|
|
Log in / Subscribe / Register

Thomas: That's it, we're quitting

[Posted March 8, 2011 by corbet]

Matthew Paul Thomas posts on the "Canonical Design" site about Canonical's intent to eliminate the "quit" button in applications. "The 'Quit' command in applications today is a relic from the days when the original Macintosh had no hard disk and couldn't multitask. Modern applications have made this command increasingly annoying. Fortunately, though, modern PCs have also made it increasingly unnecessary. Mobile operating systems have, for the most part, eliminated the 'Quit' command completely. In Ubuntu, the messaging and sound menus will help us do the same."
to post comments

Thomas: That's it, we're quitting

Posted Mar 8, 2011 15:34 UTC (Tue) by Rehdon (guest, #45440) [Link]

Does this mean that Gnome won't become a giant 'Quit' button at some point in time, as some people fear?

(Sorry, couldn't resist :)

Rehdon

Thomas: That's it, we're quitting

Posted Mar 8, 2011 15:51 UTC (Tue) by JoeBuck (subscriber, #2330) [Link] (18 responses)

Because mobile phones have eliminated "Quit", everyone has to download task-killer apps to forcibly cause applications to quit.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:02 UTC (Tue) by mcoleman (guest, #70990) [Link] (5 responses)

Yup. Taking control away from the user is terrible design.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:26 UTC (Tue) by drag (guest, #31333) [Link] (1 responses)

alt-f2 xkill

:-P

Thomas: That's it, we're quitting

Posted Mar 11, 2011 18:08 UTC (Fri) by oever (guest, #987) [Link]

The key combination ctrl-alt-esc also shows a 'kill cursor' that can be used to kill an application. This combination has the advantage over xkill that you can cancel with esc.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 0:34 UTC (Wed) by jengelh (subscriber, #33263) [Link] (2 responses)

Sounds like mwm/twm. (If there is a WM key binding or mouse way to send a close event, I'm all ears.)

Thomas: That's it, we're quitting

Posted Mar 10, 2011 18:57 UTC (Thu) by proski (guest, #104) [Link] (1 responses)

Or Windows 3.1. It only had "minimize" and "maximize". The "quit" button was "invented" in Windows 95.

Thomas: That's it, we're quitting

Posted Mar 10, 2011 19:04 UTC (Thu) by jengelh (subscriber, #33263) [Link]

But in Windows 3.x (and 2.x), you could still send a window close event by going through the icon menu (top left of the title bar of a window).

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:05 UTC (Tue) by nix (subscriber, #2304) [Link] (9 responses)

It's a good thing Ubuntu already has one, then. Also, most modern non-mobile systems have so much memory that you barely notice the memory hit of a sane application: adding special menu options to confusingly optimize this memory usage seems silly. (Note: I do not include OpenOffice in the 'sane' category here.)

I suspect what should actually be done is to have the quit options appear only if your machine seems to be short of memory, or perhaps have the system warn you of the impending shortage and ask you what to kill. But I agree with the general thrust of the article: 'Quit' is a confusing historical wart that is a bad solution to the problem it purports to solve. Something better should be possible.

(One sort of thing should still be quittable: applications with large quantities of dynamic state, like Emacs, so you can reset that state if need be. Actually, what we want there is probably a 'reset to startup' option, which does, in effect, a quit-and-restart...)

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:11 UTC (Tue) by MisterIO (guest, #36192) [Link]

No, thanks. I want the quit button.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:48 UTC (Tue) by NAR (subscriber, #1313) [Link]

It's not just the memory (and not just OpenOffice - Firefox or GoogleEarth can also use lots of memory), it's space on the taskbar or notification bar too.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:24 UTC (Tue) by jpnp (guest, #63341) [Link] (1 responses)

I suppose for a particular definition of sane application you could make that case. But, for me, looking at the apps which are considered normal desktop use (including OOo, Firefox, etc), that's not true today.

I'm seeing a cases here of the OOM killer in Ubuntu 10.04 destroying X sessions on Ubuntu 10.04 machines left overnight with only standard office/internet apps (even on machines with 6G RAM/2G swap).

Now, I don't think the quit item makes a lot of sense any more. I currently have 5 office documents and 3 PDFs open. I cannot see the case where I'd like to close the 4 that are .doc plus one that is .ods together, but not those that are PDF, but that's what I currently have.

However, I do think we need much better interfaces for managing application/resource control when we do away with File->Quit.

Android only manages to avoid app close, by telling every app that it must cope with being randomly killed at any time (so must always keep state on disk).

Thomas: That's it, we're quitting

Posted Mar 9, 2011 2:33 UTC (Wed) by bluss (guest, #47454) [Link]

We're doing something similar on the desktop. All serious applications have "recover from crash" modes whose use are basically more of a rule than an exception.

GNOME has also killed complicated session management in favor of applications that always start back into roughly the state they were in the last time.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 21:57 UTC (Tue) by zaitcev (guest, #761) [Link] (1 responses)

It's not just memory. Most of these applications sit there rifling through my disk at the most unopportunate moments, and generate the network traffic.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 14:13 UTC (Wed) by nix (subscriber, #2304) [Link]

That's a different problem. For power management reasons, if nothing else, applications should idle as much as possible.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 23:57 UTC (Tue) by russell (guest, #10458) [Link]

So, let's undo all the hard work people have done to make our machines faster. Have you mentioned that to kernel developers yet? "Oh by the way, we user space programmers are going to change everything you ever knew about memory management. Time to start again, thanks."

Thomas: That's it, we're quitting

Posted Mar 9, 2011 1:04 UTC (Wed) by quotemstr (subscriber, #45331) [Link]

In the age of unified memory caches, the concept of "short on memory" has no meaning. If there is memory pressure, the MM system will page out infrequently-accessed pages and use te newly-freed pages for more urgent allocations. Killing a process has two effects under this scheme: first, it allows the memory manager to discard private pages held by the process that just exited, adding more pages to the free pool, and decreases the reference counts of shared pages. Absent odd situations like mprotect, it doesn't increase the total number of bytes that can be stored in RAM (assuming an adequate pagefile size).

Thomas: That's it, we're quitting

Posted Mar 10, 2011 12:12 UTC (Thu) by modernjazz (guest, #4185) [Link]

Some of us do run applications that are intended to use the memory of the machine---10 gigabytes of RAM is not a luxury that lets me play more YouTube videos, it's something that lets me get real work done. Once I've finished analyzing a 6GB data set, I'd like to quit the application to free memory. But presumably "serious" applications will retain a quit button.

Still, I agree with taskbar clutter being a major problem; even with 4 desktops, once 15 or more separate applications are running, things start to be a bit of a mess. Quitting applications is a wonderful way to reduce that clutter, and decrease start-up time next time I log in.

Getting Rid Of “Quit”

Posted Mar 9, 2011 8:38 UTC (Wed) by ldo (guest, #40946) [Link] (1 responses)

JoeBuck wrote:

Because mobile phones have eliminated "Quit", everyone has to download task-killer apps to forcibly cause applications to quit.

Speaking specifically for Android, such “task-killer” apps seem to be more trouble than they’re worth. Android is carefully designed in how it manages memory—see here for the details, and search for “kill” to find the relevant bits.

Getting Rid Of “Quit”

Posted Mar 10, 2011 20:13 UTC (Thu) by qubes (subscriber, #2562) [Link]

I find it amazing that you think the "quit" box antithetical to your design but accept as fact that all applications are well written and behave wrt. memory/cpu use. Degenerating the use of app killers on android is odd when they have many real needs. Misbehaving apps being the first that almost every android sales person installs for new phone buyers.

While I will wait to see a finished product, I find it sad that removing control seems to be the primary focus of gnome3 and the unity desktop. Until the Singularity, I'd like to be in control of my desktop and servers, thanks.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 15:54 UTC (Tue) by mennucc1 (guest, #14730) [Link] (24 responses)

What about the "save" button? I always dream of removing the "save" button. In my ideal O.S. environment, a document would "autosave" every few seconds; and the "save" button would be substituted by a "checkpoint/save version" button, that would save a snapshot of the current state of the document (possibly with a comment attached); moreover some automatic snapshots would be saved for me ("yesterday", "10 minutes ago", "two weeks ago"). This would be more similar to the desktop/pencil/paper paradigm : you don't "save" your work on paper, but you may keep copies of the work-in-progress (e.g. by xeroxing). Some programs, such as OpenOffice, already have this feature; and any document can be saved in multiple versions by using VCSs - but VCSs are usually to complex for the lay-users, so I would propose of just integrating it in the usual programs (or in the O.S. infrastructure). Think of it, it would be like integrating the "Mac Os Time Machine" inside the workflow of each and any application.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:29 UTC (Tue) by vonbrand (guest, #4458) [Link] (14 responses)

"Autosave" is terrible: It saves at random points in time, not logically consistent states. I was very much relieved when I could (finally!) turn the blasted thing off in MS Office, when the system (and application) got stable enough.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:44 UTC (Tue) by DOT (subscriber, #58786) [Link] (10 responses)

If autosave leaves your document in an inconsistent state, the application is broken. Autosave should obviously be an ACID transaction.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:57 UTC (Tue) by ssam (guest, #46587) [Link] (2 responses)

suppose you are writing a piece of text. having a half written sentence is n

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:50 UTC (Tue) by DOT (subscriber, #58786) [Link]

I see your point. But here we have to make a distinction between saving as in salvaging, and saving as in publishing. In databases, filesystems, git, publishing, ACID operations are important. You want your items to move from one consistent state to another. But with salvaging, grammar consistency is less important than saving a few hours of work. You can finish your half-written sentence, but you can't get back those hours of work.

