Hacker News new | past | comments | ask | show | jobs | submit login
Introduction to Gnome 3 (gnome3.org)
139 points by audidude on Jan 18, 2011 | hide | past | favorite | 72 comments



Not that I want to beat up on a open source project, but for a desktop environment focused on ease of use, there are some very odd things overlooked.

1. How am I supposed to know what these icons do? Why even title them? https://img.skitch.com/20110118-biweei4eyuey217fisi7ks2d9f.p...

2. Eight of the eleven icons in the dock have significant title clipping. Did no one notice this? https://img.skitch.com/20110118-e1ykw7u3qf3gwmuhxaug28p5mr.p...

Usability isn't minimalism. It isn't translucent windows, rounded corners, or stark blacks and gradients. It's these details, and about a hundred others that I'll run into while using your software. These details can either add friction to whatever problem I'm trying to solve (and therefore add marginally to my bad mood), or they can remove it. If you don't ask these questions, you aren't even giving your users a chance.


People like to beat up on MS a lot, but the differences between the Windows 7 GUI (or iOS GUI) and Gnome 3 really highlight the disadvantage that a bunch of spare-time developers who say to themselves, "well, I think our users will like text truncation, let's throw it in tomorrow!" are at versus a big company with millions to spend on thoughtful UI design, graphics, research, and focus groups.

Windows 7 has its problems too, but it's clear to me that the changes Gnome 3 is bringing are nothing more than what some out-of-touch developers think is what users want, without having actually asked users. Maybe I'm wrong on that assertion but that sure is the feeling I'm getting from this project.

Gnome 3 would have benefited greatly from a series of focus groups before development started or from snagging a high-profile UI expert to rap them on the knuckles whenever they had a bright idea like using 12 different font colors, sizes, and styles on the already-overwhelming dashboard screen.


> Gnome 3 really highlight the disadvantage that a bunch of spare-time developers

You do realize you are criticizing the topmost user interface details of a theme of an extensively themable early release. I think those "spare-time developers" have been focusing their attention on the API changes that the Gnome 2 to 3 transition impose and the new concepts introduced by the shell. The core seems very good. What you complain about is usually addressed in the configuration files for the theme and the names of the shortcuts. Multi-lingual support will bring some additional problems that should be addressed in the coming weeks. There are many dozen different themes that are compatible with Gnome. All those will have to be tweaked to look right.

Have a little more faith, for just about every computer that is responsible for your internet access runs on top of software written by people you would call "part-time programmers".

I'd trust their work over the one from Microsoft's full-time employees anytime.


If they didn't want to invite criticism, they shouldn't have release screenshots of a product they don't want criticism of, complete with breathless captions about "beauty" and "usability."

Furthermore, I have better things to do with my time than spend time mucking around with configuration files and theme settings. Back in high school I had the time and inclination, and in fact I did so; now that I run a business, I'd rather spend that time doing actual development. Not to mention the moms and grandmas of the world who don't even know what a theme is. I, and they, need a UI that just works out of the box.

And finally, just about every computer that is responsible for my internet access runs on top of software written by large corporations who have hired people to develop for open source bodies. Let's not forget that it's for-profit companies with full-time programmers like Red Hat, Canonical, IBM, Intel, etc. that are more than partly responsible for the relatively wide enterprise (and desktop!) adoption of Linux today.

I'm not out to bash Gnome, Linux, or open source; I use Ubuntu on my work laptop every day. I criticize out of the desire to see the open source movement develop something better, and that means not patting everyone on the back just for "working hard." It's precisely that kind of "gold star for effort" attitude that lets crappy software exist.


> If they didn't want to invite criticism, they shouldn't have release screenshots of a product they don't want criticism of.

I am sure this will be a learning experience. It's alpha software, after all, and it is released in order to allow people to learn what works and what doesn't in the real world. As usual, with open-source, if you decide to use Gnome Shell right now, or in April, you'll be running the bleeding edge. I would say that, if you want to run it as stable software, you should wait until October. Open-source is developed in the open, as it invites intelligent criticism and grows stronger with it.

> Furthermore, I have better things to do with my time than spend time mucking around with configuration files and theme settings

Unless you want to design themes, you are not supposed to do that. I rarely do. I think I haven't done that since 2003 or so, and I was really going for a full customized desktop at the time. I did it because I wanted to do it.

> And finally, just about every computer that is responsible for my internet access runs on top of software written by large corporations who have hired people to develop for open source bodies.

I would like to point out a lot of Gnome has also been developed by a lot of full-time employees of for-profit companies too. It was you who called them "spare-time developers". I am not sure how it will play out with Novell's demise, but I trust other companies, like Red Hat and Canonical can chip in if needed.

> I'm not out to bash Gnome, Linux, or open source

I then certainly misread your post. I didn't realize you meant "spare-time developers" as a compliment. My bad. I am sure open-source has a lot to gain from the criticism of a bunch of spare-time experts.

Meant as a compliment, of course.


http://www.linuxpromagazine.com/Online/News/Red-Hat-the-Top-...

"The vast majority (70 percent) of contributors work on the project on their spare time, while an additional 20 percent of contributors do so on both a paid and voluntary basis."

Adding the percentages of the top 5 corporate contributors adds up to just over a third of contributions from ostensibly full-time developers.

So it appears to be just a fact that people are developing this in their spare time. That doesn't mean they're less competent than paid developers; of course not. It simply means that compared to a big company, like MS or Apple, they just don't have the same level of cash money resources to hire the right kinds of people--UI design experts--and buy the right kinds of feedback--real-life user testing, focus groups--that big companies do. And that very much shows in these screenshots.

Alpha or not, criticism is essential for furthering the project. If everyone just sat around and said "whew, lots of hard work, good job guys!" to whatever crap any OSS dev produced, OSS would never progress. I want Gnome 3 to be better because I use Gnome every day!


Ahem, "The study found that some 70 percent of contributors are unpaid, but that the majority of commits comes from paid participants." I'm not sure what basis you used to decide that only developers employed at the top five companies count as paid developers.


Have you tried to, instead of calling them "spare-time developers", file a bug report or a feature request? Did you get involved in any design decision? Do you subscribe to whatever list Gnome Shell developers discuss their ideas?


Canonical has and pays UI design experts. I don't know if they're the best, but I remember going to a conference given by one of them a few years ago at the FOSDEM.


  > If they didn't want to invite criticism,
If the only outlet for your criticism is Hacker News, then I would argue that it's not very constructive criticism. You're saying that criticism is what will make the product better, but that only works if your audience can directly affect the outcome (e.g. GNOME developers/contributors).


I think there's no substitute for clued-in developers.

Focus groups produced the interface overhaul to Windows Vista and 7. Rather than building on an existing, successful interface, they decided to be more mac-like, and managed to make the interface prettier but much yet less usable in the process.

I agree, gnome fonts are yuck. Even in the screenshot on the frontpage the fonts look awful - both in the text document and the dropdown menu headings.


The problem with the fonts is more than just aesthetics. It's about gamma correction.

For various reasons, the brightness of an RGB color value is not linear. This means that a checkerboard of white/black pixels will not look the same from afar as solid 50% gray. The correct shade is something like ~75%.

When anti-aliasing crisp details such as letters, this is incredibly important. Without gamma correction, there will be a noticeable change in brightness when a shape lies between pixels vs when it is pixel-aligned.

I think this is why Linux geeks tend to be vehemently opposed to anti-aliasing; they don't realize the rendering method they're using is crap.


> but much yet less usable in the process

Based on what metrics and evidence? I think it does a good job making me more efficient.


One interesting thing about the Gnome Shell design process is that it has been very open and responsive. Although there was a heavy focus on creating design principles up front, the actual design has evolved a lot over the few years that the project has been active. The overview of the current design at http://live.gnome.org/GnomeShell/Design has many relevant links that you can explore if you're interested.

I know there are plans for usability testing, but I don't know how they have progressed since this blog post http://mairin.wordpress.com/2010/02/10/gnome-shell-usability...


Interesting, I didn't know about that.

One thing that concerns me is that they only decided to do usability testing in Feb 2010, but according to the Wiki page Gnome 3 has been in development since Jul 2008.

Shouldn't a huge UI overhaul receive incremental user testing to stamp out bad ideas before significant dev time is invested? After an alpha is delivered, does anyone really think that if one of their big paradigm changes is found to be disliked, that they will go in and remove or significantly change it? Maybe Gnome would, I don't know--but from my experience, once code is delivered in a semi-functioning state, it's very hard to overcome that inertia to make huge directional changes. (And perhaps no big change is needed--but from these screenshots I can say that so far I personally don't like the new direction.)


A significant amount of Gnome 3 development time has been spent on things which will be mostly invisible to the user; overhauling libraries and such.

It looks like Gnome Shell has been in development since late 2008, but only got to a state where it was very usable in late 2009. http://gnomejournal.org/article/85/easy-breezy-beautiful-gno...


I think what's most unfortunate is that Linux desktop people follow MS' approach of integrating the windows managers and the graphic shell ever more closely.

This approach seems to requires a monolithic application.

I would imagine it could be more open-source-appropriate to separate out the Windows manager (in charge of the details of font and icons) and the shell (the taskbar, start-menu, explorer-type-application etc).

Caveat: I'm working on a shell-like project myself. I suppose I might feel some satisfaction to see Gnome get it wrong but I don't want to see a monolithic system that freezes out alternatives.


Is that a valid criticism though? I don't run a desktop, just a window manager (scrotwm), and I can still run KDE apps without having to load the entire KDE desktop environment. (Gnome I don't know about, as Slackware doesn't provide Gnome.)

Also, does 'shell' mean something specific to desktop people? What you detail as being the shell, I would call 'desktop'. But I could very well be out of touch with the current lingo.


I'd like to point out that the images are marked as being of an alpha version. These complaints are fair, but they can be forgiven for not having all the details under control at this point. I personally am reserving judgement until I have an actual, finished product to test out.


I think the clipping might be fixed. I don't see any in the dash (GNOME's term for the dock) or the application picker in the screenshot at http://live.gnome.org/GnomeShell/Design?action=AttachFile...


It does say "alpha version".

The concepts look good, lets hope they sort out these details.


I'm amazed by how dumb and severe the icon label truncation is. That screams to me that it was designed without testing it with real content.

How can anyone see Network C…, Network P…, Network P… in the preference panel and not wonder if the truncation is too agressive?


I used to use dillo as my browser of choice. It truncated tabs by dropping vowels, which I thought was really clever. I have no idea how well that would work in other languages, though.


For those complaining about the details: I think this represents the code and general work on a new interface that is coming out. The applications that it is working with are the utilities that have been included in Gnome for quite a while and the names worked fine with the old menu system; clearly the work being demonstrated is on the interface level. The current interface only takes the default application names it is being fed. I´m sure a finished ¨product¨ will take into account these details, but I don´t find it unreasonable that it looks how it does in an Alpha product. Simply changing what the name appears as in the shell isn´t going to be the long term solution.


Truncation like that also happens on the Mac when the Finder's grid spacing is set too tight. They may have overlooked it for the screenshot.


I'm not criticizing truncation. You need to do it on long names. I'm saying it looks like the designers of the GNOME UI are using truncation as a quick fix for a flawed design.

Besides, truncation in OS X is much much smarter:

http://i.imgur.com/MN3Q2.png

Notice how the tracking (spacing between letters) in the first item gets tightened to make it fit? It looks a little squished, but it's much preferable to reading dots.

The last item had to be truncated, but it happens in the middle and keeps the useful end information visible.

I was being honest when I said I was amazed at the GNOME truncation. I think an operating system UI that can't handle the names of its killer apps (e.g., OpenOffic…) lost its way somewhere.


Also, if you do this on the desktop or in the finder with grid/icon view, OSX will word-wrap for a couple notches before trying to "middle-truncate". Loss of information is their last resort.


> I'm not criticizing truncation. You need to do it on long names. I'm saying it looks like the designers of the GNOME UI are using truncation as a quick fix for a flawed design.

Yeah, I thought it looked pretty obvious because it was so bad it just hadn't been gotten to.

While I can't articulate on whether pango's ellipsizing is as smart as Apple's, it should be as simple as changing a single property from PANGO_ELLIPSIZE_END to PANGO_ELLIPSIZE_MIDDLE.


> Besides, truncation in OS X is much much smarter:

Don't forget Apple (among others) has tons of silly patents regarding text rendering that Gnome developers have to program around while everybody else can just cross-license with their own silly little patents. It may be possible they simply cannot afford the risk of doing the obvious thing.


Could you clarify this?

Do you actually know that Gnome is actively "code around" some alleged patent violation?

Programming around any imaginable patent violation actually seems really bad, especially given that software patents aren't looking as strong as they once were. I vaguely Linus or someone saying it's better not to research existing patents, it limits your liability. But just on the principle of letting corporations covertly bully you, it seems bad. It seems much better to force the companies sue you and see what happens.

My googling shows Apple has a patent on the App-bar but there are plenty of App-bars out-there. Red-hat apparently was worried to remove a dock but the screen shots we see here clearly show something like a dock. But even thread discussing that situation sounded murky. Docky still seems to be distributed for example.

Apple has patent on "Open Type" but that also is used heavily in Linux.



OK...

Well, if this is wide-spread, I find frustrating, even infuriating that it only appears in bits and pieces rather than there being a large "this is what software patents are imposing on us". The worst possible thing is to let this happen silently.



I've heard of the free-type patents and I found references to the patenting of OS dock (hasn't end free docks in Linux as far as I can see - especially, the Gnome Shell look a lot like a dock).

I oppose that Free-type patents like all software patents but they don't scare me too much because sub-pixel rendering's value is debatable and it's not something that should affect programming at a higher level.

On the other hand, in the context of the thread, when rbanffy said "Don't forget Apple (among others) has tons of silly patents regarding text rendering that Gnome developers have to program around", in response to "truncation in OS X is much much smarter", he seemed to imply there were patents on the level of simply truncating text.

Are there?

If there are, it sound horrible.


Truncation != Rendering. Truncating long strings by replacing middle by ... is pretty common.


Wrote a long comment and then the web bug ate it :-(

In a nutshell: it's more of an introduction to GNOME Shell, which is just one part of GNOME3. It's going to be used by default, by old interface will still be available for those who want it. The UX is probably the least finished part yet, since a massive amount of work first went in invisible improvements for v3 (such as gobject-introspection).

Something the HN crowd will like: the Shell is written in JavaScript, has bindings to low-level libraries using gobject-introspection, uses JS libraries like jQuery, and comes with its own FireBug-like tool for inspection and modification at runtime. Since most of the look&feel is done in JS, you can just jump right in and hack left and right. One of most impressive demos at GUADEC2010 had a couple of JS oneliners doing various special effects on the desktop. Also, the whole thing is themeable in CSS, which is a lot nicer than doing gtkstyle theming.


> Also, the whole thing is themeable in CSS

Does this mean the ludicrously generous padding will be as easy to change as

    html * { padding: 2px!important; }
That would make me very happy.


Well, HTML/DOM is not used, but you should be able to do something of the sorts to the UI elements, yeah.


That's great to hear! :)


A lot of effort has gone into this; it's a shame to see so many disparaging posts here. I think the developers should be applauded for their work.

Between this and Ubuntu's plans to adopt Wayland in place of X, there are some interesting times ahead for Linux on the desktop. I, for one, am looking forward to seeing how it all plays out.


The developers absolutely deserve to be applauded for their work, but declaring it as "simply beautiful" is almost laughable. I think that it's actually quite ugly. And even if you disagree, I don't think it's fair to call your own work beautiful unless you've proven that the overwhelming majority consistently enjoys your aesthetic.


For a lot of people, Justin Bieber is a great musician. Popularity is not a trustworthy indicator of artistic merit.


Popularity is not a trustworthy indicator of artistic merit.

How else would you define it?


Artistic quality is one of those things that defy definition. It may be that it's aesthetically pleasing, or it may be that its design communicate some other inner quality or a completely different idea. It may be that the inner quality, in itself, has some aesthetic quality, much like a beautiful mathematical proof.


I read that as more of a goal statement than a description. Whether or not you think they've succeeded, they certainly seem to have put a lot of effort towards that goal: http://live.gnome.org/GnomeShell/Design/


They're trying to sell "simply beautiful" but all of the images on the front page have serious antialiasing problems.

Huh.


Stable software or a beautiful website... Tough choice when you have limited resources... I'm glad their priorities are right.


As far as open source desktops go, I'm rather excited about both Unity and Gnome3, but for very different reasons.

I think Unity, which is my preferred desktop right now, is shaping up to be the nicest, prettiest and easiest to use desktop open source has ever had. I might even consider it to be the first open source desktop I could point my mother to and she could use it straight away. I'm fully expecting Unity to be just plain awesome in several cycles.

Gnome3 with the JS core is open for extension like we've never seen before in open source. I say that b/c of the ease with which someone could extend now...like it or not, JS has a lower barrier to entry than traditional tech used in linux desktops. I can't wait to see what people create for Gnome3 via this extension scheme.

I know I'll have both installed for sure...fun times ahead!


"Simply Beautiful". It feels weird that the fonts on the site are better than on the screen shots. The font on the top and in the "Messaging built-in" is really grandma-sized and not really beautiful.

As said before, the theme and icons are really really lacking. I'm sure however that it will work great (and look great after some modifications like adding the Ubuntu font and Faenza Icons).


Is this information still relevant?

From the Try It link (http://www.gnome3.org/tryit.html):

GNOME 3 is under active development and will not be completed until April 2010. We hope to be able to provide live USB images really soon. Watch this space!


It appears the date has changed to April 2011


>A lot of effort has gone into this; it's a shame to see so many disparaging posts here. I think the developers should be applauded for their work.

bad work is bad. GNOME developers are again ditching code that works (and it sucks).

http://www.jwz.org/doc/cadt.html >This is, I think, the most common way for my bug reports to open source software projects to ever become closed. I report bugs; they go unread for a year, sometimes two; and then (surprise!) that module is rewritten from scratch -- and the new maintainer can't be bothered to check whether his new version has actually solved any of the known problems that existed in the previous version.


Gnome 3 Alpha, codenamed "Jumping the Shark". So many things shown are ridiculous - the spacing, the truncated titles in the ribbon-thingy, the seeming changes in the concept of workspaces. Why are dropbox and the wireless-thingy broken off from other system tray icons? What is the rationale for the changes in basic concepts? At the moment, I'm still using Gnome on all computers. This version might be just what I need to finally start using a minimal tiling manager.


Ubuntu's decision to move to Unity made me go back to Debian for GNOME/Clearlooks for the first time in a while, but what am I going to do now with the way GNOME 3 looks?


It seems that you switched before you tried. I don't know much about unity, but I was following the beginnings of gnome shell and I can't wait to get it in my hands. Activities / creating desktops as you need them is exactly what I wanted. But there are other cool things included too.


No, I used unity in the netbook edition in both 10.04 and 10.10. I switched when I heard it would become the default.


I'm totally with you on this one. XFCE is an excellent option to avoid this design madness.


Xmonad


> "Simply beautiful", "Distraction-free computing", "Everything at your fingertips", "completely redesigned for GNOME 3", "And much, much more", ...

This page reads much like a corporate ad.

> "GNOME 3 is crammed full of new features."

Is that a good thing?

For Gnome, I want: more streamlining, compartmentalization, removal of deprecated features, and optimization. Make it smaller and tighter -- only add parts conservatively and if they offer great value.


Gnome 3 "Made of Ugly." How is it possible that the aliasing on those fonts in the example image is allowed on the front page of their marketing. Every third line is a different size and has a blurry appearance.


That's the result of bad post-processing on the image.


Based on the screenshots, font rendering -- the most important thing -- still seems to suck big time.


>Simply Beautiful?

Seriously? Since when Beauty is that top priority for GUI? Usability, Gnome, Usability!


I don’t really see the contradiction.


How much of the eye-candy (IM, etc.) is so woven into the desktop that it's impossible to remove?

With OS X and Windows, I add the extraneous stuff - it doesn't come pre-installed and pre-enabled.

I'll stick w/XFCE, thanks. It's based on Gnome, but it doesn't have the cruft.


What? In Windows 7, the contemporary version to GNOME 3, all the flashy effects come on by default - or at least they did on this machine. You have to turn them off if you don't want them. IIRC OSX is basically the same... isn't it a little counter-productive having a whole lot of pretty effects if they're off by default?


typo on http://www.gnome3.org/faq.html? It says "Commons Questions and Answers", where I think it means "Common ..." Didn't see a contact link so maybe a webdev will see it here.


"AN OVERVIEW AT A GLACE (sic)" ???


I'm sorry, but that Metacity theme and even the GTK+ theme are really lacking. They lack the polish that Meerkat's Ambiance theme has (everytime I post a screenshot of stock 10.10 I get comments asking what the themes are).

It's a shame, I think Gnome-Shell looks cool and I think GNOME3 brings a lot of UI/UX changes that have been needed for a while, but they're not doing themselves any favors giving the actual appearance a real chance to shine. The "GNOME" default icons are a pity, too.

edit, I guess it's not Metacity any longer... Mutter.


That is an old gtk theme. The GTK3 look fits better with the new titlebar

http://www.flickr.com/photos/jakubsteiner/5201953378/sizes/o...


It's getting better, but that screenshot has plenty of it's own problems. That titlebar, menu, icon row and tabs takes up like 200+ px. Versus like 60 with Chrome.

What's the deal?


Large objects are easier to target. (That, and Chrome doesn't have a menu bar, which would at least bump it up to 80px.)




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: