|
|
Subscribe / Log in / New account

Shuttleworth: Window indicators

[Posted May 3, 2010 by ris]

Mark Shuttleworth takes a look at another use for the right side of the window title bar. "We've carefully placed all the panel indicators on the right, and we've carefully put the window controls and window title on the left. So now we have all this space on the right. As a pattern, it would fit to put the window indicators there. Cody Russell is leading some work in Canonical around the technology which actually draws the window title bar and borders. It's called "client side window decorations". We are moving the rendering of the window decorations into the app itself, so that you don't have the window manager and application drawing those pieces separately. That simplifies certain things (of course it also makes some things harder)."
to post comments

OT: Ubuntu promise

Posted May 3, 2010 19:27 UTC (Mon) by kragil (guest, #34373) [Link] (12 responses)

Did the last Ubuntu promise change with 10.04?
"Ubuntu core applications are all free and open source. We want you to use free and open source software, improve it and pass it on."
Was it always core applications only?

OT: Ubuntu promise

Posted May 3, 2010 19:43 UTC (Mon) by mjw (subscriber, #16740) [Link] (6 responses)

The last promise has eroded a little over the years according to archive.org:

2006: "Ubuntu is entirely committed to the principles of free software development; we encourage people to use free and open source software, improve it and pass it on."

2008: "Ubuntu CDs contain only free software applications; we encourage you to use free and open source software, improve it and pass it on."

OT: Ubuntu promise

Posted May 3, 2010 20:10 UTC (Mon) by kragil (guest, #34373) [Link] (5 responses)

Ok, thanks.

The 2006 promise certainly wouldn't work for Ubuntu One.

OT: Ubuntu promise

Posted May 4, 2010 7:31 UTC (Tue) by tcourbon (guest, #60669) [Link] (3 responses)

I fail to see the relation between Ubuntu One and the fact that, for you, Ubuntu doesn't promote any more free (as in freedom) software use.

Making money isn't nasty, even when the underlying "infrastructure" is free software (again, free as in freedom).

People should stop protesting against money making attempts since at the moment no company enforce the use of paid services to run Linux, and I doubt it will ever change.

OT: Ubuntu promise

Posted May 4, 2010 8:08 UTC (Tue) by dneary (guest, #55185) [Link]

The interesting question for me is whether it's harder to make money with free (as in freedom) web services than it is with non-free web services.

You assume that wanting free services implies not wanting the company to make money. I'd like the company to make money *and* have free services. The question is whether that's a viable model - and why it isn't, if not.

Dave.

OT: Ubuntu promise

Posted May 4, 2010 11:26 UTC (Tue) by kragil (guest, #34373) [Link] (1 responses)

Chill, all I said was that you can't say "Ubuntu is entirely committed to the principles of free software development; we encourage people to use free and open source software, improve it and pass it on." and then have a proprietary service called Ubuntu One. That just doesn't work. Simple logic. No reason to wet your pants.

What's "free software" for a web service.

Posted May 4, 2010 14:42 UTC (Tue) by grantingram (guest, #18390) [Link]

As far as I can see Ubuntu One is some file space which you can synchronise with and a music store.

What exactly is required from a file space and a music store in order for it to be free software? I'm sure complex web applications blur the boundary between desktop and server but this seems fairly clear cut.

The client appears to be GPL. Perhaps the server parts aren't free software but neither is google, lwn or the bbc for that matter....

I'm very chilled but I am curious as to where one would draw the boundaries!

OT: Ubuntu promise

Posted May 4, 2010 16:33 UTC (Tue) by iabervon (subscriber, #722) [Link]

I don't see how the 2006 version would conflict with pretty much anything other than discouraging the use and development of F/OSS. There's no reason that an entity couldn't develop both free and proprietary software and encourage people to make use of all permissions given in the licenses for each.

So it seems to me that, over time, they've moved to a statement that suggests less but actually promises what it suggests. AFAICT, they've never delivered what they used to suggest, and always delivered what they now promise.

OT: Ubuntu promise

Posted May 3, 2010 19:44 UTC (Mon) by Lukehasnoname (guest, #65152) [Link]

Multiverse and Restricted have always been repositories enabled by default that have usage or source code restrictions.

OT: Ubuntu promise

Posted May 3, 2010 19:47 UTC (Mon) by jspaleta (subscriber, #50639) [Link] (3 responses)

The wording has changed..it actually changed prior to the release of 10.04 according to google cache. Though I can not tell you exactly when.

http://web.archive.org/web/20080430210306/http://www.ubun...

Here's the old text from April 2008.

The Ubuntu promise

* Ubuntu will always be free of charge, including enterprise releases and security updates.
* Ubuntu comes with full commercial support from Canonical and hundreds of companies around the world.
* Ubuntu includes the very best translations and accessibility infrastructure that the free software community has to offer.
* Ubuntu CDs contain only free software applications; we encourage you to use free and open source software, improve it and pass it on.

Now it can easily be argued that the new language about "core" is a clarification of the intent stated previous about "CDs." In fact now that Ubuntu has official EC2 images and what not... the meaning of "core" could actually be more expansive than the older language concerning "CDs." The old promise certainly did not apply to things like Ubuntu branded EC2 images. But the new language might.

But since there's no explicit definition of "core" the new language is most certainly fuzzier and open to more interpretation than the old. You only get to cry foul over the language change if they actually release CD images with non-free software "applications". You may not even get to cry foul over libraries and drivers under the old language..again depending on what definition is applied to "applications".

As an expert in cataloguing all Canonical ill-will, I have to say the language change doesn't really meet my bar (and you must admit my bar is pretty low) for calling Canonical out about it.

Though a clarification on what core actually means probably wouldn't hurt.

-jef

OT: Ubuntu promise

Posted May 3, 2010 19:57 UTC (Mon) by Lukehasnoname (guest, #65152) [Link] (1 responses)

Jef, do you actually use Ubuntu? If so, is it only to criticize the distro? I've never seen you say anything positive about Canonical or Ubuntu.

I'm not defending what Canonical does all the time.

OT: Ubuntu promise

Posted May 3, 2010 21:23 UTC (Mon) by jspaleta (subscriber, #50639) [Link]

I find it amusing that this comment comes in a reply that basically says I'm giving Canonical the benefit of the doubt about something.

Now on to your own comments.

1) Nope I don't use Ubuntu. I also don't use products produced by the R. J. Reynolds Tobacco company. Are you suggesting that I should refrain from being critical of R. J. Reynolds Tobacco until I've enjoyed the smooth refreshing flavor of a pack of Salems? Yeah sure, RJ Reynolds Tobacco did contribute money to my high school and I made use of the facilities their money helped make possible at the school. But I still have the right to be critical of the public statements they make about their products right? Surely my status as an indirect beneficiary of R.J. Reynolds Tobacco's public community building efforts don't disqualify me from being critical of their business does it? Surely the same rules apply for my criticism in Canonical.

2) I've definitely said at least one nice thing about Canonical. I'm on record saying that the Papercuts idea is very good and that I was very pleased with the amount of effort they put into making sure the addressed papercuts were incorporated into upstream projects in the first attempt of the project for Karmic. I haven't reviewed the papercuts this cycle, but considering their performance in upstreaming things last year, I'm pretty confident that I'll find they've done as well..if not better...this time.

And I've said nice things about Canonical funded work Dustin Kirkland did with regard to working with upstream to get the home directory ecryptfs support in Ubuntu working.

-jef"smoke'em if you got'em"spaleta

OT: Ubuntu promise

Posted May 5, 2010 2:34 UTC (Wed) by daniels (subscriber, #16193) [Link]

As an expert in cataloguing all Canonical ill-will, I have to say the language change doesn't really meet my bar (and you must admit my bar is pretty low) for calling Canonical out about it.

Wow, I'm proud of you! :)

So is this going to be part of Gnome?

Posted May 3, 2010 20:06 UTC (Mon) by dbruce (guest, #57948) [Link] (24 responses)

I have a couple of reservations about this. First, assuming the "windicators" are a good idea, wouldn't it be better to spend the developers' effort directly working on the upstream Gnome project, rather than on Ubuntu's custom branch?*

Second, as an "upstream application developer", the idea of "client-side window decorations" seems counterintuitive, at least to me. I like the logical idea of the titlebar and window borders belonging to the system, while the contents of the window itself belong to the application. I don't like the Mac practice of having the single menu bar change to reflect whatever application currently has focus. Plus, having the app draw the window decorations seems to be a type of "platform-specific extensions" that are problematic when trying to write cross-platform programs.

DSB

*One might equally argue that it would have been better if the Ubuntu folks in general had put their efforts into improving Debian, rather than creating their own derivative. Just saying... (Note that I'm posting this from a Lucid machine)

So is this going to be part of Gnome?

Posted May 3, 2010 20:18 UTC (Mon) by Tara_Li (guest, #26706) [Link] (13 responses)

I have to agree - I already have Google Chrome up, and it apparently for some reason just *HAS* to draw its own window decorations, telling Enlightenment not to draw a border. (Use System Title Bar and Borders leaves me with *NOTHING* to max/min/close with.)

Applications aren't going to want to take on the extra responsibility of doing things they fairly rightly leave up to the Window Manager, especially if they're *SUPPOSED* to go look up the theme information and make themselves appear to fit said theme.

This braincramp just guarantees that Shuttleworth's big push to move all of the controls to the left will end up totally screwed up. Google's bad enough - *EVERY* damn App is going to take this as permission to come up with their own unique look, leading to the exact opposite of what he wants.

So is this going to be part of Gnome?

Posted May 3, 2010 20:29 UTC (Mon) by nteon (subscriber, #53899) [Link] (10 responses)

in Chrome go to Options->Personal Stuff and at the bottom is a radio box entitled "Use system title bar and borders".

So is this going to be part of Gnome?

Posted May 3, 2010 23:02 UTC (Mon) by Tara_Li (guest, #26706) [Link] (9 responses)

You missed the part where I said that didn't help, and in fact made it worse?

So is this going to be part of Gnome?

Posted May 4, 2010 0:01 UTC (Tue) by drag (guest, #31333) [Link] (7 responses)

Sounds like your Window Manager is the part that is broken.

I've used a few different ones and with Chrome I preferred to use the WM borders (disabled the default Chrome draw-my-own) for a few different reasons and I have not had the problems you've described.

------------------------------

As far as Client-side vs WM-drawn window borders. One of the challenges that application designers face, especially with GTK, is how to make efficient and attractive use of the application controls at the top of most applications. This is a problem with most Linux desktop stuff.. they are all very inefficient space-wise and are not terribly good looking.

Client drawn cnotrols can help out application designers in these regards when used properly. Most applications probably would not care too deal with worrying about WM controls, and many developers that do would probably just abuse it, but it's something that is nice to have.

So is this going to be part of Gnome?

Posted May 4, 2010 2:12 UTC (Tue) by Tara_Li (guest, #26706) [Link] (4 responses)

If the problem of the window title bar is so bad - why not fix it in the window manager, and not have to change all of the apps to have them do it all kinds of different ways, most likely fairly confusing ways.

One of the reasons I do use e16 is that I have a theme for it that makes the window decorations very small.

So is this going to be part of Gnome?

Posted May 4, 2010 9:39 UTC (Tue) by mjthayer (guest, #39183) [Link] (3 responses)

Lets face it, in most cases it won't be the apps that are handling this but the widget library that they use. And those libraries are normally themable anyway - does it make that much sense to keep the window decorations separate from the widget library so that you have to set the look and feel you want in two different places?

And in an answer to a comment on his blog, M.S. confirmed that there will be a mechanism for the window manager (or anyone else) to render those additional icons instead of the application and disable the client-side rendering. Which means that there will be a limit on the silly things that applications will be able to do with them and still work.

I do agree that it is nice to abstract the interface away from the application, but the separate window manager decorations are breaking off a tiny bit of the interface and just abstracting that, which isn't quite so clean. I would love to see the same thing pushed the whole way, but that would probably be a more major undertaking!

So is this going to be part of Gnome?

Posted May 4, 2010 17:32 UTC (Tue) by iabervon (subscriber, #722) [Link] (2 responses)

Why wouldn't the window manager use system-wide themes? Regardless of whether the window manager or the application renders the controls, if every program is sharing the theme in order to get consistency, this ought to extend to the window manager. And if not every program is sharing the theme, I think it's better if the controls for all windows match each other than if the controls for windows that don't share the same theme match the theme and not each other.

That is, the window controls should be from the same theme as everything else. But if two windows have different themes, it's a bit awkward if one (or both) have controls that clash with the window contents, but a complete disaster if you end up with two windows on the same desktop with different controls.

So is this going to be part of Gnome?

Posted May 4, 2010 21:56 UTC (Tue) by mp (subscriber, #5615) [Link] (1 responses)

a complete disaster if you end up with two windows on the same desktop with different controls

Indeed, but what a funny disaster. Now I just cannot resist confessing to having worked for a whole couple of days with the WM (sawfish it was then) configured to assign random decoration themes to all newly mapped windows.

So is this going to be part of Gnome?

Posted May 14, 2010 16:26 UTC (Fri) by jospoortvliet (guest, #33164) [Link]

FYI, a discussion about this is going on.

read if yo're interested: http://blog.martin-graesslin.com/blog/2010/05/open-letter...

So is this going to be part of Gnome?

Posted May 4, 2010 4:17 UTC (Tue) by rsidd (subscriber, #2582) [Link] (1 responses)

With kwin (kde 4.4.1), too, checking "use system title bar and borders" gets rid of all window decorations -- no close buttons etc.

So is this going to be part of Gnome?

Posted May 14, 2010 16:25 UTC (Fri) by jospoortvliet (guest, #33164) [Link]

alt-F3 -> advanced - untick no border ;-)

So is this going to be part of Gnome?

Posted May 4, 2010 6:19 UTC (Tue) by nteon (subscriber, #53899) [Link]

yup, somehow that part didn't register in my brain.

So is this going to be part of Gnome?

Posted May 3, 2010 21:40 UTC (Mon) by AlexHudson (guest, #41828) [Link]

Does feel a lot like just moving deckchairs around, or (in this case) buttons. I just don't see the point. 10.4 has a relatively nice clean UI, and it sounds like that's going to be thrown away to make room for more icons.

I remember reading an interview with one of the graphic artists for (I think) Monkey Island on the Amiga, who made a comment about their graphical interface requiring icons for actions. Basically, the more complex the action they were trying to get across, the worse the icons became, until they came to design an icon for "Pay respects to the dead" in 32x32 or however many pixels. That's exactly what's going to happen here.

None of this stuff is going to be really user-friendly. It's going to diverge from GNOME, but worse, it's going to diverge from other Ubuntu (e.g. the netbook spin getting a global menu with presumably totally different icon behaviour). When you have more that one volume control on the screen it's a pretty clear indication that you've lost; only geeks will possibly understand that.

So is this going to be part of Gnome?

Posted May 6, 2010 12:30 UTC (Thu) by fragmede (guest, #50925) [Link]

The window manager you use is ignoring a change to the window properties, specifically, _MOTIF_WM_HINTS. Workaround is to restart Chrome after setting the property.

(Crhome bug - http://code.google.com/p/chromium/issues/detail?id=16044 )
(Openbox bug - http://bugzilla.icculus.org/show_bug.cgi?id=4250 )

So is this going to be part of Gnome?

Posted May 3, 2010 22:31 UTC (Mon) by jspaleta (subscriber, #50639) [Link] (9 responses)

I've seen a rationale for forking Debian. Debian isn't focused on what Shuttleworth wanted to do with Ubuntu. It's a distinctly different focus for the distributions and the timescales over which the distributions are released.

But...with Gnome its a very different story. Gnome and Ubuntu share very very similar usability targets...and very very similar release schedules. I haven't seen a rationale as to why Canonical can't jump in and really start being a part of the Gnome 3 process and really commit to driving code commits in the GNOME 3 timeframe.

I'll say it again. Canonical hosted the Gnome Usability hackfast. Canonical design team members had the ability to be in the room when Gnome 3 design issues were being discussed as part of that hackfest. Did they or did they not bring up the suitability of libappindicator at that hackfest?
I've seen zero discussion about a libappindicator discussion during the hackfest...anywhere. Why is that?

It seems to me Canonical has put a lot of thought into their roadmap for libappindicator... and yet... they seem really reluctant to actually have the discussion about how it fits into the design for GNOME 3. Did they discuss it at the hackfest or not? If not why not?

-jef

So is this going to be part of Gnome?

Posted May 3, 2010 23:43 UTC (Mon) by foom (subscriber, #14868) [Link] (2 responses)

It sounds like client-side-decoration support has been discussed with Gnome (and KDE).

http://jasondclinton.livejournal.com/74222.html

"The discussion moved on to client-side-decorations with Cody Russell which are already being developed on a side-branch. The work has a goal of being merged for GTK 3."

"Cody said that he has be communicating with Trolltech to ensure that Qt does it in a sufficiently similar way in Qt that there won't be compatibility issues between Qt and GTK applications that provide their own CSD."

So is this going to be part of Gnome?

Posted May 4, 2010 0:52 UTC (Tue) by jspaleta (subscriber, #50639) [Link]

Client side decorations... are not libappindicator.

And I'll note for the record that quality technical collaboration via patches on how to implement csd in gtk+ has been ongoing in the gnome bug tracking system for almost a year...pretty much from the time code was published for review. That is very encouraging to see Canonical getting upstream review involved very early on and doing the right things for a smooth path towards inclusion into gtk+. In fact Cody mentioned working on it in May of 2009 in the gtk+ devel list...giving people a heads up that code was coming for review. Gold stars all around.

That makes the lack of discussion concerning the development of libappindicator that much more of a stand out in stark contrast to how csd is being handled. Canonical clearly wants csd incorporated into gtk and got upstream review involved early in the implementation process. If Canonical was really interested in seeing libappindicator adopted by upstream..they could have done the same thing with a path towards integration into gtk...but they haven't really picked up the ball and carried it towards that goal.

Just as interesting Cory brings up the benefits of gtk+ gaining the general purpose functionality of libappindicate in the Nov 2009 GTK+ team meeting at the same time hinting about the global menu work that Shuttleworth recently announced. But unlike the case of csd.. the gtk+ patchsets for libappindicator-like functionality haven't been forth coming for upstream review.

Several months later discussion over libappindicator as an external dep proposal dances around the issue as to whether or not libappindicator should be implemented in gtk+ or be a stand alone library..even though Cody already brought it up. The situation with libappindicator is a confused mess of Canonical's own making. I was hopeful Canonical would bring it up for discussion at the hackfest like Ted implied he would get someone to do and set some firm foundations in place for a commitment to work out the specifics to see it integrated. But I can't find any evidence at all that it was discussed.

-jef

So is this going to be part of Gnome?

Posted May 14, 2010 16:30 UTC (Fri) by jospoortvliet (guest, #33164) [Link]

The KDE dev's aren't exactly positive about it however:
http://blog.martin-graesslin.com/blog/2010/05/open-letter...

So is this going to be part of Gnome?

Posted May 4, 2010 3:28 UTC (Tue) by ccurtis (guest, #49713) [Link] (5 responses)

Is it so horrible that if this thing turns out to be a dud it's only going to impact Ubuntu users and not all GNOME users? I realize that doesn't answer your question, but is it really necessary to committee everything to death in advance or is Canonical allowed to try something novel?

So is this going to be part of Gnome?

Posted May 4, 2010 4:07 UTC (Tue) by jspaleta (subscriber, #50639) [Link] (2 responses)

libappindicator has been submitted for inclusion as an external dep to GNOME. That process _requires_ a statement on how libappindicator is meant to impact GNOME 3. Canonical continues to dodge the question. I'm giving Canonical the benefit of the doubt and assuming they aren't just submitting it to waste the time of the Gnome release team. If they are serious about it, then they need to seriously answer the question about its relevance to GNOME 3, specifically gnome-shell. Ted Gould specifically said that he was going to find someone to bring it up at the hackfest as part of discussion sourronding the submission as an external dep. So... did that discussion happen or not?

What is Canonical's intentions with regard to seeing the functionality of libappindicator integrated into Gnome 3? If they want to just go off and do their own thing...that'd be fine..but then why submit it as an external dep? But from where I sit, it looks like they are sort of half-assing a conversation and stringing Gnome upstream along about this stuff as a suitable technology. Not cool. Commit to collaboration, don't start a conversation and then block technical discussion by being negligent in responding. Canonical needs to go all-in and commit to finishing the conversation and really work through the technical issues that have been brought with libappindicator.

-jef

So is this going to be part of Gnome?

Posted May 4, 2010 4:26 UTC (Tue) by foom (subscriber, #14868) [Link] (1 responses)

> So... did that discussion happen or not?

That seems like an easy thing to find out. Maybe you should ask the people involved and report back.

But I guess if you did that it wouldn't be possible to cloak accusations of ill-will in the guise of questions?

So is this going to be part of Gnome?

Posted May 4, 2010 16:47 UTC (Tue) by jspaleta (subscriber, #50639) [Link]

Unlike the average blogger or for what passes for technology laypress on the internet, I'm not inclined to broadcast private off-the-record comments into public as fact by quoting anonymous sources. I require someone to go on record publicly or I need to find a public archive that answers the question that I can cite as a source. Its called sourcing your information. Something journalists used to be required to do when the words journalistic and integrity were used in close proximity in the same sentence.

I ask the questions here, because the track record with regard to factual information from quoted sources in the articles is very good here.

I also know that the readership includes people with the information I'm requesting...because some of them have responded to me in the past here.

I also know the subscribed readership here is somewhat discerning and would never let me get away with "anonymous sources tell me..." when making a critical claim with regard to Canonical's action.

You want me to shut up about it? Help me find someone who will stick their neck out and go on record with a public summary of the discussion of libappindicator at the hackfest.

-jef

So is this going to be part of Gnome?

Posted May 4, 2010 8:11 UTC (Tue) by seyman (subscriber, #1172) [Link]

> Is it so horrible that if this thing turns out to be a dud it's only going to impact Ubuntu users and not all GNOME users?

We've had security holes introduced by distribution-level patches in the past and one would think that packagers would have learnt their lesson in this regard.

So is this going to be part of Gnome?

Posted May 4, 2010 13:57 UTC (Tue) by marcH (subscriber, #57642) [Link]

> Is it so horrible that if this thing turns out to be a dud it's only going to impact Ubuntu users and not all GNOME users?

Not even all Ubuntu users are going to be impacted.

People seem to forget that Ubuntu does definitely not force GNOME on all its users (I would not use Ubuntu otherwise).

Shuttleworth: Window indicators

Posted May 3, 2010 20:13 UTC (Mon) by wonder (guest, #64293) [Link]

is it me or they actually will not support the new gnome 3 layout since they are forking gnome-panel or whatever they do there?

Shuttleworth: Window indicators

Posted May 3, 2010 21:43 UTC (Mon) by dskoll (subscriber, #1630) [Link] (24 responses)

So Shuttleworth wants to throw out a 25+-year-old X Window system design decision? He'd better hope he's right... but I doubt it. For all the criticism X has come under, it's withstood the test of time.

It seems to me that Shuttleworth is a De Icaza wannabe, except he wants to turn Linux into Mac OS instead of MS Windows.

Shuttleworth: Window indicators

Posted May 3, 2010 23:38 UTC (Mon) by khc (guest, #45209) [Link]

I think client side window decoration is actually a GTK+ project, ubuntu will just take advantage of it to do windicators

Shuttleworth: Window indicators

Posted May 4, 2010 3:35 UTC (Tue) by jbailey (guest, #16890) [Link] (3 responses)

I used to work for Canonical, and Shuttleworth is crazy in a lot of ways, but he's not pushing to turn Linux into anything. He talked a lot about new people to Ubuntu, and a large portion of those people being potentially new people to computing. The notion of what current users are comfortable with is less interesting than a desktop that just works and just feels good.

The discussion threads here seem to think that Ubuntu should go to Gnome and participate is committee discussions about how the platform could look. Why is it important that the desktop look one way or another? Ubuntu is carrying out a vision that they have, and Gnome is free to do the same thing or not. Or even better, both can learn from one another until maybe one day they converge on an idea.

The world is richer for the experiments being done on both sides.

Shuttleworth: Window indicators

Posted May 4, 2010 6:44 UTC (Tue) by jengelh (guest, #33263) [Link] (1 responses)

>The notion of what current users are comfortable with is less interesting than a desktop that just works and just feels good.

Then we should just keep the default "[system menu] title [minimize] [maximize] [close]" layout. Because that's what they are comfortable with (from the various Windows versions) and it also works in principle. For feeling good, I have added "[depth]" (bring-to-bottom) and "[rollup]" buttons, but the rest is still the same.

Shuttleworth: Window indicators

Posted May 4, 2010 20:23 UTC (Tue) by Tara_Li (guest, #26706) [Link]

Truth be told, I right-click my title bar, most of the time. Sure, on the left, I have the minimize and shade buttons, and I use the minimize one from time to time, and on the right I have sticky, close (right-clicking that one force-closes the window, may leave the app running), and a flyout that has a whole host, including sending the window to other desktops, raising & lowering the window, and toggling max width, max height, and max size. But right-clicking on the title bar gives me a nice menu of the same things, in text format, for easy picking.

I love Enlightenment 16.

Shuttleworth: Window indicators

Posted May 4, 2010 10:33 UTC (Tue) by NAR (subscriber, #1313) [Link]

The notion of what current users are comfortable with is less interesting than a desktop that just works and just feels good.

So he's prepared to lose his current users to gain - exactly what? Some goat-herds in western Kenya who never saw Microsoft Windows yet? It could be a charitable mission and it's perfectly in his rights to burn his money on whatever he wants, but I don't think it's really useful. I mean car makers are not "experimenting" with changing the brake with the accelerator for "usability"...

And for the record I've just managed to ban a girl with nice bottom on MSN in pidgin, because some "usability expert" moved the "OK" button to the left, instead of the right. So I'm not in that good mood.

Shuttleworth: Window indicators

Posted May 4, 2010 7:44 UTC (Tue) by jmm82 (guest, #59425) [Link] (2 responses)

You could say the same for sysvinit -> upstart. I will be the first to say I did not want to learn a new init system, but I am starting to like it.

I do think it is slightly crazy and it is still unclear why the new buttons could not go on the left....anyways I am not following it much. I installed Ubuntu 10.04 and moved my buttons back to the right.

I will personally give the Shuttlemaster(lame space reference) the benefit of the doubt. I've been using Ubuntu for 4 years. Worst case, Ubuntu gets too crazy and I switch back to Fedora.

Beyond the surface, it basically boils down to apt-get vs. yum and finding where the config files were moved to.

If everyone was narrow minded we would still be using punch cards. I'm sure a few "crazy" people dropped their stack of punch cards one too many times and decided "Hey, we can do better!"

Shuttleworth: Window indicators

Posted May 4, 2010 9:43 UTC (Tue) by mjthayer (guest, #39183) [Link] (1 responses)

> You could say the same for sysvinit -> upstart. I will be the first to say I did not want to learn a new init system, but I am starting to like it.

I'm waiting for systemd to reach maturity... that potentially looks like something to outclass both of them.

Shuttleworth: Window indicators

Posted May 14, 2010 16:34 UTC (Fri) by jospoortvliet (guest, #33164) [Link]

While systemd looks good, I don't get a good feeling from yet ANOTHER replacement for an infrastructural piece of our favorite OS. I've read the rationale, and understand it, but like with Pulseaudio (ALSA still feels brand new to me, darn, I've got some OSS apps still) I feel it's a little too much too fast.

Then again, I'm part of the KDE Krowd and we know all about this choice and often also go the way of introducing something new and innovative if existing choices don't cut it... So maybe I shouldn't criticize it. Meh.

Shuttleworth: Window indicators

Posted May 4, 2010 13:31 UTC (Tue) by freemars (subscriber, #4235) [Link] (11 responses)

So Shuttleworth wants to throw out a 25+-year-old X Window system design decision?

In fairness to Shuttleworth, the near-universal switch to 16:9 displays is a bit of a game-changer. Anything that takes up vertical space on screen now must be considered very carefully. Office workers could probably benefit if they turned their shiny new LCD display to portrait mode, but Shuttleworth's (and Balmer's) U.I. designers can't count on that for more than ~25% of their users.

I'll be following along to Ubuntu 10.4, but my next update may well be to Debian or ??.

Shuttleworth: Window indicators

Posted May 4, 2010 19:04 UTC (Tue) by jhardin@impsec.org (guest, #15045) [Link] (5 responses)

"the near-universal switch to 16:9 displays is a bit of a game-changer."

So how does that mean the applications should now be responsible for drawing window decorations? If vertical real estate is now precious, start making 16:9 desktop manager themes that put the decorations in the "correct places".

Having the applications decide how to place decorations is _not_ a step forward. It does not improve UI consistency.

Shuttleworth: Window indicators

Posted May 4, 2010 23:55 UTC (Tue) by ringerc (subscriber, #3071) [Link] (4 responses)

With most hardware, using rotation disables acceleration - and not just 3D. It's not appealing. Curse 16:9 monitors!

Shuttleworth: Window indicators

Posted May 5, 2010 1:09 UTC (Wed) by foom (subscriber, #14868) [Link] (1 responses)

Hm, I can see how screen rotation might disable the old special cased acceleration frameworks, but *not* 3D (or any of the modern acceleration built on top of the generic 3D hardware).

Shuttleworth: Window indicators

Posted May 5, 2010 5:43 UTC (Wed) by ringerc (subscriber, #3071) [Link]

It seems to absolutely ruin performance on my machine (GF4-based video, nouveau) but indeed it may be that with fully 3D accelerated displays it's not longer an issue. It never worked well with the nVidia drivers, but perhaps with free 3d-enabled drivers things would be better.

That'd be nice, as a vertical 24" 16:9 display would be a whole lot more useful than a horizontal one. Alas, none of my hardware likes running in portrait much.

Shuttleworth: Window indicators

Posted May 5, 2010 15:31 UTC (Wed) by jhardin@impsec.org (guest, #15045) [Link]

...I wasn't thinking of physical rotation of the monitor to make a landscape monitor into portrait, if that's what you're referring to, though hearing that doing so disables 3D acceleration is indeed disappointing.

I was suggesting designing a theme with the title bar and window controls on one of the sides of the windows rather than on the top.

Is there some limitation to theming capabilities that prevents doing this?

Shuttleworth: Window indicators

Posted May 5, 2010 21:18 UTC (Wed) by odie (guest, #738) [Link]

I use a pivotable display connected to a Radeon 9250 using the standard open source driver, and performance is perfectly acceptable in portrait mode (meaning I can't notice any slowdown at all). I might not have used 3D apps in portrait (I hardly ever use openGL apps), but 2D acceleration certainly works, including xv.

Shuttleworth: Window indicators

Posted May 4, 2010 20:27 UTC (Tue) by Tara_Li (guest, #26706) [Link] (2 responses)

You know, with E16, this is not a problem - pick an E theme that has the title bar and controls on - TADA!!! - the side!

Shuttleworth: Window indicators

Posted May 6, 2010 4:27 UTC (Thu) by thedevil (guest, #32913) [Link] (1 responses)

On the side? How is the title actually displayed? Is it like a thin limb sticking out on the side of the window? Then it wastes either all the pixels further to the side of it, or those beneath/above it, depending on how you place the adjacent window.

You *don't* really mean one of these horrors with the TEXT actually rotated, do you?

Shuttleworth: Window indicators

Posted May 8, 2010 23:02 UTC (Sat) by nix (subscriber, #2304) [Link]

Yes, the text is rotated. (I didn't realise E was still going. E17 was due to be Out Soon and would Rock the World back in 2001 or thereabouts, and the nothing happened, and nothing happened, and everyone fell asleep. Is it still alive?)

Shuttleworth: Window indicators

Posted May 5, 2010 11:50 UTC (Wed) by sorpigal (guest, #36106) [Link] (1 responses)

"In fairness to Shuttleworth, the near-universal switch to 16:9 displays is a bit of a game-changer. Anything that takes up vertical space on screen now must be considered very carefully."

My 16:9 monitor has just as much vertical space as my old 4:3 monitor had; more, actually, because 900 > 768. Vertical space is no more expensive now, it's just that horizontal space got a little cheaper.

Vertical space is generally infinite and horizontal space at a premium, or so people always told me when they criticized the way that I sacrificed 48px of the right side of my screen for an always-visible dock and gkrellm.

If vertical space is really important why not ignore client side decorations and simple use a WM (or WM theme) with decorations on one side, or both sides, of the window? A far more sane, simple and standard solution.

Shuttleworth: Window indicators

Posted May 14, 2010 16:28 UTC (Fri) by jospoortvliet (guest, #33164) [Link]

Yep. Or automatically hide borders and decoration when an app is maximized - used in the KDE Plasma Netbook interface. Will also be an option in Kwin on the Plasma 4.5 or 4.6 desktop (not sure if it made the cut for 4.5).

Shuttleworth: Window indicators

Posted May 5, 2010 10:08 UTC (Wed) by cortana (subscriber, #24596) [Link] (3 responses)

Does this mean that we're importing the bug in Microsoft Windows where the windows of an application that has frozen are unable to be moved/resized/closed?

Shuttleworth: Window indicators

Posted May 5, 2010 15:33 UTC (Wed) by jhardin@impsec.org (guest, #15045) [Link] (2 responses)

That's my greatest single hatred of Windows and my greatest fear of Shuttleworth suggesting design changes of this nature. Let's not adopt fundamental flaws of Windows - such are not improvements.

Shuttleworth: Window indicators

Posted May 5, 2010 16:31 UTC (Wed) by cortana (subscriber, #24596) [Link] (1 responses)

Having said that, if the window manager could realise that a client has frozen or is otherwise misbehaving, it could steal back responsibility for the window borders and let the user resize the window...

Shuttleworth: Window indicators

Posted May 5, 2010 18:47 UTC (Wed) by oak (guest, #2786) [Link]

> if the window manager could realise that a client has frozen or is otherwise misbehaving

Normal window managers do this by pinging the applications (see _NET_WM_PING message in EWMH spec) when user clicks to the close button. Or by counting how many times user clicks the same close button in close succession.

The freezing detection needs to be tied to some user action where it's reasonable to check the app. Just polling them all the time would ruin battery life.

Without any visible controls with which user could interact with the application that's quite hard though. Using keyboard shortcut (e.g. Windows key) for popping up a menu for window actions would work, but this would be hard to discover for normal users. Another possibility is doing this kind of freeze detection just when switching back to the application (focusing it, clicking on its icon in panel etc).

The good point about application doing the resizing controls and window frames is that this it gets rid of of the synchronization issue between window manager and application screen updates. When OpenGL HW acceleration is used, it additionally gets rid of interleaved GL commands which can be (at least a performance) problem for some low end GFX chips.

Shuttleworth: Window indicators

Posted May 3, 2010 23:13 UTC (Mon) by smokeing (guest, #53685) [Link] (5 responses)

Sorry for being a little off-topic, but blimey, is it Comic Sans I see on the mockup??

Shuttleworth: Window indicators

Posted May 4, 2010 16:21 UTC (Tue) by frazier (guest, #3060) [Link] (4 responses)

I don't know if it is Comic Sans or not, but the font is being used intentionally to try to make the wireframe appear to be hand drawn instead of computer generated. They don't want to give the impression of a polished, final product.

Shuttleworth: Window indicators

Posted May 4, 2010 19:04 UTC (Tue) by smokeing (guest, #53685) [Link] (3 responses)

> I don't know if it is Comic Sans or not

http://en.wikipedia.org/Comic_Sans

A fleeting appearance of Comic Sans has the effect of a strong purgative for anyone even remotely involved in graphic design. It tells of utter incompetence riding on unabashed self-conceit. There are just too many other, suitable casual fonts, and those who use Comic Sans either don't care or don't have a clue.

And this incompetence shines through. If Shuttleworth was so carried away with whatever UI design concepts seemed revolutionary to him as to never mind the fact his mockups are done by some monkey on a Windows machine, then what I hear is child's babble, I am not listening past this point, and altogether not interested.

Shuttleworth: Window indicators

Posted May 4, 2010 20:53 UTC (Tue) by baran (guest, #28429) [Link] (1 responses)

Do pontificate some more. I'd like to see brighter examples of your wisdom's pearls.

comic sans just won't die

Posted May 4, 2010 21:34 UTC (Tue) by smokeing (guest, #53685) [Link]

Thanks for the appreciation :}

Anywhere else, Comic Sans just would mean a sloppy manner of expression; something just 'casual'. But it is none other than the spiritual leader of one major Linux distro who's speaking. And here, in this context, the use of an essential attribute of Windows looks inappropriate to say the least.

But all this doesn't really belong here; I have said 'sorry' already at the beginning of the thread.

Chill out

Posted May 4, 2010 21:03 UTC (Tue) by dbruce (guest, #57948) [Link]

Really, who gives a crap if the font in these concept drawings is comic sans or anything else? Does it have even the slightest relevance to the validity of his argument?

For that matter, do font selections in general have relevance to much of anything of importance at all? Sure, some fonts look nicer than others, but on the whole I barely care.

DSB

Shuttleworth: Window indicators

Posted May 4, 2010 3:07 UTC (Tue) by xnox (guest, #63320) [Link]

Ayatana team wants to do what they want to do on their schedule. It's a different approach with consistent long term goals and incremental releases which they reveal during each Ubuntu cycle. I don't see point in pushing it to Gtk so soon getting it wrong and maintaining bad API for so long. Doing this on a branch allows you to experiment and getting in right when it is time to merge.

It's like with linux kernel and android fork. Linus doesn't mind so what is your problem lwn.net readers =)

If you are a distro and want to have this too merge from launchpad.net =)

Shuttleworth: Window indicators

Posted May 4, 2010 8:13 UTC (Tue) by zuki (subscriber, #41808) [Link]

From Shuttleworth's post:
> I really liked the Chrome browser's use of overlay status messages, so
> kudos and thanks to them for the inspiration. The net result of those
> two steps, in apps where we can, is to save about 5% of the vertical
> space for your stuff - real content.

The overlay status messages are one of the things chromium does great - without them the fullscreen mode is useless. And when browsing on a netbook, fullscreen statusbar-less bookmarks-bar-less menu-less mode is a must.

The same goes for the screen real-estate at the top - the current layout is just to sparse. I'm not sure what is the best technical solution to do this rearrangement, but somehow it must be done.

I would make use of this. Seems a logical place for some additions

Posted May 4, 2010 9:22 UTC (Tue) by amtota (guest, #4012) [Link]

This is one thing that used to be easier to do on MS Windows (albeit ugly and much more limited in scope).
And despite this laudable effort, it is likely to remain so until *all* the distros, and the most commonly used window managers pick this up…
Let’s hope it doesn’t take too long, because I think this is a great idea. It makes a lot of sense to keep the volume controls and such in a location that makes it more intuitive to the user.

This is something which would be very useful to me.
I have written a tool (http://shifter.devloop.org.uk) which allows you to suspend, resume and send windows to other desktops (apps running via NX, VNC or xpra), so my use case would consist of:
* a button to suspend the window (it disappears but is still running on the server – which may be the same machine)
* a button to kill the session (not essential) – this is used to kill the process, different from closing the window since this may not be running on the same machine
* a button to send to another user, this should provide a drop down with a list of users, or alternatively call my code to take over.

The real difficulty here is that the windows that my app controls are not actually managed directly by the application: my application launches the NX client (or VNC or xpra) and from then on it is that process which will connect to the window manager.
I would love to be able to overload this behaviour to allow my application to inject its own widgets into the client processes it launches. But I think this is going to be non-trivial?

I actually filed a bug 6 months ago:
http://shifter.devloop.org.uk/trac/ticket/24
Saying I had little hope of seeing progress in this area, so thanks!

(I posted this message to his blog and got flagged as spam, WTF?)

Shuttleworth: Window indicators

Posted May 4, 2010 10:41 UTC (Tue) by Hanno (guest, #41730) [Link] (4 responses)

Ubuntu 10.04 looks very very nice, a great job by the designers. The new theme is quite an improvement. However, I moved the window buttons to the right after installation...

The indicators look like a nice idea, but actually, it seems to be a grand effor to solve a non-problem. Most screens are wide these days and I would have preferred a sensible way to put system trays and process lists in vertical stripes instead of horizontal.

Let's see how it turns out. As another commenter already mentioned, users can always switch to different distros if Ubuntu fucks up this time.

Shuttleworth: Window indicators

Posted May 4, 2010 14:42 UTC (Tue) by kirkengaard (guest, #15022) [Link] (1 responses)

Agreed - the effort is to solve a manufactured problem, of how to put a new feature in. As I type this from a 16:9 screen, and take note that wherever the buttons go, the bar is the same height, I'm wondering why this feeping creature is being used with the screen dimension problem as justification.

"Window indicators" have gone wherever the task is already understood to be. Task bar, task menu, whatever, the callout is the same, and the UI handles it. The window is part of the UI, and so he wants to move the functionality to the window. Fine. We do understand the task to be in the window, so that's reasonable, except that moving that function to the window still requires some manner of solution for when the window is obscured.

My real question is, is he dumb enough to mean what he says about moving that functionality "into the app itself," taking it from a useful centralized UI abstraction back down into reliance on every single application to do "the right thing"?

It sounds like Shuttleworth really intends what I've been hearing in the technical discussions, about having a new abstraction to draw CSD, and placing those in the title bar and then in the panel for maximized application windows. So he's offering applications a new UI system for displaying common user notifications, and asking application developers to use it; which he will then place in the title bar in the arrangement that's irritating so many people, and then also wants to maximize into the panel, voiding any use for the panel besides status indication because this needs to co-opt the space we all use for widgets and application shortcuts. But all he's saying is the last part, and calling it "moving this functionality into the app itself."

Shuttleworth: Window indicators

Posted May 4, 2010 16:06 UTC (Tue) by mjthayer (guest, #39183) [Link]

> My real question is, is he dumb enough to mean what he says about moving that functionality "into the app itself," taking it from a useful centralized UI abstraction back down into reliance on every single application to do "the right thing"?

The vast majority of the applications' UIs are in the applications already, and people expect them to get it right. Having the title bar and buttons rendered by gtk+ with its current theme inside the application rather than gtk+ in the window manager seems like a minor move in that context. And as gtk+ is themable, you can still select a theme that has a title bar like you would have had with the window manager (or that disables the client drawn title bar for that matter).

As I wrote in a previous comment, I am all for abstracting the UI from the application, but breaking off a small bit of the UI in the form of the window decorations and just abstracting that doesn't seem such a big thing to me - and why not just have it match the rest of the applications UI, instead of fighting to get the application's and the window manager's theme to more or less match? The user is still free to vote with their feet if applications mess up the title bar (rather than, in particular, just letting their toolkit handle it).

Shuttleworth: Window indicators

Posted May 4, 2010 14:56 UTC (Tue) by grantingram (guest, #18390) [Link] (1 responses)

I may be the only one but I like the new window button position!

Especially on the laptop it seems to involve a lot less scrolling on the touch pad to the top right and back to do something useful with the mouse pointer.

I don't really understand why everything has to be the same in all distributions. If Ubuntu are not doing something different - why do they exist?

Shuttleworth: Window indicators

Posted May 4, 2010 15:10 UTC (Tue) by sharms (guest, #57357) [Link]

I also like the new button positions


Copyright © 2010, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds