GNOME Developers Hate the Mouse, Remove Middle-Click Paste
news.softpedia.com
news.softpedia.com
Some of us even hoped that maybe someday the situation would reverse itself, and GNOME could once again become a viable desktop environment. Unfortunately, it has become clear over time that this is not the case, and likely never will be.
It's disappointing to see a project that was once quite useful, yet still with a lot of potential, be destroyed so quickly and unnecessarily. And it's perfectly acceptable and understandable for us to voice our displeasure with further degradation of what GNOME once stood for.
Most recently, GNOME Terminal, ironically enough (it's about as non-GUI as you can get for an application, but it's a reasonably nice terminal emulator, and does a few things my trusty old rxvt doesn't: dynamic rescaling, dynamic color themes).
Rhythmbox comes to mind. A few others.
I really don't care where the hell my apps come from, in terms of toolkit. But if they break very-well-established Linux / Unix behaviors, it really pisses me off.
I've also seen GNOME disease utterly fuck up perfectly good applications (the Galeon Web browser comes to mind).
Worse, occasionally unaffiliated developers look at what GNOME is doing after a few too many hits on the crack pipe themselves and think it's a good idea.
GNOME's utter user hostility and imperviousness to any level of logic or reason had already gotten old a decade ago.
It can get old. Just use something else. I tried gnome3 when it happened and jumped ship like a sinking Titanic. I'm now heavily invested in kde, and version 5 isn't going to change direction into this minimalist nightmare, its still all about choice, except under the hood it gets much faster and cleaner with qt5/ gl accelerated qml.
Me, I've gone i3wm, but it's not for the faint of heart; And it still depends on lots of gnome utilities, like the network manager or the settings daemon.
It really depends on your WM. Some of the heavier desktops, yes. With the lighter ones (again, I use WindowMaker), I'm up and running in a couple of seconds (there are some apps and utilities which take a few seconds to launch).
The mouseless / tiling WMs (such as i3wm) are pretty cool as well.
No, the problem is that perfectly serviceable applications get utterly buggered by the brain-death of the underlying toolkit. By changing functionality as fundamental as middle-click mouse behavior. And it's reached the point that being a GNOME application is a liability: I'm going to look at it cross-eyed and search out an alternative if at all possible.
As another poster noted, his use of a venerable old window manager, in his case, Enlightenment DR16, means he isn't subject to an entire class of insults: "creeping featurism/bloat slowing down or interfering with overall responsiveness, or this recent distressing trend to remove well-established features and behaviors."
The simple fact is that the basics of a GUI desktop were established 40 years ago[1], and most of the changes since have involved price, performance, graphics rendering quality, and a very small number of behavioral modifications.
Technology offers diminishing returns to scale, and there's only so much benefit any graphical environment can offer by changing. And in doing so it loses its key benefit: familiarity. NO graphical environment is fully intuitive, it's all learned. Tossing out that learning on a regular basis is an anti-feature.
And so pdkl95 uses the decade-old E16, I use WindowMaker, clone of a 1989 desktop. A highly skilled hacker and repeat entrepreneur I know uses twm -- a very heavily hacked configuration, but twm all the same. It BLAZES on modern hardware. First released in 1987.
Even Apple has made relatively few modifications to its OS X interface, Aqua (and most were met with much gnashing of teeth). Microsoft's UI changes are a large part of the negative response to its recent Windows and Office offerings. Really, the improvements simply aren't worth the cost.
RHEL 7 will ship with a Gnome desktop on desktop clients, and whatever Red Hat release will be cloned into CentOS, Scientific Linux and Springfield Linux. That, I gather, in total, is rather a large number of screens. The users of those screens in various organisations will not be able to install a different DE as they (may) not have admin rights and (may) not have access to repositories with alternative DEs available. Red Hat have already suggested that they will be releasing the RHEL 7 desktop with gnome classic as the default to minimise training costs.
If anyone here knows of a RHEL type managed system with a large number of screens, it would be interesting to see how RHEL 7 is accepted over the next few years (of course, they may decide not to upgrade until RHEL6 is EoL).
I don't really have a big ax to grind with Gnome, but I don't understand why it remained relevant after the KDE/Qt licensing issue was resolved.
> The middle-click will be used to start selections, and provide text contextual menus (such as word definitions, sharing, etc.)
This seems wrong. They're removing a slightly obscure feature to replace it with a completely obscure one. I wouldn't call myself an UX expert, but this seems to violate at least two important principles:
1. Affordance/User expectations: How is anybody going to figure this out, if not by accident? Worse, users who knew the middle click will probably be irritated by this.
2. Consistency: Are only Gnome applications going to support this? It very much seems like that. So their new function for middle clicking won't even work consistently across applications.
I know I'm being an armchair critic here, but I honestly wonder why they think it's a good idea.
In Keyboard 2.1, the caps lock key was removed as you can use the shift key. The Windows key was surplus to requirements, there's a perfectly good alt key.
Fast forward to Keyboard 3.0. There is no longer a control key (hold down alt for 2 seconds!), numlock is gone (the top row has numbers), cursor keys gone (you can use a mouse to move around).
In Keyboard 3.1 they removed the shift key as stylistically capitals aren't necessary.
in keyboard 3.2 they removed the comma and semicolon. these really aren't necessary as you don't need to pause when reading.
in keyboard 3.3 there is no need for single quotes because you should not use contractions. that just causes confusion and besides the amount of debates and derision around it were getting ridiculous. justification: world peace and an end to grammar nazis.
in keyboard 35 they are seriously considering removing the enter key the exclamation mark and any characters that require umlauts also they have removed the full stop because ideas dont exist in a vacuum so why should sentences as these are constructs of a narrow mind not to mention this will help with a cleaner text rendering codebase
http://superuser.com/questions/87470/copy-on-select-paste-on...
As you said, scrolling under pointer is a huge boon, but also clicking on background windows (in most well-behaved apps) usually focuses the window without sending the click to the window (possibly triggering an unintended destructive action).
But the best part is that cmd-click allows interacting with windows without focusing them, allowing all sorts of things like translating your IM window out from under a movie player without covering the video.
I didn't know cmd+click. I can't think of a good use case in my world, but I'm pleased as punch that it exists, and that I've learned something new about OSX today.
When I tried it, this feature made me feel very unsafe, in that a small, "unrelated" movement of the mouse could lead to huge modifications to my current desktop workspace (and snap me out of concentration at the very least).
Seems a feature that leads to surprise, a bit like if "i" worked as a toggle both in command mode and insert mode in vim.
I'll note that I'm not bashing the feature, nor a personal preference. I'd like to know what advantages are there, from a knowledgeable user.
I use it to switch between windows without having to fiddle with something like alt-shift, or clicking on the target window each time I want to move. It's about speed, not safety. Once you get used to it though, I don't think you end up in the wrong window very often.
But I guess if middle-click never does anything useful for you (Windows, Mac), you might prioritise it less in your hardware design :|
Option "EmulateWheel" "on"
Option "EmulateWheelButton" "2"
Option "XAxisMapping" "6 7"
Option "YAxisMapping" "4 5"
Which is exactly the way I scroll: Hold the middle button and move the trackpoint.I would submit that this is not a problem with middle click to paste, but rather a problem with a bad hardware design for the particular mouse you are utilizing. The true solution is to replace the badly designed mouse with a better mouse.
If the mice you're trying don't have those, and it's very easy for you to inadvertently press the middle button, then that does seem like a pretty major design flaw. If you're always buying mice from the same manufacturer, perhaps investigating a different brand is in order?
If getting better-designed hardware isn't an option, then I guess the alternatives would be improving your finger control so that you don't accidentally click the button while scrolling, or disabling it in software so that it doesn't matter. Honestly though disabling it and missing out on that functionality entirely seems like the worst of those choices.
The only time I ever intentionally use middle click paste is when using the terminal where, you know, Ctrl-C and Ctrl-V are already taken. An option to disable middle click paste for everything but terminal emulators could only be a benefit, removing a rare but ever present source of confusion and mistakes.
Opera does that, too, and I love it. No need to focus the address bar after selecting text, just click and go :)
Gnome devs don't like to give the user options.
I initially assumed this was only familiarity with my current environment and dependence on the various settings, tweaks, and shortcuts that build up over the years. Jumping into any other WM was obviously going to take some time to get used to, and finding out how to re-enable stuff like my preferred keys for switching virtual-desktop should only take a weekend to figure out and port over.
All of the replacements ended up being reverted back to my working E16, usually because of either: creeping featurism/bloat slowing down or interfering with overall responsiveness, or this recent distressing trend to remove well-established features and behaviors.
What's particularly distressing is that these problems seem to be getting worse with time. I gave up attempting to find a sane configuration for GNOME years ago due to their overall bloat and confusing and rarely-consistent style. Misfeatures like this mouse button change suggests they are now embracing the Microsoft-style "Principle of Maximal Surprise" style of software design.
If you told me a decade ago that I'd be using the WM traditionally described as "unnecessary eye-candy" because it is faster than the other options, I wouldn't have believed you. If crap like changing the mouse buttons becomes wide-spread, though, I guess I'll be stuck using E16 for _another_ decade.
At least I have a (mostly) working WM to fallback on, I suppose. The feature-removal drama in Firefox is quickly becoming far more problematic...
[1]: "modern" being defined as "some time after Rasterman started teasing us with very-early E17 prototypes"
And I've been using it for 16 years. That's a lot of familiarity.
This seemed to be a trend 10 years ago or so - I haven't used GNOME on a daily basis since then - and it seemed the GNOME team has almost always had a "this is how it will be, no arguments" attitude. It's probably a perception more than anything else, but perception is reality, don't they say? Trying to make radical changes then hiding behind "Human Interface Guidelines" - even when the new behaviour is demonstrably bad - seemed to be their MO all those years ago, but my memory may be a bit hazy now.
So... we'll have a thread here of people saying "no, don't do it!" or "give me an option to toggle!" but I suspect the damage is done and this is how it will be. What decisions have been overturned by community feedback?
Then somewhere around GNOME 2.12 or so, Nautilus fell in love with the "spatial" concept. So each directory opened would open its own window and also remember the window position so it would be opened in the exact same location the next time. Leads to the user having 30+ windows open if you want to drill down to some files a few directories deep.
No way to undo the suckyness except to edit some gconf registry key.
The User Interface Designers thought it was great. Everyone else thought it was a disaster. Fortunately it was changed to something more sensible a few releases after and the spatial desktop horse is (for now) dead and buried.
Btw, while I sound really critical of GNOME, I'm not. I think it's great that they try new and wacky ideas that no one likes. Sometimes they strike gold and then progress is made benefitting all of us.
If they had a 'wacky new ideas' branch for people to test on, I suspect the whole thing would be received a lot better.
A few, AFAICS.
E.g. Gnome 3.10 is returning to a conventional multi-option power dialogue box[1], after the merging of 'suspend' & 'power off' into one menu item was roundly rejected by everyone (the extension giving them back as separate options is the 3rd most downloaded of all time on extensions.gnome.org).
Most obviously, Gnome 3.8's 'classic mode' (which is just a collection of extensions which replicate gnome 2 interaction models) wouldn't exist without community feedback (possibly combined with pressure from RH with an eye on conservative RHEL customers).
[1] https://raw.github.com/gnome-design-team/gnome-mockups/maste... (feature page: https://wiki.gnome.org/ThreePointNine/Features/SystemStatusM... )
1) Ctrl-alt-L: Give me a cursor at the end of all selected lines.
2) Select the line paces you want to paste using normal editor keyboard commands.
3) Ctrl-C
4) Go to the place in your lines where you want to paste (or click somewhere to get a new cursor and paste everything newline separated)
5) Ctrl-V
Scales to however many lines you want to do this to, with 'constant programmer time' ;-).
Actually, mine has five:
Three at the top (corresponding to the mouse buttons), generally used with the trackpoint.
Another two at the bottom (two-button mouse), used with the trackpad (which I've, of course, disabled in BIOS like any sane user would).
There should only be one clipboard.
As a multi system user with only one k/m input (windows/red hat), I love the differentiated buffer capability. It allows discrete usage of cntl x/c/v across systems and STILL allows me to manipulate c/p actions while multitasking.
Check yourself: http://tronche.com/gui/x/icccm/
I'll use these to "read in" files:
xc < filename
And then in vim (faster than toggling paste mode): r ! xp
or r ! xpc # for CLIPBOARD buffer "+P
?It's also a double pinky keystroke with a shift required (on QWERTY). Really awkward to type.
Try ':0r ! xp'
It does pretty much roll off your fingers.
Clipboard selection gets copied to a clipboard application and is accessible as long as you don't overwrite it (or even then, if you have more sophisticated clipboard app).
I have been using MATE to remain productive (gnome 3 and Unity really don't fit my workflow), so this won't really affect me, but it puts another nail in the coffin of the idea that I might transition to gnome 3 in the future.
It's ennoying enough in the "normal" case when pasting doesn't work with the middle-click, but in the past I've had several cases with X-forwarded windows where Ctrl+Insert used the paste buffer of the remote machine, and middle click used the paste buffer of the local machine. Removing the middle-click paste means I have no convenient way to copy&paste into "remote" applications anymore.
It's kind of puerile, but for whatever reason this gave me a good laugh: https://news.ycombinator.com/item?id=6282043
You mean, like Plan9's UI, where buttons and menus are "just" text? That'd be something I'd love to see come to another desktop environment...
So now in the future i have to select test in a terminal with the mouse, right click, click copy and then hit ctrl-v or right click again, whereas now i just select text and hit the middle mouse?!
Please, leave the middle-mouse-paste an option or atleast make it a gnome-extension :(
btw, I also use clicking the scroll-wheel to paste all the time
Almost every tab I close is by clicking on my scroll wheel.
Some inferior scroll wheels it seems don't click but I've bought quite cheap mouses and they all seem to have scroll-click.
At my last workplace (linux only environment), one day, somebody showed it to a second person while a bunch of other people were in the room. Two days later, everybody was using it as if they couldn't live without it. Turns out that before nobody knew the feature existed.
Double left-click and it's immediately in the buffer, because why else would you double click text? Middle click and you're done. Easy peasy, and I would hoist the black flag and start slitting throats if it disappeared.
Fortunately for me I use LXDE, where they seem to have resisted making a career out of churn. I really don't care what's in front of me except for what I'm working on, and I don't want to be aware of my DE. To me it's like adding features to a coffee cup. Once you're drinking coffee why would you care at all, as long as the handle still worked.
Well now you know several people who do. I use it all the time, and hate using windows which doesn't have the feature.
You can't use keyboard shortcuts in terminals, so that's not a replacement of middle click.
Start using it - you'll quickly find it indispensable.
Everyone with a terminal uses the feature. I've been using it every day for the past 17 years.
It's somewhat inconsistent and unreliable between applications, but it's a huge timesaver.
At first glance it seems like a backward move to me, but i'd be at least willing to give the idea a try.
GNOME 3 takes a lot of heat, they're one of the few actually pushing the envelope so it's to be expected i suppose.
A little more encouragement and a little less boring complaints "stating the obvious" would go a long way.
Many users, myself included, have railed at the removal of things we have grown accustomed to. Some time ago, however, I decided to just give up and go along for the ride. Why not? These people surely have more UX experience than I do. Perhaps I don't really know best in this regard, even though I know what has worked for me in the past. Perhaps there are better ways.
In fact, I find GNOME 3 is actually quite usable if one can manage to remove one's expectations based on prior versions. After the initial pain of "giving up" features I had become used to over the years, I found myself enjoying its simplicity.
GNOME is a singular case study in its dogged determination to rapidly cull beloved features; indeed, I can't think of any other UI project of this size that has been so aggressive and so blind to user input. There are few things the GNOME team seems to hold sacred, and likewise they seem more than willing to alienate existing users and violate user expectations in the name of progress. I find this kind of blind determination and adherence to design goals at the cost of popularity admirable in its own way, although it seems that the project is doomed to find only a small niche due to this approach (which, honestly, is kind of a shame, since I find it quite pleasant to work with).
Like everybody else who has integrated middle click paste into their workflows, I'm going to feel the pain from the removal of this feature. But I'm pretty sure I'll still be able to get my work done without it, and eventually I'll forget about it completely, and I'll be OK with whatever replaces it.
(In reply to comment #2)
> Breaking a default that has been in X forever is not desirable. In this case,
> it is a good default. Leave it alone. Bastien, please do not do this.
Disabling broken behaviour by default is a good thing. It would be a GTK+ specific XSetting, which would only be disabled by default based on your XSettings manager. If you use gnome-settings-daemon, then it would be disabled by default.
The people that paste their passwords to IRC will thank us.
---
And there you have their use case. People who use IRC. Instead of giving an option of disabling middle clicking in the IRC application (or per-application) they will, instead, be removing this commonly used feature, because "won't someone think of the users?"
Personally, I've long been planning to switch back to a tiling wm anyway.
Its a myth, a fantasy, an excuse for all things gnome and other "lets make a product for idiots" does to butcher itself into oblivion, many startups do the same mistake.
The idea "for the average user" or "for your grandmother" is a mind-trap that leads to poor choices and poor products. Dont make products for someone not interested in using it, make them for someone like a child instead of an old person like a grandmother. Just give a kid a smartphone or KDE and look how they learn and use the devices (Ive seen it with my own eyes, a kid use KDE to start his games, save pictures form the internet, view them in dolphin etc, and he was just 4-5 years at the time.), and give a smartphone or KDE to a grandmother and see basically nothing happen.
I wonder how you copy and paste with a touch device? And how easy it is.
For me at least, this feature significantly improves the UX. I use many times every day.
What I learned from my own experience with GNOME 3 (I really like it) and from observing other people's reactions, is that we tend to overestimate the importance of our old habits and are willing to put more effort to protect them than it would take to change them (e.g. spending all day figuring out how to customize the UI while it would take an hour to get used to the defaults).
Nothing has been changed. Those of us in the GNOME project were working towards a slightly different solution, which would allow middle-click paste like functionality. That solution didn't work out for the next release (3.10), so we reinstated the old behaviour.
Everyone needs to calm down. We're not going to simply remove middle-click paste.
This closely resembles the herd of servers at my disposal. I find it a highly productive desktop environment / user interface.