The Grumpy Editor's GNOME 3 experience [LWN.net]

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 1:26 UTC (Wed) by sramkrishna (subscriber, #72628)
In reply to: The Grumpy Editor's GNOME 3 experience by jcm
Parent article: The Grumpy Editor's GNOME 3 experience

Of course! Being rude would accomplish nothing other than make things worse. Besides, as I said you guys are all being nice to me so I'm happy to discuss these things.

OK - so I think we're coming down to some of the core issues here. I'm pretty sure what you've written down is shared by a lot of people.

Regaridng GConf (or DConf) - that was given a bad rap and we didn't engage with anybody to fully explain that bit. Let's put it this way, if you wanted a persistent key value store that would allow apps to react when the values change, how would you do it? In this case, having a name space, and then a simple key/value with a set of types is really all GConf is and a server to maintain state. It's NOTHING like the windows registry. I think in DConf they've tightened things up a bit in terms of namespace. It is the name space portion that could be abused. GConf was damned by it's superficial likeness to the windows registry. You will note that nobody is talking about GConf in forums all that often. I think someone just writes some kind of gui tweak tool to do all that stuff. Looks at the gconf today with gconf-editor, can you claim it is a mess as the windows registry is? We've had it for ten years now.

For all the negatives that came out of the 1.x to 2.x migration it wasn't so bad.. and we were cutting options left, right and center in those days. Today, a lot of you appreciated.. some of you claimed it is perfect. :-)

You might also consider that our drive for "Just works" has improved the linux eco system tremendously with hal, udev, improved x drivers, wayland... there wouldn't be any drive for those projects without a project like ours demanding those features.

As for trendy, people are redesigning UI because there are new devices out there. Tablets, smart phones, medical devices, gps stuff.. everything is all getting connected to the Internet.. so projects like firefox, chrome are all out there trying to capture those markets. New types of embedded devices are coming out. We are winning there. Linux is _the_ embedded OS of choice. It is my blu-ray player, my television.. and it continues. It's trendy because there is a trend. :-) If we want free software to make inroads on these devices then we're going to need projects like GNOME, Meego, KDE to be able to adapt and be able to get into those areas. So we need to re-envision what the user interactions would be, and we need to push the rest of the Linux eco system to support it. Technology doesn't stay still it marches on, baby.

Users will be willing to lean a new interface if a user gets something out of it. GNOME 2 isn't going away immediately, there will be a transition state where users will be able to get their feet wet and their pace. In the mean time, users will go to KDE4, XFCE, FVWM2 and in that time GNOME will continue to enhance and polish itself. People by that time might come back see that GNOME has something to offer that would give them a reason to learn the new interface. It's on us to meet the challenge entrancing you back to our platform. If we fail.. well we failed, but it isn't for lack of trying. But I'll tell you this, I spent two weeks dogfooding GNOME 3 and I can't go back to GNOME 2. It surprised me because there was a lot of things I loved about GNOME 2, specifically my beloved gtop applet.

I installed NetBSD on my Amiga.. (when I wasn't running AT&T SysV on my Amiga) I've been doing some kind of building and testing for the past 14 years. I would spend hours tweaking my .fvwmrc2 file juuuuust right. Then change it again next week.

What you want and what GNOME envisioned 10 years ago (before it was cool to think having a consistent user interface with minimal tweaking ) and here we are. Maybe you don't want to change, but in order to continue to be relevant we have to.

Skip the first release, come back for 3.2, by then we should have a couple of new extensions, fixed a lot of bugs, and changed UIs around to accommodate the feedback we had in the first iteration. I'm saying all this in order to leave the door open so that those of you disappointed in the changes might relent and come back after we've added some features, or put back features whatever you prefer.

I will note that we didn't re-implement anything but the user experience. But our development platform is exactly the same for the most part. In contrast, KDE4 completely re-did their internal plumbing and maybe it required it. But for those with large projects re-implementing their application was probably a pain they would like to fore go. It took the GNUCash guys 7-8 years to port to GTK2. Very painful. We have too many people depending on our codebase to be doing anything like that again.

Don't get too comfortable with that Mac, the next iteration will use IOS. :-) Which of course follows the same trend we are!

From all you've described, you seem the prime kind of person we want as a user for GNOME 3. If we offered you a stable UI and you got used to the interface and added your launchers, (and fonts for you Jonathan Corbet) I don't see why you couldn't stick with GNOME. :-)

to post comments

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 1:45 UTC (Wed) by jcm (subscriber, #18262) [Link] (15 responses)

On the d/gconf front, I understand why it's not "just text files" (though these days we have all kinds of file monitoring and watching APIs that are getting a lot better for just using text files) but I am very uncomfortable with the idea of hiding the real guts such that one needs to edit these settings directly. If they really were just out of the way, and had a nice UI for changing them, ok. But it really does seem like things are heading into the Microsoft realm of "kinda sorta works" out of the box, with all kinds of obscure knobs you can only tweak if you look up the right schema.

I wasn't a huge fan of the 1.0->2.0 migration either, but it didn't churn my stomach as much as this transition has. I'll agree that having shiny things has advanced the Linux ecosystem (side note: I'm one of those "you can tear my real X server out of my cold dead hands" types when it comes to replacing it with something that is theoretically better but loses the UNIX networking heritage), and that some of these advances were necessary. But I think many of them would have happened anyway. Take automounting, hotplug, whatever, these are all things people like to have working and pretty much independently of the desktop environment they are using.

On the UI trendiness, you didn't really answer my point about retaining relevance to existing platforms :) I'm all for having a tablet or smartphone UI, but in my personal opinion it's simply inappropriate trying to run the same interface on both classes of device. Many others have learned that this is the case through bitter experience and failure in the marketplace, and I think GNOME 3 is going to try to offer what many other projects are already doing (tablet/embedded UIs) while not really catering to those of us who want a traditional desktop. As others said, 2.x is stable and mature. At that point, why not just sustain it, and have those interested in gnome-shell like UIs go and work on a dedicated tablet/embedded UI instead? That surely would have been better IMO.

Why does GNOME have to change to be "relevant"? Others have asked this elsewhere in these comments. This is a *fundamental* issue, and one where I think we very likely strongly disagree. I consider the current GNOME 2.x to be pretty "perfect" as a daily use system. It has some warts, there are things I would change, but I've been logged in for 100 days in this current session and aside from gvfsd doing its usual not understanding network routes changing and needing a kick, everything has been fine. It sounds to me like "relevance" is code for needing to have something for people to work on, or an answer for new kinds of devices. These are worthy goals, but they could (IMO) be better served as sub-projects, optional UIs, and the like, while retaining a great experience that works well. After 10 years, we can finally say Linux has a compelling and great UI, and now is not the time to be completely replacing it with something else :)

I *was* the prime kind of person for GNOME. But "new GNOME" is like "new Coke" to me. It looks great, passes all the tests, but the real proof will be when it's released next month and many times as many users start to have the same kind of reaction we're talking about here.

Thanks,

Jon.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 2:12 UTC (Wed) by sramkrishna (subscriber, #72628) [Link] (12 responses)

Hi Jon,

The whole GNOME ui everywhere on all devices is my own perception and doesn't reflect what GNOME itself is trying to do. They might just simply focus on the desktop experience. I think big. It's a bad habit. :-)

The separation of the UI from the desktop API itself is a feature, that means we can come up with our own UI for any particular device. What's important is the underlying layers of GTK+/Cairo/Linux as the platform. Separating the UI and allowing us to use javascript to create and control the UI experience gives us the flexibilty of being able to create other UI for other devices.

As for your question why is change relevant. Change is always relevant. Why do you need so many different kinds of filesystems? Isn't ext4 good enough? If we just accept status quo how does anything get done? Why do we need a car? Isn't a horse and buggy good enough? Every time you move to the next level you move to a different set of realities that give you new avenues to pursue. It's in our nature to always pursue what somebody will always consider "flights of fancy". But it is the engine that moves us forward.

All people crave change, but not all crave change in the same area. Desktop people want to push the limits of user interaction. We have no other purpose. No desktop project is going to accept status quo for long. No kernel developer is going to accept using the same scheduler for long. There is always some new hardware that drives making a better scheduler to take advantage of it.

Maybe we've lost you for now.. but guaranteed, those other projects are doing the same thing we are. Maybe some want to give you even greater control of your desktop, give you lots of options, others might want to take away some, put others.. we fiddle with things, that what we do. You fiddle in the areas you're interested in.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 11:41 UTC (Wed) by nix (subscriber, #2304) [Link] (7 responses)

All people crave change

Here's where you should start rethinking. Most people crave stability and dislike change. They may not loathe it, but they don't like it. Change for the sake of change is a bad thing.

(Look at the core Unix tools: fundamentally unchanged other than new features and the removal of restrictions since the 1970s, even though they are definitely not perfect. Why? Inertia. People dislike change.)

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 15:23 UTC (Wed) by me@jasonclinton.com (subscriber, #52701) [Link] (4 responses)

That's not the reason that GNOME 3 was launched. You can read what the reasons were here: https://live.gnome.org/ThreePointZero/DesignHistory

* Finding windows was frustrating and difficult
* Workspaces were useful but not easy or natural to use
* Launching applications was labour-intensive and error-prone
* The panel suffered from over-configurability; applets were little used by most users

In short, we saw real problems which needed real solutions.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 16:25 UTC (Wed) by nix (subscriber, #2304) [Link] (3 responses)

It sounds like the amount of mouse movement involved now has made at least two and possibly three of these cases *worse* than they were. (I wouldn't know: I'm a KDE and fvwm2 man. I used to be a GNOMEr but the loss of funtionality in the early GNOME 2 days turned me off for good. Software that won't change to act as I want it, rather than vice versa, gets dumped at once.)

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 16:27 UTC (Wed) by me@jasonclinton.com (subscriber, #52701) [Link] (2 responses)

> I wouldn't know

Then stop commenting?

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 17:44 UTC (Wed) by nix (subscriber, #2304) [Link]

My apologies for daring to point out how I as a human being use the machine, and why GNOME 3 would apparently not suffice. If ex-users are considered to be lost causes, so be it.

The Grumpy Editor's GNOME 3 experience

Posted Mar 24, 2011 14:16 UTC (Thu) by nye (guest, #51576) [Link]

>Then stop commenting?

You have yet to make a single comment on this article that isn't outright trolling, but you seem to feel compelled to randomly pop up in every subthread to hurl some abuse.

Commenters like yourself are the reason that I consider LWN to be a largely hostile, poisonous community, and consequently make less of an attempt than I should to behave in a way that I might in a forum I respect.

The one potential positive outcome of your posting is that I resolve to try harder to act as if LWN is more like the place I want it to be, on the grounds that if everyone did that, perhaps it would be.

If only it were possible to killfile entire categories of articles as flamebait, the overall experience would probably be much improved.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 20:38 UTC (Wed) by jcm (subscriber, #18262) [Link] (1 responses)

Exactly!

We have (weirdly) in recent times gotten to a place where there is a big push (I don't mean to single out GNOME here, because it's general) to undermine our UNIX heritage, and throw it all away in the name of shininess. Now, that push may not be malicious, and I'm sure everyone is well intentioned, but the newer generation aren't necessarily respecting what came before, or the value of long-term, stable, backward compatibility.

Why do Microsoft have a strangle hold? Sure, lots of reasons that are not entirely kosher, but they also have win32 (and later stuff) that people have been able to target for about a billion years in writing to a standard platform. It is *high time* we re-embraced our POSIX, SUS, LSB, and other standard routes, had *one* well supported, ingrained and established GUI platform, and could be that shining Linux alternative. It might not be perfect, but by not re-inventing the wheel every 5 years we will finally give others a chance to really target our platform, then we can improve it slowly, evolving as dictated by real world use.

This is what I was alluding at in my other comment(s). I used to care about building up random platform bits for the sake of it, but after 15 years I'm growing tired of the same old stuff. I want a computer that just works and I can take for granted the platform bit. There's been enough time. Sure, I no longer have to run isapnpdump or APS Magicfilter, but in the last few years I've had to fiddle with (not mentioning names) let's just say "way too many" wheel re-inventions of core Linux technologies. All of these are interesting projects, but I would rather we had stuck with what we had before and new problems had been solved instead of doing the same old thing that was being done 10, 15, or 30 years ago.

In a few years, the current generation of folks trying to change everything will be 5 or 10 years older, get to the same realization, and the entire story will repeat itself. I suppose the important thing is not to get too worked up about it, which is why I also run a Mac now (not because I think it's any better, but because it's "good enough" for surfing the web and doing desktop non-server stuff at home). I might have just chuckled to myself and ignored this story, but it did work me up slightly because of the amount of time I've had to waste especially last weekend undoing all of these changes to get back to where I was.

Jon.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 20:43 UTC (Wed) by jcm (subscriber, #18262) [Link]

I may also need to get a rocking chair, and a porch and start muttering, I'll accept that :)

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 13:53 UTC (Wed) by dskoll (subscriber, #1630) [Link]

All people crave change

Really? Could you back up that assertion with a little bit of evidence? My (admittedly anecdotal) experience is that most people hate change when it comes to major changes to the way their computer works. This is based on observations of my kids, parents, siblings and co-workers.

My kids use KDE and found the KDE 4 transition extremely jarring; it took them a couple of weeks to get back to their normal level of productivity.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 22:34 UTC (Wed) by dmadsen (guest, #14859) [Link] (2 responses)

There's a difference between forcing change and enabling change. Using filesystems as an example as you did is very apropos -- if I want to use an old existing filesystem type, I can, and there are tools to help me migrate when/if I want to.

The changes happen *at my pace and my option*, not anyone else's.

I'm an old fossil. I don't want change. I want my hands/fingers to know what to do on my desktop so that I am free to think about the actual task I'm trying to accomplish. A changing desktop is a distraction, an annoyance, a hindrance to productivity.

And when I have a few minutes and want to try out a specific new feature, I should easily be able to. And if I like it, then I'll use it. Someday maybe I'll use the entire feature set. Whoo-hoo!

Someone may love desktops and it may be their whole life. But it's a tool for me. Don't ever confuse yourself with your user base. And God forbid, don't think that everyone likes Macs or weird cellphone menus or adding lots of mileage on their mouse.

So sure, add all the new stuff you want; rewrite anything you want; have a great time. But when you do it, *don't* mess with what I have!

---dcm

The Grumpy Editor's GNOME 3 experience

Posted Mar 17, 2011 16:46 UTC (Thu) by Frej (guest, #4165) [Link] (1 responses)

I doubt anyone but you can change what you have.

The Grumpy Editor's GNOME 3 experience

Posted Mar 17, 2011 21:39 UTC (Thu) by mpr22 (subscriber, #60784) [Link]

Indeed - but pinning your DE at the previous major version, which is what's necessary to keep it around when you find the new major version utterly unacceptable - tends to make your package manager very, very unhappy very, very quickly.

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 2:20 UTC (Wed) by foom (subscriber, #14868) [Link] (1 responses)

> On the d/gconf front, I understand why it's not "just text files" (though these days we have all kinds of file monitoring and watching APIs that are getting a lot better for just using text files)

Except that KDE's kconfig has done exactly that for something like 10 years now: configs stored as plain text ".ini" files, and an efficient binary cache that is regenerated if you change the config files. It seemed to work fine for them...

The Grumpy Editor's GNOME 3 experience

Posted Mar 16, 2011 9:14 UTC (Wed) by mjthayer (guest, #39183) [Link]

> Except that KDE's kconfig has done exactly that for something like 10 years now: configs stored as plain text ".ini" files, and an efficient binary cache that is regenerated if you change the config files. It seemed to work fine for them...

I seem to recall that one of the dconf selling points was that the storage backend was pluggable. Perhaps someone could write a backend that does the same thing as KDE, or even uses kconfig...

The Grumpy Editor's GNOME 3 experience

Posted Mar 24, 2011 14:04 UTC (Thu) by nye (guest, #51576) [Link]

>In this case, having a name space, and then a simple key/value with a set of types is really all GConf is and a server to maintain state. It's NOTHING like the windows registry.

Could you elaborate on what you mean by this? That seems like an accurate description of the Windows registry so I don't get why you believe it's nothing like it.

I think what people hate about the registry is the hiding of options behind an arcane interface, the proliferation of entries with no idea of what they do, the non-textual storage mechanism, and the fact that random applications add new entries and often don't bother to remove them. However almost anything conceivably changeable is documented by Microsoft, typically at least as well if not better than most GConf keys, and all of the other problems still seem to apply to GConf, except perhaps the last, so what makes GConf so different?

The Grumpy Editor's GNOME 3 experience

Posted Apr 8, 2011 7:40 UTC (Fri) by tjavailable (guest, #74182) [Link]

Hi friends its happy for me to announce that Gnome 3 has been released. These are the steps to install Gnome 3: 1.Grab the ISO image based on Open SUSE or Fedora: 1.Pros: No need to replace your Gnome 2.x in Ubuntu 10.10 2.Cons: Need to install another distro that stays away from your Ubuntu 10.10 installation. 3.You can get the ISO image from here. Then place it on a CD or your USB stick and enjoy the Gnome flavour! 2.PPA for Ubuntu 11.04 1.Pros: Enables to try Gnome 3 from your own installed Ubuntu or in top of your Gnome 2.x 2. Cons: Your current Gnome 2.x might get bloated and you will feel why I'm using this as my default one since Ubuntu is shipping Unity from 11.04! For more check Here