So there should be a distinction between autosaves and manual saves. LibreOffice has the ability to tag versions of your document. Normal saving then becomes the salvaging type, and "Save New Version" becomes the publishing type. Some UI work should make salvaging automatic and publishing more prominent.

You're asking for the difference ...

Posted Mar 8, 2011 22:05 UTC (Tue) by kmself (guest, #11565) [Link]

... between application save state, and a versioning system.

The more I'm using tools such as git, the more I'm getting convinced that this is what we want underneath a lot of user interactions.

I'd rather have an edited file saved in an inconsistent state than not saved at all. If the cursor point is saved as well, you'll open and go directly to the point of your most recent edit. All very graceful.

Also all managed by that highly advanced GUI Joe-sixpack user word processor, vim.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:59 UTC (Tue) by nteon (guest, #53899) [Link]

having the bits be consistent and the logic/ideas/grammar consistent are two separate things. I suspect your talking about the former and vonbrand talking about the latter.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:04 UTC (Tue) by jpnp (guest, #63341) [Link] (5 responses)

I think the point is that you, the author of a particular document or file, know what are good "consistent" points in your work to save/restore from, not that the application itself can't write a parsable file.

E.g. you'd rather save a text document between sentences (or more likely between expressing particular thoughts), or a code file between updating functions. A file auto-saved halfway though fixing a bug touching several places would be a dangerous thing.

Most applications know only about your actions and very little about your intent.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:22 UTC (Tue) by dgm (subscriber, #49227) [Link]

This is just another face of the end-to-end principle that most UI designers just ignore.

Paraphrasing Einstein: Make everything as simple as possible, but no simpler.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 18:37 UTC (Tue) by Cyberax (✭ supporter ✭, #52523) [Link] (3 responses)

That's why the "save" functionality must be removed and "checkpoint" functionality must be added.

So you won't lose any work (including undo/redo state and clipboard ring) in case of accidental alt-f4, but you'd still be able to tag meaningful checkpoints.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 15:04 UTC (Wed) by sorpigal (subscriber, #36106) [Link] (2 responses)

What's in a name? If you 'save' to a storage system that automatically creates new versions, then that's 'checkpointing', right? You still haven't removed the "save button" or the process of manually saving something if you merely change the label.

Adding automatic versioning is not a bad idea, removing "save" is.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 15:54 UTC (Wed) by jpnp (guest, #63341) [Link] (1 responses)

So without user input, how do you propose that work is "checkpointed" between conceptually useful chunks of work, rather than just at random time points?

Unless the application knows what are relevant chunks of work, it's going to require user input. My version control doesn't automatically check in code every hour, I tell it when I've finished a particular task, because only I know, and because only that's useful.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 16:34 UTC (Wed) by mennucc1 (guest, #14730) [Link]

In my proposal, I never stated that everything would be "without user input". My proposed paradigm would be : (1) autosave/checkpoint w/o user input (saving undo info as well) (2) save finished, conceptually useful chunks of work, when the user requests so.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:11 UTC (Tue) by mennucc1 (guest, #14730) [Link] (2 responses)

The workflow paradigm that was forced on us all in last ~20 years is: there is a "file on the hard disk", and there is a "open version of the file" in the application; we work on the "opened" version, and then save it in the hard disk.
I think that this paradigm is not at all natural, for good reasons,
  • when the application is not running, you have "one concept" of the file content; when the application opens the file, you have "two concepts";
  • I clearly remember trying to explain it to people that never used a computer, and it IS complex and unnatural to grasp at first,
  • indeed it is far away from the everyday experience: any physical container contains "one content and only one content" at any given moment; and when you change the content, there is no intermediate moment from when you "change it in the application" and when you "change it in the hard disk"
The above workflow paradigm was forced on programmers by the slowness of old storage (i.e. floppies) and they forced it on us, and it lasted. But if you could throw it all away and think and design it from scratch, what would you do? My idea (and I have had it for many years) is that we should change it all, and do it away with the "save" button, to be replaced with a "save version" button, "autosave" (and file locking, to avoid that different application may edit the same file... but this is mostly done nowadays).
The fact that "autosave" would not save at "consistent states" is not relevant. (Note also that this is again what happens with physical media/container : when you write to a piece of paper, at a given moment you may have a non-finished content). The purpose of autosaving is not "I want only to save well-finished work" : you use versioning for that. The purpose of autosaving (and file locking) is there is only one present content of the file, and it is what you are looking at.
The fact that you use saving where versioning should be used was forced on you, and us all, by old programmers using obsolete harddware. What happens nowadays is that you were given saving and not versioning, and for 20 years you used what you could get to implement what you need. (I know indeed people that then create copies and copies of the same file, adding to the filename the comments that explain "the version"). Then someday someone with enough horsepower (Ubuntu? Apple?) will come to think of it, and they will change it.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 18:29 UTC (Tue) by vonbrand (guest, #4458) [Link]

That is similar to the workspace idea in i.e. APL: You don't load a file, you start a whole world in which you can manipulate pieces of code, run them, etc; on exit all is saved so you get the same state back when you return. It sure is a neat idea, but if you do several different tasks, you'd need one "world" for each, and sharing state between tasks becomes... interesting. Plus saving a state for later backtracking (if some experiment turns out bad) or sharing pieces with others doesn't look easy.

Disclaimer: I fooled around with APL a bit way back, perhaps I didn't "get" it completely.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 22:54 UTC (Tue) by gmaxwell (guest, #30048) [Link]

Versioning is _very_ useful even without abandoning the file/open-file paradigm and as you mention the hardware today is easily capable of good versioning support, yer we don't have integrated versioning.

Changing paradigms is difficult especially when a _hard_ prerequisite for the new paradigm is missing.

Why should anyone even bother worrying changing from the file/open-file model until after versioning is ubiquitous?

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:28 UTC (Tue) by foom (subscriber, #14868) [Link] (3 responses)

You might check out Apple's list of features for Lion. Amongst them...
Auto Save Past perfect. Say good-bye to manual saving. Auto Save in Mac OS X Lion automatically saves your work — while you work — so you don’t have to. Lion saves changes in the working document instead of creating additional copies, making the best use of available disk space. The lock feature prevents inadvertent changes from being saved and automatically locks documents after two weeks. And the revert feature returns you to the state the document was in when you last opened it, so you can feel free to experiment with confidence.*

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 15:11 UTC (Wed) by sorpigal (subscriber, #36106) [Link] (2 responses)

This fails when I have opened the same file twice on purpose, which I will do sometimes for configuration files and for other purposes. You might say "Just lock the file and don't allow two copies to be opened for writing" but now you're saying not "versioning is a better paradigm than saving" but "I think you don't need functionality that you use" and at that point you've gone off the rails.

If it's the same but better, great. If it's similar but less, no thanks.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 15:17 UTC (Wed) by foom (subscriber, #14868) [Link] (1 responses)

"Sorry but you don't exist", as another commentator said -- that's such an uncommon use case, that OSX's default applications can easily get away without supporting it. In fact, you *already* can't open the same file twice in the GUI apps -- it just brings forward the existing window.

That said, I'm sure emacs and vi won't start using this functionality anytime soon, so you should be safe with your config file editing. :)

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 17:50 UTC (Wed) by nix (subscriber, #2304) [Link]

Emacs has long moaned if you try to open the same file twice in different instances, and brought forward the existing file if you open it twice in the same instance. Of course, you can take control or ignore things in about fifty different ways: this *is* Emacs...

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 19:30 UTC (Tue) by nicooo (guest, #69134) [Link] (4 responses)

> This would be more similar to the desktop/pencil/paper paradigm

People use computers because this paradigm sucks.

I'd like every program to work like vim. Swap file + persistent undo + undo branches.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 19:57 UTC (Tue) by drag (guest, #31333) [Link] (2 responses)

Document editing applications should be backed by a distributed document revision system...

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 8, 2011 22:33 UTC (Tue) by nicooo (guest, #69134) [Link] (1 responses)

You'll have to wait until MS Office implements it. =]

LibreOffice imitates office to the point of using the same fail fonts.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 11:59 UTC (Wed) by mpr22 (subscriber, #60784) [Link]

I'm confused. What's wrong with Arial? Sans beats serif for readability at small-relative-to-resolution point sizes, and Arial looks better than any of the Free alternatives I've actually seen.

no "save", Re: Thomas: That's it, we're quitting

Posted Mar 9, 2011 14:17 UTC (Wed) by nix (subscriber, #2304) [Link]

In the interests of religious-war equality, Emacs has long had an autosave that works *right*, i.e. that saves after N keystrokes and when left idle, but *not* into the original file, but rather into another one in an easy-to-find location in case of disaster, and provides a 'recover everything that was running when my editor crashed' command (which perhaps should be run automatically but by default is not).

(Now its default undo/redo is pretty horrible compared to vim's. Of course there are packages to change it...)

Thomas: That's it, we're quitting

Posted Mar 8, 2011 15:57 UTC (Tue) by rsidd (subscriber, #2582) [Link] (1 responses)

Actually, I hate the fact that clicking the "close" button on the only open window of an application doesn't always quit the application these days.

So I use the "quit" in the menu. Now Ubuntu wants to take that away from me?

Next they'll remove xkill from the base system. And then the "ps" and "kill" commands.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:23 UTC (Tue) by dskoll (subscriber, #1630) [Link]

Yes, because the user-friendly way to handle these things is to reboot. That's what Microsoft does, and we all know that Microsoft is far more user-friendly than Linux.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:08 UTC (Tue) by MisterIO (guest, #36192) [Link]

Yeah, Epiphany doesn't have a quit button and you know what? I'm not using it! I use saner apps.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:11 UTC (Tue) by mpalamara (guest, #69530) [Link]

Removing quit is utter nonsense! Sometimes I actually want to stop an application. Ubuntu has so many problems and this is what they are thinking about! Fan-frickin-tastic!

Did you ever get the feeling...

Posted Mar 8, 2011 16:22 UTC (Tue) by dskoll (subscriber, #1630) [Link] (4 responses)

... that the GNOME developers, Ubuntu developers and Perl 6 developers have all discovered the same mind-altering drug? Or that they're all playing an elaborate hoax on us?

Did you ever get the feeling...

Posted Mar 9, 2011 3:18 UTC (Wed) by mrons (subscriber, #1751) [Link] (3 responses)

Maybe someone slipped me some of these drugs but I don't get the connection with perl6.

I doubt that perl6 is lacking the equivalent of a "quit button".

Did you ever get the feeling...

Posted Mar 9, 2011 13:57 UTC (Wed) by dskoll (subscriber, #1630) [Link] (1 responses)

The inclusion of Perl 6 in my list is because of its wild experimentation, gratuitous incompatibilities, and disconnect from real-world users.

Did you ever get the feeling...

Posted Mar 10, 2011 18:40 UTC (Thu) by jmorris42 (guest, #2203) [Link]

The difference is the Perl people are only half mad. Yes they have spent a decade dreaming and don't have a runnable implementation yet but they kept Perl 5 going while they work. And at this point the language has changed enough they really should call it something else because calling it Perl will just confuse people into thinking they can port existing code to it. But if they ever ship a working product it looks like it will be worth the time and effort. They wisely decided if they were going to break backward compatibility they might as well well and truly break it and go ahead and fix everything that years of hard use had revealed as weaknesses in the original design.

The GUI people on the other hand are shipping their half thought out concept pieces downstream into mainstream distributions. I finally downloaded F15 Alpha so I could see what the yelling was all about with GNOME3. Oh. My. God. So not ready for prime time words fail. But it is going to ship next month. I put one bug into bugzilla and and declared myself done with GNOME until this madness passes.

Did you ever get the feeling...

Posted Mar 9, 2011 14:31 UTC (Wed) by nix (subscriber, #2304) [Link]

Perl 6 has 420 buttons labelled 'quit' in different languages, all of which do something subtly different and none of which actually quits.

The button which quits is labelled 'xyzzy'.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:26 UTC (Tue) by petegn (guest, #847) [Link] (7 responses)

well maybe there needs to be a BIG quit button attached to Canonical and Gnome

Now KDE rules the roost

Thomas: That's it, we're quitting

Posted Mar 8, 2011 19:27 UTC (Tue) by welinder (guest, #4699) [Link] (6 responses)

The KDE that, at some point, decided that you couldn't possibly want
to auto-mount your usb hard disks?

Thomas: That's it, we're quitting

Posted Mar 8, 2011 20:29 UTC (Tue) by nybble41 (subscriber, #55106) [Link]

Which version of KDE was that? Excluding some of the early point releases in the KDE4 re-write for which the functionality may not have been implemented yet, I do not recall any version of KDE since the feature was first introduced which did not allow you to auto-mount your USB disks. Certainly the current version does, although the default is to only auto-mount disks which have been mounted manually at least once for safety reasons. It can also be trivially configured to auto-mount all removable disks, or to never auto-mount, or to auto-mount only certain disks of your choosing. (These settings are accessible via the Device Notifier applet.)

Thomas: That's it, we're quitting

Posted Mar 9, 2011 0:19 UTC (Wed) by GrueMaster (guest, #57504) [Link] (4 responses)

Actually, automounting is disabled by default, but can easily be enabled. I do a lot of development work that requires swapping SD cards constantly and raw writing images to them. If my system automounted every time I inserted one (which ubuntu/gnome does), I would scream.

But this thread is slightly off topic from the original post. To that, I say get rid of quit and see how it goes, but allow an option for users to re-enable it. Then collect feedback to see if that was indeed a good decision.

Personally, I don't like the idea, as I have a desktop computer with dual monitors and don't shut down when I close the lid. My desktop is usually up 24/7 with the only reboots being for release updates/upgrades. Being able to close an app and all it's windows is good for me. Otherwise I get too much clutter.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 10:03 UTC (Wed) by dany (guest, #18902) [Link] (3 responses)

Actually it does make more sense to enable automounting by default and for power-users like you to have option to disable it.

Eliminating quit buttons in apps? Horrible idea. Why to fix something, that is not broken?

Automounting

Posted Mar 9, 2011 15:07 UTC (Wed) by hmh (subscriber, #3838) [Link] (2 responses)

No, automounting random crap by default is not a good idea. It doesn't take a genius to know this: you wouldn't drink for a bottle that a random stranger stuck to your mouth either. Most of us won't do it even if it looks like a bottle of a beverage we're used to, and those of us who do, will learn the hard way not to sooner or later.

Automounting random crap is in fact an EXTREMELY BAD idea. The only thing that would be worse than that is actually accessing any data on the automounted volume (which is exactly what will happen most of the time). You don't even need "autorun" for it to be dangerous. We've had exploits of this kind in the past.

It is not even just a security issue. If it is not something that should be mounted at that time (e.g. a pendrive with a damaged filesystem you want to repair, etc), it can cause data corruption and data loss.

Automounting previously-known media is a compromise, although nearly as unsafe from a security PoV. It will at least prevent some mistakes. Still not a good default IMO, but less outrageous than automounting random crap by default.

Automounting

Posted Mar 12, 2011 1:25 UTC (Sat) by cras (guest, #7000) [Link] (1 responses)

How about adding a new icon showing that there's a new USB drive inserted into the system. Kind of like it already shows (I assume). The difference being that it would only get automounted when you (double)click the icon. Security & user-friendliness all there.

Automounting

Posted Mar 12, 2011 1:41 UTC (Sat) by hmh (subscriber, #3838) [Link]

This already exists in KDE, and it is the default mode of operation AFAIK. Surprisingly enough, it works really well, too.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:37 UTC (Tue) by danielpf (guest, #4723) [Link]

Incredibly shortsighted. Do these amateurs realize that desktops may need to run indefinitely because are either multiusers or are often on heavy computing load, so reboot is not an option? And sorry killing a job after ps/top is definitely not user friendly.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:40 UTC (Tue) by mikov (guest, #33179) [Link] (5 responses)

So they are talking about applications which have multiple _independent_ top level windows and "Quit" closes them all.

The natural way to handle this is to have the application "quit" when the last window is closed. The user doesn't need to know that the separate windows happen to "belong" to the same "main" process, because that is an optimization.

So, yes, technically there is no need to have a separate "Quit" command. "Exit" or "Close" still need to be present and they close the window. The only question is whether the application continues running after all visible windows have been closed. The answer is that it shouldn't.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:50 UTC (Tue) by ewan (guest, #5533) [Link] (4 responses)

That model doesn't deal with the situation where you want to quit the application without first closing all the windows; I do that routinely with Firefox when I want it to go away, but save the running state to reload next time.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 20:34 UTC (Tue) by mjthayer (guest, #39183) [Link] (3 responses)

> That model doesn't deal with the situation where you want to quit the application without first closing all the windows; I do that routinely with Firefox when I want it to go away, but save the running state to reload next time.

That sort of session management doesn't necessarily have to be implemented individually in each application though. For that matter I sometimes wonder why tabbed document management has to be implemented in individual applications rather than at a higher level, window manager or perhaps better something in-between. (I know that tabbed window managers exist, but Firefox still rolls its own.)

Thomas: That's it, we're quitting

Posted Mar 9, 2011 1:52 UTC (Wed) by tetromino (subscriber, #33846) [Link] (2 responses)

> That sort of session management doesn't necessarily have to be implemented individually in each application though.

For non-trivial applications, session management absolutely must me implemented in the application individually. This is because there is more to an application's state than just the list of documents open. At the very minimum, the application needs to scroll to the place in the document that you were at when you closed it. For a web browser, there are also things like HTML forms that you have partially filled in but have not submitted yet.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 1:53 UTC (Wed) by tetromino (subscriber, #33846) [Link]

s/must me/must be/

Thomas: That's it, we're quitting

Posted Mar 9, 2011 8:55 UTC (Wed) by mjthayer (guest, #39183) [Link]

> For non-trivial applications, session management absolutely must me implemented in the application individually.

You are right, and I did not express what I meant well. Of course the applications need to support session management, but not necessarily to include the session manager as well. I think what is annoying me is the multi-tab paradigm which has succeeded MS's multiple document interface, where a single process (or paradoxically with modern web browsers one process with many sub-processes) manages multiple documents/web pages whatever in a way that I think would be better done by multiple copies of the application with appropriate window manager, session manger etc infrastructure.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 16:53 UTC (Tue) by jimparis (guest, #38647) [Link] (9 responses)

I understand their concerns about usability -- particularly in their examples about inconsistency, like comparing LibreOffice and the web browser versus more standalone programs.

But I hope they avoid the temptation to control application lifecycle automatically. In my experience, that's by far the worst aspect of Android. Android tries to blur the distinction between switching away from an app and closing an app, and lets the system choose between the two automatically. I had the G1, which is relatively RAM limited. Trying to switch between a SSH session and the web browser was an exercise in frustration, because it would frequently decide to kill SSH when the web pages started using too much RAM. Multitasking was essentially unusable.

The "workaround" was to frequently switch between the two apps in a futile attempt to keep them "fresh" in the mind of some mystery heuristic! This wasn't a special case like the OOM killer, it's the normal mode of operation for Android.

I've since switched to the N900 and it's a breath of fresh air -- I can open up five browser windows and two xterms, and never have to worry that they'll magically disappear. Sure, things might slow to a crawl as it starts to hit swap, but that's a much more graceful degradation, and I still get to choose what to close to get things back to normal when I need to.

But that doesn't mean I need a "quit" button. My N900 doesn't have many of those. Closing all of the Firefox windows quits Firefox. I think that as long as Ubuntu is working on just eliminating the distinction between close and quit, and not trying to drastically redefine how app lifecycles work, it's a good thing.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:27 UTC (Tue) by ewan (guest, #5533) [Link]

I think that as long as Ubuntu is working on just eliminating the distinction between close and quit, [...] it's a good thing.

Trouble is, it's a real distinction. 'Close window' and 'Quit whole app' aren't the same thing on apps with multiple windows, nor on things like IM clients and music players that run in the background.

You can try to eliminate enough flexibility to make them the same, but you'd have to eliminate quite a lot. The lack of thought that's gone into this i neatly shown by the suggestion from MPT that an appropriate replacement UI for when you really do mean 'Quit' is to kill the task from the System Monitor app.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 20:39 UTC (Tue) by alankila (guest, #47141) [Link] (7 responses)

It seems to me that the SSH program did not flag itself as a "foreground service", which should tell android that killing the application would be disruptive to user experience.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 21:16 UTC (Tue) by jimparis (guest, #38647) [Link] (6 responses)

Maybe. I browsed the source and don't see a call to setForeground or startForeground. But it looks like it's still a mess. Should every app explicitly state "I'm more important than everyone else"?
  • setForeground was "not advertised much because it should not be used by most applications". If ssh doesn't fall into the category of behavior that matches their idea of "most applications", then they misunderstand how I want to use my phone.
  • The G1 ran Android 1.5. In Android 2.0 "setForeground() now does nothing" and you have to use startForeground instead. Clearly it was not a function that was intended to be used much if at all if they're willing to make it a no-op.
  • I doubt it would have helped much anyway. The Process Lifecycle docs claim "in a foreground state ... It is still theoretically possible for the service to be killed under extreme memory pressure from the current foreground application, but in practice this should not be a concern".
Guess what: my phone was always under extreme memory pressure. That's due to the hardware guys, sure, but the system design of "we send SIGKILL willy-nilly and don't have swap" really doesn't help. Their bizarre concept of how application lifecycles should work is exactly what I'm hoping that the Ubuntu folks will avoid.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 4:43 UTC (Wed) by cmccabe (guest, #60281) [Link]

It really sounds like startForeground is what your ssh client should do, at least on Android 2.0 and better.

> Their bizarre concept of how application lifecycles should work is exactly
> what I'm hoping that the Ubuntu folks will avoid.

I've been told that some of the process lifecycle ideas came from BeOS. It seems to be hard to actually find any information about BeOS, though.

You don't exist...

Posted Mar 9, 2011 8:23 UTC (Wed) by khim (subscriber, #9252) [Link] (3 responses)

If ssh doesn't fall into the category of behavior that matches their idea of "most applications", then they misunderstand how I want to use my phone.

Wow. Talk about hubris. We have canned answer to such rants in our office: you don't exist. Meaning: you belong to such a small category of users that your needs should not affect design of an app (OS, etc). Sure, guys at Google do understand how you want to use your phone (because they are quite similar to you, in a sense) - and they explicitly ignore you. Yes, most applications don't need to be kept in memory at all times and these which do must be split in headless service and non-persistent UI process. It's explained in Application Fundamentals!

Now, if you do want to control everything you can use N900. Cherish it because it'll go away: "You don't exist" so catering to your needs don't bring money (that's why Nokia ultimately failed: developers were allowed to design the interface... most users hated it)... Note that iOS approach to multitasking is exactly the same as Android's approach to multitasking - because it's the only one approach which makes sense for 90% of users.

Their bizarre concept of how application lifecycles should work is exactly what I'm hoping that the Ubuntu folks will avoid.

Well, Ubuntu is in serious disadavantage: it has many existing applications and people extect to continue to use them. Thus they can not just use "sane" (see above) approach and will try to use palliatives. This is dangerous road: today Ubuntu is attractive for 5% of people "who don't exist" and atrocious for 95% of "normal users". It's pretty clear what they should do to cater for the needs of normal users - but in the interim there are risk of creating something which will repulse all users: not enough control for people "who know what they are doing" and not enough automation for "normal people"...

You don't exist...

Posted Mar 9, 2011 23:14 UTC (Wed) by alextingle (guest, #20593) [Link]

My wife uses Ubuntu, and says it's perfectly fine. And she's well outside your "5%" of geeky users.

You don't exist...

Posted Mar 10, 2011 20:32 UTC (Thu) by qubes (subscriber, #2562) [Link] (1 responses)

...
Thanks for finding computer use hard atm, but all of my users find current Ubuntu _easier_ to use than current releases of Mac or Windows. Granted I don't support alot of users, but the paradigm you disparage is rather good for most usage modes.

I think I would agree with you more if *all* of linux was headed to a very constrained platform -- which given the innovation on the hardware side I think would be a mistake. CPU, screen size, memory and storage space are all rising faster than well written* software can use it.

You are sure?

Posted Mar 13, 2011 16:15 UTC (Sun) by khim (subscriber, #9252) [Link]

Granted I don't support alot of users, but the paradigm you disparage is rather good for most usage modes.

Funny that. You've said it yourself:

Thanks for finding computer use hard atm, but all of my users find current Ubuntu _easier_ to use than current releases of Mac or Windows.

Perhaps. It depends a lot on what kind of support is availabl for Windows and Ubuntu. But the key word here is Ubuntu. Ubuntu tries to make Linux easier to use - and often does it by removing choice so if your users are Ubuntu users and not Arch Linux or Gentoo users, then my point still stands. Perhaps Ubuntu is already streamlined and trimmed enough to fight Windows and Mac and I just missed that point, but the fact is: it all become possible because Ubuntu decided to concentrate on creating working desktop and not on number of options available.

I think I would agree with you more if *all* of linux was headed to a very constrained platform -- which given the innovation on the hardware side I think would be a mistake.

Oh, sure. Distributions for geeks will exist for a long time yet. But they are no longer mainstream and they affect direction of Linux mainstream development less and less.

E16 is still out there but it's not longer the default Window manager for GNOME - these times are long gone and I doubt they'll ever return.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 11:00 UTC (Wed) by alankila (guest, #47141) [Link]

I personally like the illusion presented by android that applications are always alive. I wouldn't doom this concept just because you had an early Android phone that did not have enough RAM to make the illusion work well.

Why is "quit" annoying?

Posted Mar 8, 2011 17:02 UTC (Tue) by tjc (guest, #137) [Link] (3 responses)

I admit that I'm easily annoyed by bad UI design, but I've never been annoyed by the presence of a "quit" menu item or button. I tried several times just now, but somehow I just can't be annoyed by this, although I'm pretty sure I could be annoyed by the absence of the same.

Does anyone know anyone who is actually annoyed by this? It sounds like something out of a Monty Python skit to me. "Quit that! Ooh, I said it again..."

Why is "quit" annoying?

Posted Mar 8, 2011 18:29 UTC (Tue) by PO8 (guest, #41661) [Link] (2 responses)

Oh, the "Quit" menu item drives me *crazy*. You pull down the File menu and there it is, staring you in the face. You can't ignore it. You certainly can't *not use it*. It just sits there, taunting you. Saying "I am a forbidden relic of a bygone age, inviting mere humans to meddle in that which they do not understand and that with which they should not interfere."

Can't you hear it whispering in your sleep?

Why is "quit" annoying?

Posted Mar 8, 2011 19:57 UTC (Tue) by cyd (guest, #4153) [Link]

Ph'nglui mglw'nafh Quit Button R'lyeh wgah'nagl fhtagn

Why is "quit" annoying?

Posted Mar 8, 2011 20:05 UTC (Tue) by cmccabe (guest, #60281) [Link]

http://www.youtube.com/watch?v=ku2wFaaPAzI&feature=re...

Thomas: That's it, we're quitting

Posted Mar 8, 2011 17:23 UTC (Tue) by Trelane (guest, #56877) [Link] (2 responses)

hmm

could this be combined with per-process hibernate?

Thomas: That's it, we're quitting

Posted Mar 8, 2011 21:55 UTC (Tue) by error27 (guest, #8346) [Link] (1 responses)

I doubt the per process hibernate (check point, restart) is going to get merged any time soon. It seems like something that won't work 100% and so it will be a support nightmare.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 1:07 UTC (Wed) by Trelane (guest, #56877) [Link]

Dang. It was quite an interesting idea.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 18:01 UTC (Tue) by Tara_Li (guest, #26706) [Link] (4 responses)

Yet another step along the way to Microsoft's "We know where you want to go today." And an example of why I'm using Enlightenment 16 for my window manager, instead of the default, on Ubuntu 10.10.

With E16, I can - raise windows, lower windows, minimize, maximize (height, width, or all), restore to previous state, close, *FORCE-CLOSE*, and I've occasionally even had call to change the border style from what the application/dialog chose, to the default border style so I can resize the dialog, or even close it.

A computer isn't a car - I can see the Canonical people deciding that the ignition key isn't needed in a car, because you can just leave it running and lock the door behind you with the neat little beeper.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 23:11 UTC (Tue) by PaulWay (guest, #45600) [Link] (3 responses)

Heh. You've never driven a car with a turbo timer, then. For some turbochargers you have to leave the engine running so the turbine can spin down to reasonable speeds with oil still in the bearings. Turning the engine off while the turbine is still going at 100,000 RPM and the bearings have no oil is a good way to ruin it. So they have a timer on the engine that keeps it running even when you've taken the key out of the ignition and locked the car so the oil pressure is kept up to the bearings, then turns it off after a minute or so when the turbine has spun down.

And how many people leave the fridge door open between getting the milk out and putting it back? Do we all close the door to keep the heat out every time, no matter how quick the two opening events might be?

Do we all turn off the light as soon as we've left the room if we're going to leave it longer than 17 seconds (for fluorescents) or 1/10th of a second (for incandescents and LEDs)? Or do you just leave it running because you're "not going to be out of the room that long" and "it doesn't hurt to leave it running for a bit".

And yet we're all madly keen on managing the memory of our computers in a desperate attempt to be better at it than the kernel.

Well, you have your wooden horses to ride. Go have fun dragging them around.

When I first got my Android phone I worried obsessively about quitting applications and killing them off. Then I realised it made no practical difference - battery life seemed relatively the same no matter how many applications I was running (it mattered more what they were doing). Why worry about whether the application has tucked itself away in the background or quit or whatever, when most of the time it doesn't matter? If I need to kill a rogue application I can, with the in-built tools for the job.

I suspect rogue applications - something that we need to kill because it is taking up too much memory or is doing something that we don't want it to do - are going to be the major problem to face. And for that I'd just invent a new gesture - double-click on the 'close' button or something - to mean "operating system, kill this process". If that turns out to close every Firefox tab I have open, then it's my choice at that stage, and I'm asking for a good reason rather than just because that's the default way to make the application go away.

Anyway, I think it's a fascinating idea, and I can see it's going to face a huge amount of resistance from all those die-hard hackers who love manually controlling everything.

Have fun,

Paul

Thomas: That's it, we're quitting

Posted Mar 9, 2011 4:27 UTC (Wed) by cmccabe (guest, #60281) [Link] (1 responses)

> When I first got my Android phone I worried obsessively about quitting
> applications and killing them off. Then I realised it made no practical
> difference - battery life seemed relatively the same no matter how many
> applications I was running (it mattered more what they were doing). Why
> worry about whether the application has tucked itself away in the
> background or quit or whatever, when most of the time it doesn't matter?
> If I need to kill a rogue application I can, with the in-built tools for
> the job.

Android was designed from the ground up to operate with a different process lifecycle than Linux on the desktop. When the Android Memory Manager is low on memory, it asks certain Activities to save their state to a Bundle (essentially a few bytes of state.) Once it has saved that bundle, it terminates the process, freeing up more memory. For example, a Bundle for an e-reader program might just have the name of the book the user was reading and the page number he was on. The application has to be written so that it can be frozen at any moment. Later, the Activity will be restored from the bundle if the user switches to it (or under certain other conditions.)

Android applications don't have a single point of entry like main() on desktop Linux. Instead, there's these things call Activities which take up the whole screen. Activities are not supposed to do very much unless they're in the foreground. There's also Services, which are more similar to daemons in desktop Linux. To share data between Activities, you must use a ContentProvider.

On desktop Linux, if you run out of memory, the VM swaps out pages of memory when it thinks they won't be needed for a while. Switching to a process that has been mostly "paged out" of memory often takes longer than opening a fresh instance of that program. You couldn't really write something like the Android MemoryManager for Ubuntu, because it's not clear what the relationships between applications are. Applications can pass data back and forth along named or unnamed pipes, shared memory, writing to files in /tmp, sending TCP messages to localhost... etc. This is historically why the OOM killer has been such a disaster. The kernel can't choose a sensible set of processes to kill unless you explicitly tell it how using something like cgroups.

Sticking feathers up your butt does not make you a chicken. Taking away the quit button doesn't magically give you the equivalent of the Android Memory Manager.

Thomas: That's it, we're quitting

Posted Mar 10, 2011 7:24 UTC (Thu) by ajf (guest, #10844) [Link]

Sticking feathers up your butt does not make you a chicken. Taking away the quit button doesn't magically give you the equivalent of the Android Memory Manager.

You've nailed what troubles me about a number of upcoming UI changes that I've read about recently, where it seems UI ideas from a fashionable but very different platform (especially a mobile phone platform) are adopted without regard for the underlying principles that make them work.

Thomas: That's it, we're quitting

Posted Mar 18, 2011 10:22 UTC (Fri) by cas (guest, #52554) [Link]

Do we all turn off the light as soon as we've left the room if we're going to leave it longer than 17 seconds (for fluorescents) or 1/10th of a second (for incandescents and LEDs)? Or do you just leave it running because you're "not going to be out of the room that long" and "it doesn't hurt to leave it running for a bit".

And yet we're all madly keen on managing the memory of our computers in a desperate attempt to be better at it than the kernel.

Well, you have your wooden horses to ride. Go have fun dragging them around.

you know, you're absolutely right.

light switches ought to be abolished - outlawed! - and replaced with automatic timers and motion sensors.

nobody ever really *needs* to manually switch a light on or off, they just think they do out of habit. a timer or sensor is a MUCH better judge of their needs than they are.

and when we've got rid of all the light switches, we can make a start on all the other on/off switches...none of them are necessary either

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 8, 2011 18:14 UTC (Tue) by PO8 (guest, #41661) [Link] (13 responses)

Quite aside from whether removing "Quit" is a good idea (it isn't), how does Canonical propose to bell this cat? I am confident that most major applications won't accept a patch that removes Quit from their file menu. Does Canonical propose to fork (or at least maintain a patch for) every application the user is likely to encounter? They could just remove it from the "most popular" apps; wouldn't that make the user experience much more confusing rather than less? I guess that they could try to hack Gtk+ or Gnome to automatically detect and remove Quit menu entries; this seems like a plan fraught with many perils, not the least of which is that it probably wouldn't work for everything.

Removing "Quit" is an interesting fantasy, but (fortunately) I think the situation is out of Canonical's hands on this one. Looks to me like they'll have to content themselves with gratuitously moving buttons around on their default window manager.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 8, 2011 20:03 UTC (Tue) by Tara_Li (guest, #26706) [Link] (11 responses)

Ah, but they're aiming to pretty much do just that - they want to remove the menu bar from applications, and do as Apple does in OS X where there's one big menu bar at the top that changes as you move from application to application. There was, in fact, talk about changing GTK and the other toolkits to do that automatically.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 8, 2011 20:41 UTC (Tue) by cmccabe (guest, #60281) [Link] (10 responses)

I feel a little bit bad saying this, but I think the time for creativity in desktop user interface design is over. Users want their desktops to look somewhat similar to Windows. All of these re-arrangements and re-envisionings are a step backwards and a distraction from the main task, which is polishing the Linux desktop. And if they result in the desktop looking different than Windows, they put up yet another barrier to switching.

You can try to look like Apple, but it's a bad choice. There are fewer Apple users than Microsoft users. Linux can compete somewhat effectively with Microsoft in the corporate space, but really can't compete with Apple on ease of use. Even if Linux could, Apple has a powerful brand and Apple users rarely consider the situation rationally.

I think the real story is that the Canonical and Gnome guys have seen the rise of Android and Meego. They feel like they're facing irrelevance, and they're ready to "bet the project" on desperate measures that they wouldn't have considered before. This makes me sad, because I do use the Linux desktop at home and at work. It does matter to me.

I guess as long as the Xfce mouse continues to scuttle through the ruins, things will be ok. Actually, I checked out the Xfce code and it looks very nice too. There's no confusion, either at the design or programming level, over in that quarter.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 8, 2011 23:30 UTC (Tue) by MisterIO (guest, #36192) [Link] (7 responses)

Since I took a look at gnome3 and unity(I never liked kde, so kde4 didn't change anything), I switched to xfce and frankly I'm still searching for a reason to go back to any of them! If these new plans become reality, I'll have even less reasons to still even think about them!

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 1:50 UTC (Wed) by dgm (subscriber, #49227) [Link] (6 responses)

I had one: you cannot use smb and sftp uris in Thunar until recently. Maybe next XUbuntu 11.04 is going to become my new default.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 3:14 UTC (Wed) by MisterIO (guest, #36192) [Link] (5 responses)

But you can still use nautilus on xfce.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 8:57 UTC (Wed) by dgm (subscriber, #49227) [Link] (4 responses)

Of course, and even Dolphin or Konqueror, but what would be the point then?

What is Gnome besides Nautilus and a panel? The libraries (most of them, at least) would come along with Nautilus. Kiss goodbye to any memory savings!

I don't mean that Gnome is that bad memory-wise these days. My current desktop is happy with 250 MB, with some terminals, a couple of Nautilus and Geany open. But we can do better. Specially application start-up times should be MUCH better.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 16:10 UTC (Wed) by MisterIO (guest, #36192) [Link] (2 responses)

Citing you: "you cannot use smb and sftp uris in Thunar _until recently_". My point was, in the meantime use xfce with nautilus(or whatever else) for the things that the current thunar can't do, then when thunar will have support for it, use it.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 10, 2011 10:13 UTC (Thu) by dgm (subscriber, #49227) [Link] (1 responses)

Maybe I had to be more explicit: I run Ubuntu in a VM with *only* 512 MB or RAM. Yes, half a gigagyte. Yes, this is at work, so 10 hours a day.

So, for me saving memory is a must. Running both Thunar and Nautilus at the same time is such a waste. I would run one or the other, but not both.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 10, 2011 21:33 UTC (Thu) by cmccabe (guest, #60281) [Link]

> Running both Thunar and Nautilus at the same time is such
> a waste. I would run one or the other, but not both

You can run Xfce without running thunar. It's just an application. If you don't start it, then it is not running. If you quit it, then it's not running any more.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 17:12 UTC (Wed) by sorpigal (subscriber, #36106) [Link]

Who cares about memory? Strike that, don't answer it. To those annoyed with GNOME and Ubuntu UI and UX decisions saving memory is not the impetus for using XFce. On this topic, therefore, lightweight < functional.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 11:05 UTC (Wed) by Seegras (guest, #20463) [Link] (1 responses)

That's why every Device and every Website looks like it's running Windows... Well maybe not.

There are _loads_ of different user interfaces, and the do _not_ need to look like Windows, nor do they. And they still are used by people accustomated to Windows.

So, no, a GUI does not need to look like Windows, and the time for creativity in desktop user interface design is far from over.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 19:58 UTC (Wed) by cmccabe (guest, #60281) [Link]

Read my comment a little more carefully. I am not talking about "every Device and every Website," I'm talking about the interface for desktop computers. You know, those large boxes you connect a mouse and keyboard to? The things we used to use before the pods and the pads came out?

Most of us still do use desktop computers, and will continue to do so in the future. And if we're being honest with ourselves, most of us do want our GUIs to look like Windows. In the same vein, most of us want to use the QWERTY keyboard layout, even though there are more efficient ones out there, and continue to speak the english language, even though a lot of words are spelled funny and parts of the language are ambiguous.

A lot of the creativity in user interface design has gone into the new mobile devices. The old WIMP (window, icon, menu, pointing device) style of interaction just doesn't work well on these devices for physical reasons, and people had no pre-existing expectations. Sadly, this window for creativity is closing too. People are already blasting Android because of some minor UI differences it has with the iPhone. It doesn't take long for people to get set in their ways.

"You can't win, you can't break even, you can't even quit the game"

Posted Mar 9, 2011 1:16 UTC (Wed) by rsidd (subscriber, #2582) [Link]

The rise of Meego? I seem to have missed that story.

As for Android, nobody is using it on normal computers, even netbooks. Google themselves are pushing Chrome OS instead.

The biggest problem of OSS desktops

Posted Mar 8, 2011 20:09 UTC (Tue) by alecs1 (guest, #46699) [Link] (6 responses)

Sweet Jesus...
Sometimes it feels like there's something really wrong with the world. This is like the OSS desktop has solved all it's hard problems, and we can come to easy ones.

Did all these things happen last night while I was asleep?!
-Ubuntu is now able to install on all random PC configurations, unlike yesterday when two friends with rather common configurations weren't able to finish installations, let alone seeing that "useless" button.
-3D acceleration works at 115% performance of the Windows countepart on all chipsets.
-KDevelop or any other IDE got to the point were you can debug any executable you attach to when you get bored, autocomplete works perfectly, no crashing or freezing, and it has all the configurations you dreamed of, not to mention how damn easy it's to write a plugin for it.
-KDEPIM stabilised and all features work flawlessly.
-GIMP made Photoshop obsolete, while nobody remembers Skype because an OSS IM replaced it.
-PowerBash kicks PowerShell ass.
-all the projects from this list http://www.fsf.org/campaigns/priority-projects/ are now a success.
-Xorg limitations, like the 255 keycodes are like they never existed, etc, etc.

The biggest problem of OSS desktops

Posted Mar 8, 2011 21:32 UTC (Tue) by zlynx (guest, #2285) [Link] (5 responses)

Complaining about all your favorite broken things that aren't being fixed has a serious logic problem: You're assuming that all developers are interchangeable.

They're not.

The Gnome GUI developers are useless debugging kernel DRM drivers for 3D cards. The kernel developers working on archaic IDE driver bugs could care less about KDevelop.

Should they all stop working and sit around doing nothing until your favorite bugs are fixed?

The biggest problem of OSS desktops

Posted Mar 8, 2011 21:50 UTC (Tue) by alecs1 (guest, #46699) [Link] (3 responses)

Don't take it literally, those are not my favourite bugs, just the important things I could think of in a few minutes, Apart from KDEPIM and KDevelop I don't really care about anything on my list. I'm just making a point about how much attention these changes get.

I'm pretty sure these Gnome or Ubuntu developers are not experts uniquely in panels, minimise and close buttons and usability.

Without any digging, I'm sure there are valid bug reports with both GNOME and Ubuntu which could be fixed with the same resources engaged, and would really make an advancement to the respective application, a bigger advancement than some so-called "usability" gimmick.

The biggest problem of OSS desktops

Posted Mar 9, 2011 0:01 UTC (Wed) by ovitters (guest, #27950) [Link] (2 responses)

People work on what they want to work on.

The biggest problem of OSS desktops

Posted Mar 9, 2011 0:49 UTC (Wed) by faramir (subscriber, #2327) [Link] (1 responses)

>People work on what they want to work on.

Sure they do. But when you work on what is essentially infrastructure for the entire GUI a certain amount of stability (humility?) is in order. This kind of change is (IMHO) the equivalent of deciding that the C programming language doesn't really need the for(;;) loop construct, because you can always do the equivalent with a while() loop. Sure a few programs (user UI habits) will have to change, but its all in the interest of simplification of the language (UI).

The biggest problem of OSS desktops

Posted Mar 9, 2011 22:09 UTC (Wed) by ovitters (guest, #27950) [Link]

I was responding to alecs1 notion where resources should be spent, not about what was actually done.

Arguing about what was done, ok. Arguing that efforts should go towards 'x'.. pointless IMO. For me personally, I work on whatever I want to work on. And often I give priority to whatever other contributors might need. However, if I worked on something, I would find it pretty odd to hear back 'effort should have been spent on X instead'.

The biggest problem of OSS desktops

Posted Mar 9, 2011 13:25 UTC (Wed) by pboddie (guest, #50784) [Link]

Complaining about all your favorite broken things that aren't being fixed has a serious logic problem: You're assuming that all developers are interchangeable. They're not.

But aren't we also talking about work done by a company who could quite possibly employ people to tackle each of these challenges? That it's not just a question of people volunteering to work on things they want/like/have time for? And where the company would see greater adoption of their products if the big challenges were dealt with?

Personally, I can see the point of having persistent, versioned documents (eliminating the "save" function), and having persistent tasks (eliminating the "quit" function, but also the "run" function), but the thing that probably scares a lot of people is having lots of tasks running and not being able to discard them, even if that means just not showing them. Being able to "reset" tasks is also a capability you'd want to have.

It occurs to me that the OLPC/Sugar people have considered some of these issues, although whether the Ubuntu people have bothered to look at that work remains unknown to me. When people think that they're innovating, they should always assume that someone else has done at least part of that work before, because that is almost always the case.

Thomas: That's it, we're quitting

Posted Mar 8, 2011 20:54 UTC (Tue) by xorbe (guest, #3165) [Link]

Man, Gnome and KDE are actively driving people away. Way too much change, just because.

Dear GNOME

Posted Mar 8, 2011 22:11 UTC (Tue) by kmself (guest, #11565) [Link] (2 responses)

That thing you see under your feet is the shark you're jumping.

Dear GNOME

Posted Mar 26, 2011 22:40 UTC (Sat) by mfedyk (guest, #55303) [Link] (1 responses)

So true.

This reminds me of the abortion that was "Spacial Nautilus". Every gnome based distro defaulted to it (until recently) except for Ubuntu. I count that as one of the reasons why Ubuntu became so popular...

Dear GNOME

Posted Mar 27, 2011 9:30 UTC (Sun) by paulj (subscriber, #341) [Link]

I have to say, I really liked the spatial part of nautilus. A really good use of connecting placement with content. I know some really unsophisticated users, who can struggle a bit with computers, who it seemed to be helpful to.

Makes sense, but not for Firefox please

Posted Mar 8, 2011 23:23 UTC (Tue) by mastro (guest, #72665) [Link] (4 responses)

Use case for the quit command in applications like Firefox:
  1. I have a lot of open windows, each with a lot of open tabs.
  2. There's a security update for Firefox.
  3. I close all the windows at once by quitting FX.
  4. I install the security update.
  5. I restart FX by clicking on the funny red and blue circle.
  6. Voilà, all my windows and tabs are restored to where they were.
  7. ???
  8. Profit!
But, yeah, in applications like evince, gedit, eog or LibreOffice an implicit quit when all the windows are closed may be good enough. I think evince and gedit have worked this way for many years and nobody complained.

Makes sense, but not for Firefox please

Posted Mar 10, 2011 5:30 UTC (Thu) by josh (subscriber, #17465) [Link] (1 responses)

Alternative: you upgrade Firefox while it runs, it notifies you when done upgrading that you should restart, you click the "restart now" button, all the Firefox windows go away, and come right back under the new version. No "Quit" needed.

Makes sense, but not for Firefox please

Posted Mar 10, 2011 16:44 UTC (Thu) by cdmiller (guest, #2813) [Link]

And by leaving Firefox running the security flaw, a remote access nasty, makes it look like your updates are applied but really aren't. Game over.

Makes sense, but not for Firefox please

Posted Mar 18, 2011 10:53 UTC (Fri) by cas (guest, #52554) [Link] (1 responses)

But, yeah, in applications like evince, gedit, eog or LibreOffice an implicit quit when all the windows are closed may be good enough. I think evince and gedit have worked this way for many years and nobody complained.

i've complained. I hate the way evince and epiphany require you to manually click close on every single window rather than just have a Quit button.

and i absolutely loathe and detest the way that evince won't even let me configure the default window size and settings (like Continuous scroll vs Paged) for documents, but instead uses heuristics based on my history of opening documents to automagically give me a window size that i don't want and the exact settings that I *always* unset. Why do they have to insist that they know what I want better than I do? Why can't they just give me a config option?

What makes problems like these even worse is that the response from Gnome about any bug report on any missing (or worse, deliberately deleted from the latest version) functionality is, as someone mentioned earlier in this thread "You don't exist". except that's the *short* version of the phrase - what they're really saying is "You don't exist, so fuck off"

here's direct quote in response to a bug report about the recent removal of gdm from debian in favour of gdm3:

I'm sorry if the changes in gdm3 don't please you, but you're not really the kind of user we're interested in anyway.

I'm beginning to suspect that Gnome and everyone associated with them has been taken over by a Microsoft conspiracy to deliberate alienate users and piss them off

(i'm only 99% joking about that last sentence. sad but true. nah, really, i think it's just the insidious effect of groupthink)

Makes sense, but not for Firefox please

Posted Mar 18, 2011 15:16 UTC (Fri) by bronson (subscriber, #4806) [Link]

I suspect it's due to the law of powers. When a feature appeals to 80% of users but alienates 20%, the gnome devs go for it ("you're not in our target audience"). Repeat this for 10 features (0.8^10) and suddenly gnome appeals to 11% of users.

Obviously an oversimplification that doesn't take into account similarity among users but you get the idea...

Thomas: That's it, we're quitting

Posted Mar 9, 2011 0:15 UTC (Wed) by xorbe (guest, #3165) [Link]

We should also get rid of the launch menu. The app wouldn't be installed if we didn't want it running in the background already, right?

Thomas: That's it, we're quitting

Posted Mar 9, 2011 4:27 UTC (Wed) by drag (guest, #31333) [Link] (20 responses)

I think that the majority of comments here didn't actually bother to read the article and have no idea what is being said on it or what they are talking about.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 8:47 UTC (Wed) by sce (subscriber, #65433) [Link] (19 responses)

I agree. I skimmed the article, and I must say it was pretty long considering it basically boiled down to "the quit menu item behaves differently across applications, so we'll remove it and make quit implicit after the last open window of an application closes."

I also find it interesting that people are so negative to changes in UI in general. For years the Linux community has been criticised for only following and not innovating, but when someone actually tries to innovate (Ubuntu with Unity and Gnome with Gnome Shell) it seems the criticism became even louder.

Personally I think the last Ubuntu releases were pretty boring compared to their earlier days, however I'm genuinely looking forward to Natty. I'm not entirely sure I'm going to like Unity, but I'll give it a wholehearted try. I'm also going to give Gnome 3 a try, so that I can pick the one I like the better.

In any case I'm open for change and innovation because I'm pretty sure we haven't found perfect yet.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 9:09 UTC (Wed) by dgm (subscriber, #49227) [Link] (6 responses)

> I also find it interesting that people are so negative to changes in UI in general.

This is not true. See the threads discussing the new scrollbar improvements the Ayatana team is making: http://www.markshuttleworth.com/archives/615

Funny that apparently only negative news about Canonical and Ubuntu seem reach LWN. Maybe a it's a sign of Fedora bias?

Thomas: That's it, we're quitting

Posted Mar 9, 2011 11:07 UTC (Wed) by zzxtty (guest, #45175) [Link] (4 responses)

I wouldn't put this down to ubuntu/fedora bias. That said I have switched from ubuntu to fedora, the reasons:

1. Moving the minimise/maximise/quit buttons from rhs to lhs. I still fail to see the point behind this move, on top of this it ruins my work flow, I stack windows and cant access the buttons any more.

2. The single menu bar, I've experienced this on macos and it drives my up the wall. I don't want to have to drag my mouse point across the whole desktop to get to a menu option.

I looked at the scroll bar changes you liked, again pointless. Most modern computers have wide screen monitors, horizontal pixel space is not in short supply. Granted it may be of use for tablet devices, but then one solution does not fit all, if you go down that route you end up with the worst of all possible designs.

We appear to be trying to make the linux desktop idiot proof. This is like taking a perfectly functional swiss army knife and removing all the options so that only the screwdriver is left. I'm using a swiss army knife because I want the functionality of a swiss army knife. If someone wants a screwdriver let them go buy a screwdriver.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 14:46 UTC (Wed) by szh (guest, #23558) [Link] (3 responses)

> Moving the minimise/maximise/quit buttons

Ubuntu 10.04:
# gconftool-2 --type string --set /apps/metacity/general/button_layout "menu:minimize,maximize,close"

Thomas: That's it, we're quitting

Posted Mar 10, 2011 20:46 UTC (Thu) by qubes (subscriber, #2562) [Link] (2 responses)

I haven't been able to find the configfoo to get the button layout I find most natural: close on the left, and all the rest on the right. I have to install KDE to get that option.

Thomas: That's it, we're quitting

Posted Mar 10, 2011 23:28 UTC (Thu) by jonasj (guest, #44344) [Link] (1 responses)

Doesn't the following work for you?

gconftool-2 --type string --set /apps/metacity/general/button_layout "close:minimize,maximize"

Thomas: That's it, we're quitting

Posted Mar 18, 2011 10:56 UTC (Fri) by cas (guest, #52554) [Link]

"apt-get install openbox" works for me.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 15:10 UTC (Wed) by foom (subscriber, #14868) [Link]

Those scrollbar changes look exactly like the ones Apple is making in 10.7. (first disclosed in Oct 2010). :p

Thomas: That's it, we're quitting

Posted Mar 9, 2011 14:06 UTC (Wed) by dskoll (subscriber, #1630) [Link] (4 responses)

I also find it interesting that people are so negative to changes in UI in general.

Why do you find it interesting? I find it perfectly natural. Most people get used to a certain way of doing things and really resent being forced to change. (Count me in that group, for sure!)

I don't care if GUI developers go on an acid-trip fooling around with the GUI as long as I can still configure it the way I like. The problem with GNOME and Ubuntu is that the developers seem to decree the One True Way a GUI should work and make it difficult or impossible to go back to the way you like.

People are very resistant to change. The QWERTY keyboard is still ubiquitous, even though it was originally designed to keep common letters far apart so mechanical typewriters would not stick. It's almost certainly a sub-optimal layout, but alternatives have failed to take hold.

Perhaps UI designers should learn to live with human nature rather than trying to decree it like some Communist central planning committee.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 18:24 UTC (Wed) by drag (guest, #31333) [Link] (3 responses)

> Perhaps UI designers should learn to live with human nature rather than trying to decree it like some Communist central planning committee.

A communist central planning committee is were you have people that are not doing any work are the ones that are making decisions on how the people doing the developing should do.

With this the people that are doing the work are making the decisions.

Despite everything it's still open source. They are just trying to boil the UI down to it's required features. They are minimalists.

With things like 'Elementary' versions of Nautilus and other features that third parties are working on it is proof that Gnome isn't forcing anybody to do anything in any particular way.

Want a different looking nautilus? Want to embed your shell into Nautilus in a better way? Just use PPA and install it. That's just a example, but there are mods and all sorts of stuff. Once Gnome 3 solidifies and people get used to it there will be even more stuff floating around. It takes a few minutes and even a Linux novice can handle it if they have good directions.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 20:02 UTC (Wed) by dskoll (subscriber, #1630) [Link] (2 responses)

With this the people that are doing the work are making the decisions.

Yes, you hear this a lot in relation to free software projects and it drives me crazy. The attitude is "We don't need to listen to our customers" or even "We don't have customers because it's open-source, it scratched my itch, and anyone who doesn't like it can fork it or shut up."

I run a software business, producing both free and non-free (ie, traditional proprietary) software. In both cases, I have to listen to my customers to stay in business or stay relevant. If free software developers choose not to listen to their customers, then they put the lie to the oft-stated belief that open-source development is a superior methodology to proprietary development.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 23:11 UTC (Wed) by nybble41 (subscriber, #55106) [Link] (1 responses)

Free software developers *do* listen to their customers. How could it be otherwise? They *are* the customers.

Others may use the software they develop, but are they, in fact, contributing to the project in any way? It's hard to consider someone a "customer" who isn't actively contributing to the development process, financially or otherwise. The term "customer" implies a two-way relationship; passive users need not apply.

The accepted way to contribute to most free / open source software projects is to either work on the project yourself or pay someone else to do so for you, at which point you *are* a free software developer, as well as a customer, and you can "listen to" your own specific needs as much as you wish. Moreover, others are more likely to respond to your suggestions if you show signs of actively contributing to aspects of the project they happen to be interested in.

Thomas: That's it, we're quitting

Posted Mar 10, 2011 1:08 UTC (Thu) by dskoll (subscriber, #1630) [Link]

I agree with what you say. I've contributed financially to a number of free software projects. I've sent patches to a number of others and I lead a few others myself.

Without a doubt, I've found the GNOME developers to be the most hostile (of those Free Software developers with whom I've interacted) to suggestions and the most certain that they know they're right. For this reason, I am no longer a GNOME "customer", so I'm mostly watching this one from the sidelines.

I still think it's a shame, though. These kinds of shenanigans can turn a viable desktop environment into a train wreck.

Thomas: That's it, we're quitting

Posted Mar 9, 2011 17:23 UTC (Wed) by sorpigal (subscriber, #36106) [Link] (6 responses)

Innovation is fine. I love experimentation and trying new things. What I don't love is when new, radical ideas get included *instead of* known and used approaches rather than being introduced as optional "enable if you want it" features. If some of these ideas were made available to those who want to try them for a couple *stable releases* as toggleable options in a user-discoverable GUI, then when the decision was made to make them the default there would be some basis for that.

What we have smells like "force it down their throats" rather than "the power is yours." Give *me* the option to try new things, get it widely tested, get feedback, and *then* switch to it if it seems like a good idea. That's the way innovation should work. You put out something new and its usage grows grassroots-style based on its merits.

Of course my stance is the antithesis to GNOME's fundamental principles since I think that, in general, more options are better.

Thomas: That's it, we're quitting

Posted Mar 10, 2011 13:16 UTC (Thu) by bluss (guest, #47454) [Link] (5 responses)

'Everything must be user setting' is a huge burden and many developers rightfully refuse it.

It sounds like you should use Debian (I do).

Thomas: That's it, we're quitting

Posted Mar 10, 2011 15:14 UTC (Thu) by dskoll (subscriber, #1630) [Link] (3 responses)

'Everything must be user setting' is a huge burden and many developers rightfully refuse it.

Maybe so, but I think it's reasonable to ask: "Every feature that used to work but is disabled should be possible to re-enable."

It's called backward compatibility and is frankly one of the weaker areas of the Free Software development model.

Thomas: That's it, we're quitting

Posted Mar 10, 2011 21:39 UTC (Thu) by nix (subscriber, #2304) [Link] (2 responses)

Some parts of the model get it right. Emacs, if anything, goes too far, deprecating keybindings in 1995 and finally removing them in 2007. People still complained, even though if they wanted to fix it, it was one line of code.

Thomas: That's it, we're quitting

Posted Mar 11, 2011 3:25 UTC (Fri) by jonabbey (guest, #2736) [Link] (1 responses)

Ah! Is that what happened to my beloved Meta-G?

Thomas: That's it, we're quitting

Posted Mar 11, 2011 20:38 UTC (Fri) by nix (subscriber, #2304) [Link]

No, I don't believe that was ever deprecated. RMS just chose to rebind it one sunny day in 2005 :)

Thomas: That's it, we're quitting

Posted Mar 11, 2011 14:23 UTC (Fri) by sorpigal (subscriber, #36106) [Link]

Not coincidentally I do use Debian (gosh, it's been a decade now!)


Copyright © 2011, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds