
Pie Menu FUD and Misconceptions
https://donhopkins.medium.com/pie-menu-fud-and-misconception...
Simon Schneegans eventually implemented his own pie menus for Gnome, but they were never adopted into any mainstream Gnome desktop user interface:
https://schneegans.github.io/news/2015/04/18/gnome-pie-061
He later made Fly-Pie:
https://www.youtube.com/watch?v=sRT3O9-H5Xs
And he's currently making cross platform pie menus with Kando:
For me these two were the only improvements in the user interface of Windows 95, but they were indeed very important. There has been no other user interface element of Windows 95 that I have considered useful and which I had not already used earlier.
The Start menu offered a much more convenient alternative for starting programs than starting them by clicking icons, like in the previous versions of Windows.
Nevertheless, I cannot understand how it was possible to patent a "Start menu" when it had only trivial content modifications in comparison with the menu used in twm, some 8 years earlier.
The second improvement was the auto-hidden taskbar, due to which the minimized windows no longer wasted screen space.
This is the only user interface element that I had not seen before using Windows 95, so perhaps it was patentable, even if it is a pretty obvious solution that could be found by anyone in the instant when it was recognized that the screen space occupied by minimized windows is a problem.
This is a general problem that affects many patents which are the obvious solutions for some problems and they do not require any kind of ingenuity for their discovery. The only reason why they are novel, is because those problems have not been encountered earlier, so nobody had thought to patent solutions for them.
So I believe that with the possible exception of the auto-hidden taskbar, it should have been easy to fight any claims of patent infringement by other graphic desktops, because for all features the differences from prior art were trivial. If the story from the parent article is true, then perhaps RedHat and al. have been scared by the legal expenses so they have chosen a workaround even if most Microsoft claims should have been invalid.
It isn’t, this “article” is just bad-intentioned bullshit.
This is ad hom, and it is bad and wrong, it is hurtful, it is mean, and it is against the spirit of HN.
You keep repeating this, but it's not up to others to prove what you claim is wrong. It's up to you to prove it is correct.
I’m commenting in multiple places because fighting disinformation is very hard (compared to producing it), I’m just trying to make it right, before people believe the falsehoods you state.
Your intentions might be good, but that doesn’t change the outcome.
What is the relevant patent number, and how did Gnome's design infringe on it, and how would removing it require a fundamental re-design of Gnome to avoid infringing?
Here's what I've found: They do have a patent for a "taskbar with start menu", number US5920316A. The abstract focuses on the taskbar, and the main claims are all linked to having a taskbar and to the taskbar being tied to a specific location (though changeable) and being permanent or autoiding in that location. Ignoring whether or not Gnome actually infringed - there are a lot of additional details to the claims - replacing the taskbar with iconified windows or using a click or hotkey to pop up a UI element instead would both be enough to avoid infringing on all the main claims of that patent. But there are so many additional details that it's likely you could easily avoid that patent while retaining a full task bar in any case.
What is clear is that it'd be ludicrous for Gnome to take as long as they did to remove the taskbar given this patent if they were being actually threatened by this and given how simple it'd be to work around.
The claims regarding the start menu are excruciatingly detailed in the general description, but very little actually appears in the actual claims. This is a common tactic to try to give the appearance of a far more comprehensive patent than you actually have, but what is covered regarding anything resembling a start menu in the actual claims of the patent is near nothing.
Given you're being directly contradicted by people more closely involved, unless you have something substantially better, these claims looks like total hogwash.
> This is the only user interface element that I had not seen before using Windows 95
Then I suggest that you need to look deeper.
I did, and I analysed it in some detail nearly a decade before that article:
https://www.theregister.com/Print/2013/06/03/thank_microsoft...
> if it is a pretty obvious solution that could be found by anyone in the instan
That is not even slightly true. Just look at a series of the Win95 beta screenshots and you will see Microsoft slowly and with considerable difficulty iterating their way to this difficult and non-obvious solution.
https://www.betaarchive.com/forum/viewtopic.php?t=3329
> for all features the differences from prior art were trivial
They really are not. I have spent days looking. I do not think you have.
https://www.betaarchive.com/wiki/
And here's a link to the first build with a fully functional start menu:
https://www.betaarchive.com/wiki/index.php?title=Category:Ch...
That's b81...there were 400+ more to follow that (800+, if you include the follow up service packs). Pretty much every person who worked on Chicago that talks about the start button/taskbar says it's one of the first things that was finalized, which the builds obviously confirm.
The "iteration" that you refer to is simply taking the combined Presentation Manager and Program Manager code and reconfiguring it into the taskbar.
For someone so "aware" of software development, you seem to expect unreasonably long timelines (5+ years for GNOME3) and unreasonably short timelines (what? 4-8 builds? to completely reconfigure a major UI component and paradigm) when they're convenient to making your wholecloth narratives work.
My own memory of ~20 years ago wasn’t great but I recall the only big legal issue regarding a windows UI was the distro called “Lindows”, aimed at making the transition to limix easier for users, which had to be renamed Linspire.
If you can refute my points, please do so, with evidence and citations.
At any rate, demanding extensive "evidence and citations" to "prove you wrong" when you haven't done the same yourself is profoundly hypocritical.
Surely as a self-professed "atheist skeptic" – according to your profile here – you must know how this works? You can't just claim all sorts of controversial things with no evidence and shout "prove me wrong!" when people object.
Also, there ARE "evidence and citations" right here in this thread. Several people who were directly involved with the projects at the time are saying "that was not a consideration, you are wrong". Are they all misremembering something so large? Unlikely. Are they all lying? Even less likely – why would they? So why do you disbelieve them?
[1]: e.g. "Microsoft was threatening to sue all the Linux vendors shipping Windows 95-like desktops" isn't at all substantiated with the provided link, which just talks about generic threats to sue. It's been a long time, but IIRC one of the criticisms was that Microsoft would claim "Linux is illegal; we'll sue you!" without ever specifically claiming what exactly was illegal. Classic Balmer-era Micrsoft FUD. AFAIK there was never any specific threat to sue "start menu like" environments, and I don't recall it ever being brought up as a concern. A good citation to prove this claim would be a discussion on the Gnome or KDE developer list discussing these concerns, or something along those lines.
If you would like to update your article with primary sources and evidence that can be verified, please do so instead.
The way Microsoft played at that time was through SCO. The lawsuit, the patent pool deals with IBM, Groklaw and Pamela Jones.
And of course you need to use it if you need modemmanager, joy.
I get they're not written to run outside of Gnome (and that a lot distros will run dbus user sessions on non-Gnome setups too), but it's nevertheless infuriating.
I don't particularly like dbus, but I also don't particularly dislike it. So it's not like I went out of my way to strip out a user session bus.
My current setup just happens not to spawn a bus for the user session because nothing else I care to run actually cares whether one is running or not, until I ran into Evince.
It mostly just annoys me the user hostility of hanging like that for a feature that the app can run just fine without - popping up a warning if a user happens to trigger something that is degraded without it would be far less user hostile.
Ah, herein lies the hangup. GNOME is fine on its own but I find its tentacles incredibly annoying when they make their way into i3. Maybe I just need to learn how to make my own Ubuntu fork and shove my setup into that.
Does macOS allow that? I haven't used it in ages. I'm asking because at some point it seemed all GNOME wanted was to follow Apple guidelines.
Yes.
Any threats from MS over anything - that MS might have been sabre-rattling I can buy - would need to have gone deeper than that.
I also wrote this nearly a decade earlier:
https://www.theregister.com/Print/2013/06/03/thank_microsoft...
I find your response incoherent. Do please explain.
1. Ubuntu didn't exist until nearly a decade after Windows '95. If there was an issue relating to Windows '95, Ubuntu
Separately,
2. Gnome was trivially modifiable in ways that mean that addressing the kind of superficial issues you bring up did not in any way require the drastic changes brought by subsequent versions of Unity.
If MS patents had been an issue relating to the specific features you listed, stripping them out or replacing them without significant other changes would have been trivial. Indeed, MS own justification for the taskbar points out their perceived downsides of pre-existing iconification (and ironically massively overeggs it, by arguing the problem was double-clicks while ignoring that they could trivially solve that by allowing activation with the same single-click they did in the taskbar)
Your description is simply wildly implausible, especially entirely absent evidence (patent numbers would be a start) and contradicting statements from people much better placed to know.
I'll also note that I remember MS saber rattling about Linux, and it was not constrained to Gnome and KDE, or even directed at them, but at Linux in general. Meanwhile, the MS-SuSE/Novell patent agreement was not the agreement an MS that believed it had an actual case would make - it involved MS pre-paying Novel $440m for Linux support coupons, and Novell paying MS $40m for the patent pledge. Coming out $400m ahead when accused of patent infringement is pretty good. I'd like someone to make those kinds of patent accusations against me. Pretty please, anyone?
I think you're getting your timelines and versions muddled up.
The article is not about GNOME. It is about GNOME 3.
GNOME 1.0 was in 1999.
GNOME 2 was 2002.
Ubuntu was 2004 and launched with GNOME 2 and nothing else.
Microsoft's campaign of legal threats began around 2006 and intensified around 2007.
Unity was released 2010 and first became default in 2011.
GNOME 3 was released 2011.
The article title literally mentions "GNOME and GNOME 3".
Denial means saying "that is not true!" Debunking means showing something isn't true.
Also, as the article of the piece, I want to point out 2 things:
1. I stand by it. Nobody's refuted it yet, including today.
2. I've rarely had so many ad-hominem attacks as for this, both last time and this time. Everyone making them, including some much bigger names than me, should feel ashamed for their poor behaviour around this.
But calling out corporate BS and FUD often gets one attacked by fanboys. It still hurts, but it is important to do.
That someone is saying No, I did it because of Y reason" , and your reply is to ask them to prove it?
Let's see, who would I choose to believe.... mhmmm
The same thing happened to many existing web sites, with "responsive web design" and "mobile-first design" starting to be heavily hyped at the time.
The usability of such web sites in desktop browsers was typically severely degraded, and the mobile and tablet experiences were also often still pretty bad, as well.
I see the Slashdot redesign of that era as being a good example of that happening. That web site became unusable with every device I had at the time.
Even today, I find web sites like that to be awkward and unpleasant to use. It was a much better user experience when separate desktop and mobile versions of a given web site were more common.
Thank you.
Not sure that it matters as the desktop that matters sits in our hands and is touch sensitive but it’s nice to know.
I am still waiting for refutation rather than mere rebuttal.
(To rebut: to say something is wrong. To refute: to prove something is wrong.)
I want to see the receipts. I want evidence.
I note that executives and project leaders from large companies and organisations subject to legal threats over intellectual property issues are required to deny infringement because to say anything else constitutes evidence: it could be construed as legal confession of the copying or other IP infringement of which they are being accused.
Or, to put it more briefly: MRDA. Mandy Rice-Davies applies.
https://en.wikipedia.org/wiki/Well_he_would,_wouldn%27t_he%3...
> I want to see the receipts. I want evidence.
Show us the receipts and evidence for your claims first. You're the one who started this without presenting a shred of evidence.
To then imply the people contradicting you are lying when they present their description of their experience is at best wildly distasteful.
If the people who were involved are directly telling you "no, you're wrong, that's not why I did it" don't constitute evidence you'd accept then what would?
> I note that executives and project leaders from large companies and organisations subject to legal threats over intellectual property issues are required to deny infringement because to say anything else constitutes evidence: it could be construed as legal confession of the copying or other IP infringement of which they are being accused.
But saying "no, patent concerns were not involved in our decision" is not what people changing their product for patent concerns would do because that would be lying, and the lie would almost certainly be caught during discovery. Instead the magic formulation is along the lines of "no comment" or other variety of "shut up." Not to mention that it can be perfectly to make a change for legal concerns while still disagreeing with those legal concerns because litigation is fucking expensive.
In any case, if the developers of GNOME 3 made the changes because they were afraid of patent threats from MS, coming out and saying "no, we didn't do it because of that" would be among the dumbest things they could do.
Please explain.
Go on, then.
Again, prove that you are right. It should be easy, no?
Even a release team member at that time who says you're incorrect only resulted you in saying "citation needed" instead of acknowledging what is said.
'Debian' is mentioned twice, in quick succession, including a suggestion that Debian 'adopted GNOME 3 and systemd' (at the same time).
Apart from, you know, NextStep which had a dock for system tray/start menu as well as my computer and network neighbourhood.
You can find screenshots of NextStep 3.3 showing this such as https://winworldpc.com/product/nextstep/3x from 1992 and it's possible they're also in earlier versions - NextStep 1.0 was released in 1989.
https://www.theregister.com/Print/2013/06/03/thank_microsoft...
NeXTStep not only had menus but tear-off dockable menus/sub-menus that are superior to any menus we have in apps today. It's clearly present in the screenshot I linked to.
The theme was that Canonical was not sponsoring as many engineers to work on GNOME and other essential projects. I think Canonical had trouble getting UI design changes pass through GNOME.
Canonical wanted to pursue a different UI with Unity. It was supposed to be the UI even for mobile devices. They spent a lot of money.
[1] http://wingolog.org/archives/2008/06/07/gnome-in-the-age-of-...
Edit: Here's one of the founders of GNOME's opinion on the article from the last time this was posted:
> This is nonsense.
Meanwhile users [me] were pleased as punch that GNOME was stable, useful and usable, just great all around and certainly not in need of major changes...
This wasn't an option for Red Hat whose contributions were all over GNOME.
If you designed a menu with just 4 options it would look something like.
[F1] - new game
[F3] - high scores
[F5] - settings
[F7] - credits
After seeing it just once reading wasn't really required. You would just hammer F1 twice for [F1] new game > [F1] easy mode You might see the second menu flash for a frame or not at all.
If the menu got more complicated you could chose to use the shift+F key for options used less regularly or you could move the option under a sub menu, or a sub sub sub sub sub menu.
The weird thing was that if you accessed a sub sub sub sub sub menu a few times it would get mapped into muscle memory. I remember hammering out something like [F3][F3][F5][F7][F1] you press the second button twice then the one below, then the bottom one below and end with the top one. If there was any hesitation the visual menu would show up and trigger the memory before even parsing the text.
It would almost instantly get me to the correct action out of a seemingly unlimited number of options.
Not counting the unusual even options, if the menus have 4 options each 4x4x4x4x4 is 1024 things!?!
I remember the awe seeing other people instantly plop things onto the screen.
Of course with such a limited computer there was not much need for such elaborate navigation but it was truly fantastic. My hands would navigate menus I didn't even know I knew.
Trying to navigate windows with the keyboard was a big disappointment by comparison. I now had to repeatedly think how to do things which on C64 was almost never the case. After finding something my mind had to noticeably get back to what I was actually doing.
[0] - https://www.nightfallcrew.com/wp-content/gallery/c64_brown_f...
>They are like twins, and switching between them is very easy.
>Unity, alienated a lot of people who only knew how to use Windows-like desktops
Unity alienate a lot of people because it was a mess. It was not nearly as poilished as macOS counterpart and at the same time failed to provide the old experience and you couldn't do much about it.
But obvsiously this is just one "IMHO" vs another one.
KDE is close enough (and menu bars have mostly disappeared), but it's just not the same.
Or their anti trust actions as listed here: https://en.wikipedia.org/wiki/United_States_v._Microsoft_Cor....
But any Microsoft/big tech apologist will tell you all that is in the past. Except then there is the $28.9 billion they owe in taxes which they refuse to pay.
We don't care, we don't have to care we are {Microsoft, Apple, Google, FaceBook, Elon Musk, Jeff Bezos, Bill Gates, AT&T, et al}
20 people in clean rooms faced with the task of figuring out a way to show all currently running applications in a small area of the screen would have come up with the same thing in 15 minutes. The fact that someone came up with it first shouldn't give him the same patent protection as say the blue LED.
the patent deals with what it is, how it interacts with running and exiting applications, how the user interacts with it, how it can be configured, in what ways it can be configured, even the start menu and a plethora of other things. it is an invention that can be patented and was granted a patent. was not even controversial. you don't need a blue led type breakthrough to patent something. if you can't see how coming up with all of the above from a blank slate and iterating on it until it fits with the system and user expectations to become the most successful UI innovation ever (remember, something that obvious was mysteriously not present on earlier systems, even when they supported multitasking) then again, I don't know what to tell you.
Windows 3.1 had minimizing programs' icons in bottom-left corner of the desktop, and so did NextStep, and similar icon space can be traced all the way back to Windows 1.0.
Extending that to all running programs and/or dedicating an always-visile area doesn't appear to be as fundamental a you want it to appear. Integrating it into the OS is obviously not meaningful part of the innovation as Windows approach would not apply elsewhere.
The essence of it is someone having an idea that all program's icons should be always visible on the screen, the rest is implementation, and ideas shouldn't be patentable. I'm also pretty sure adding a task bar to KDE/Gnome was done not by reading the patent, but rather by implementing this basic idea from scratch.
Going back to blue LEDs, it's as if the concept of a blue LED was patented itself, and nobody could do a blue LED with a different kind of semiconductors and doping.
https://www.theregister.com/Print/2013/06/03/thank_microsoft...
Just look at the development history to see how hard it was to invent:
Can I patent a steering wheel?
Its wild. Where does the line start and stop.
This is why Apple sued Digital Research -- over the GUI to its OS:
https://www.cnet.com/tech/tech-industry/apple-in-the-courtro...
https://ctrl-alt-rees.com/2023-05-13-how-apple-ruined-gem-an...
https://www.osnews.com/story/26322/apple-vs-dri-the-iotheri-...
Where does it stop? Specific precise implementations.
Menu bar at the top of the screen? Too general.
Menu bar at the top of the screen and menus are opened by left clicking on the title, after which the menu appears (termed a "pull down menu"): yes, that was Apple IP.
So, everyone else changed their designs:
* DR moved to drop-down menus instead: they appear on mouseover, without a click.
* Commodore moved to a status bar that only becomes a menu bar on a right click in Amiga OS.
* Microsoft moved to menu bars inside the window underneath the title bar (although in accordance with Fitt's law, they are much harder to hit. But MS had dedicated keystrokes for this, which the others did not.)
MS licensed this design to the Open Group for Motif, and today, most Unix GUIs use it, although now many are moving to hamburger menus.
There are also industry standards for this stuff, which software designers are intended to follow and therefore 100% will not get you sued.
https://en.wikipedia.org/wiki/IBM_Common_User_Access
This is all documented history. The fact that people today forgot this is a problem, because there are reasons things are the way they are. This stuff makes sense. It got the way it is, not by accident or coincidence or random drift, but because of multi-million dollar lawsuits.
GNOME 3 threw out a lot of good design because the designers did not know there was a difference between "stuff that is Microsoft IP" and "stuff that are industry standards".
But the GNOME designers didn't know what bits were perfectly safe to use, and which bits infringed on MS patents... so they tried to build something new and different, and took inspiration from phone designs, which were not meant to be driven by a keyboard and pointing device.
There are reasons things are how they are. I think that matters. History matters.
As Santayana said:
« Progress, far from consisting in change, depends on retentiveness. When change is absolute there remains no being to improve and no direction is set for possible improvement: and when experience is not retained, as among savages, infancy is perpetual. Those who cannot remember the past are condemned to repeat it. »
It's usually misquoted as "those who don't understand history are condemned to repeat it."
Which is why David Spencer said of Unix:
"Those who do not understand Unix are condemned to reinvent it, poorly."
It is a mystery to me why the FOSS and Unix community react so negatively and hostilely to being reminded of their history... but that is not going to stop me doing it.
I can sort of understand getting uptight if you're a huge fan of tiling, for instance.
Really?
They could have gone with any sensible approach and they choose an absolutely braindead one.
Luckily I was always a KDE user so I believe anybody who refuses KDE switch also has no footing to complain.
> So it went its own way with Unity instead: basically, a Mac OS X rip-off, only IMHO done better. Myself, I still use both Unity and macOS every day.
Why write this post to complain then?
The Docky/Plank thing has always mystified me, so can I ask: why?
An ordinary Xfce panel will do that and if you add the "docklike taskbar" plug in it integrates really well, without needing additional 3rd party tools and the extra RAM they use.
So... just: why?
Why? That's a harsh thing to say to me when all I'm trying to do is explain some history a lot of rich and influential people would prefer was forgotten.
> Why write this post to complain then?
I'm not complaining. I am trying to counter FUD that is obscuring documented history.
But you haven't actually shown any proof of your claim. You claim to "explain some history a lot of rich and influential people would prefer was forgotten", while offering no proof. While demanding proof that you aren't wrong.
Both GNOME and KDE had their highs and lows but in general they ended up OK.
Apologies if I was too harsh.
> The "why" part seems to be forgotten now: because Microsoft was threatening to sue all the Linux vendors shipping Windows 95-like desktops.
Reminder: anything other than denial means admission, and that exposes them and their employers to legal threats.
Why do you keep attacking a stranger? Do you hate me for some reason? Because it certainly feels like it.
I do not see that person attacking you. He's saying the article is bs, that is about the article, not you as a person.
> You are very hostile and keep calling me names
I haven't seen one reply from that person that is about you as a person. I see loads of replies saying the article is bs. I see loads of replies from you saying that a claim is true unless proven otherwise. Not sure, but the aggression is not one-sided.
I was. I do. Were you?
Cinnamon is based on GNOME 3. So, logically, it could not have been affected, because it did not exist at the time. It is based on something that happened following these legal threats.
Budgie came along after that.
Legal threats: 2006-2007.
GNOME 3 and Unity: 2011.
Cinnamon: 2012.
Budgie: 2014.
- the classic desktop design is simply ridiculous. Let's start with icons: they are launchers organized mostly in a free form, so they can't be many, but they also get covered anytime you open a windows. Like desktop backgrounds they are an absurdity coming from the idea that the digital desktop must resemble a physical desktop where icons are "tools laid out on the desktop";
- menus, another absurd coming from the paper era where people know labels and hierarchies of suspended folders;
- generic "bars" not different than physical desktop pens, paper clips, ... small containers of "many small stuff" to work with.
This as a simple first timid step toward a less absurd design. The the floating windows model, born clearly to resemble paper stuff you can move and laid out on a desktop, that's why tiling WM have become popular again after decades.
IT standard users at the time of Unity/first Gnome SHell (the double capital is not a typo) have already tried various launchers bar on the left or right side simply because screens from 4:3, 16:10 than 16:9 and now the start of 21:9 makes standard users in need of more vertical space, while the horizontal one start to be abundant. For similar reasons someone have made "sidebars" in apps and some have started the concept of vertically stacked tabs that remain readable even when are many and do not waste vertical screen space. Most standard users back than was already inhabited to "quick launchers" mimicking a CLI as a launcher with some extra/adopted completion, that's why menus disappear substituted by "dashes" and Ubuntu rightly and ignored by many have pushed the HUD for in-app menus as a quick search&narrow.
But most are reactionary, at first they fear to "loose something" in a search&narrow setup, while they have already ditched "web directories" for search engines a decade or more before. So the evolution is terribly slow.
Do you want to know the desktop of the future than? Well active documents, like org-mode where you can click links that can execute code (elisp:) or specific actions (ol-* link types). Something we have partially had in the past, like with PostScript GUIs or even before with Smalltalk-based workstations and at a slightly limited extent with LispM UI and Plan9 UI where you can just compose menus, buttons etc typing a name in the right place.
The modern web winning other widgets-based UIs is the proof we need docs-UIs, but modern web means docsUI in hands of someone else, org-mode is local, the future will be trying to bring back some local aspects in the modern web style and the desktop will be more and more a 2D REPL.
The commercial gimmick of Windows that copy Kde and before have tried to patent some GUIs elements to deter FLOSS devs does not really count much in the evolution timeline IMVHO.