Lea: The Unity design process (and how you can play a part in it) [LWN.net]

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 17, 2012 16:28 UTC (Fri) by ean5533 (guest, #69480)
In reply to: Lea: The Unity design process (and how you can play a part in it) by cmorgan
Parent article: Lea: The Unity design process (and how you can play a part in it)

And now several more people will respond to you with "justifications" for their chronic rants. Popular arguments will include:

-Unity is being forced upon us by Canonical and it's not fair.
-KDE/Gnome/XFCE/LXDE/awesome/WindowMaker is way better than Unity and Unity is stealing attention from it/them.
-Unity is just fragmenting the user space and confusing users with too much choice.
-Unity's poor design will (somehow) infect other existing DE's.
-Canonical is ruining the spirit of the Linux community.
-I see other people ranting about Unity in comments so obviously no one likes Unity, and since no one likes it then it shouldn't exist.

Some of the responses will have valid points, but they will be lost in the flood of irrational hatred.

to post comments

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 17, 2012 21:30 UTC (Fri) by rahvin (guest, #16953) [Link] (5 responses)

Irrational? When I loaded Unity for the first time I couldn't figure out how to open a terminal window. Do you have any idea how frustrating that was? The second version (I'm not good at tracking the version number but I think the first was 11.04 and the second was 11.10) fixed some problems so that it was actually possible to find the terminal icon and click it. I've been using computers since I was 8 years old, my first PC that I purchased myself was a top of the line 386 with 4 megs of memory and an 80meg hard drive so that should give you a clue how old I am.

Unity was so anti-intuitive for a regular computer user that it wasn't even funny. I like that they are experimenting but Christ, they need a button that says "Flip back to regular interface" on the toolbar so people that want to get actual work done can do so. This is why people are complaining. People are upgrading and can't even open a program anymore, it's like turning advanced users into first time users and it's VERY frustrating. Particularly given the lack of any type of help when you first run it.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 8:08 UTC (Sat) by Rehdon (guest, #45440) [Link]

I have 'converted' quite a number of people to Ubuntu, I even gave away CDs to my students, and guess what? when Unity came out all of them gave up on Ubuntu. All of them. Gnome Shell got a similar reaction, no way people could feel comfortable with the new UI Paradigm.

And what was Gnome/Unity devs' reaction to people complaining and explaining why they wouldn't use a desktop that they considered a major regression? Utter denial:

- if you were lucky, you were encouraged to "embrace change", to give it time, to give up to your "old" ways because change is good, and there are billions of people already loving GS/Unity; your friends stopping using Ubuntu or Linux altogether? just anecdotal evidence, don't worry about that (as if a UI study conducted on a total of 15 people might be considered something different than 'anecdotal evidence' ... at best);

- when they saw that many users just didn't buy the party line, the "irrational hatred" straw man popped up: some devs singled out the most offensive and inflammatory reactions as "hatred" and brushed off all criticism.

So we've got our own "reality distortion field" in the Linux world, and it ain't pretty. Me, I'm using Linux Mint and Cinnamon at the moment, and following GS/Unity development all the same, just in case they realize that they've really alienated a good part of their existing user base (not holding my breath there), but my Linux evangelizing is done for good, I've had enough of "WTF" reactions.

There you go, file this under "irrational hatred of a previous Gnome/Ubuntu user and fan" if you want.

Rehdon

Opening a terminal

Posted Feb 18, 2012 17:17 UTC (Sat) by boog (subscriber, #30882) [Link] (3 responses)

A friend who is a programmer but a first-time user of linux made exactly this complaint - unable to launch a terminal in ubuntu. As a KDE user, I couldn't even help him. Possibly not quite discoverable enough for the chosen target group of people who don't know what ubuntu is?

What is the answer?

Opening a terminal

Posted Feb 18, 2012 18:20 UTC (Sat) by sladen (guest, #27402) [Link] (2 responses)

Ctrl+Alt+t

The hot key is the fastest way for a power-user; an intermediate speed method is <super> t e r m <enter>. Once running, one can right-click on the Launcher Terminal icon and tick "Keep in Launcher" to make it omni-present. Note that both Ubuntu and Mac OS X do not show a Terminal by default, but one can find it by searching.

Opening a terminal

Posted Feb 18, 2012 18:29 UTC (Sat) by boog (subscriber, #30882) [Link]

Thank you.

"Note that both Ubuntu and Mac OS X do not show a Terminal by default, but one can find it by searching."

This does seem to say something about the ubuntu attitude to *existing* (and loyal) users, who presumably have grown used to the command line. Current users of linux are probably not pining for the "hidden terminal" feature.

Opening a terminal

Posted Feb 20, 2012 11:30 UTC (Mon) by etienne (guest, #25256) [Link]

In my experience, the intermediate is more:
<super> term
then look at the screen to note that the first three letters have been missed and only thing starting with "m" are displayed, then backspace and retype "t e r m" and then check that the first found application is "terminal" and not "terminal - supervisor" so that either you press <enter> or get the mouse and click on the second icon...

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 17, 2012 22:29 UTC (Fri) by tshow (subscriber, #6411) [Link] (11 responses)

Personally, my complaints are:

1) My tech support burden with my dad went through the roof when he upgraded his ubuntu box and wound up running unity. He cut his teeth on windows (well, actually he cut his teeth on an IBM 1620, but that's another story), and was doing fine in xfce, but unity threw him completely. He's getting the hang of it now, but he's no more productive than he was in xfce, and the switch has cost the two of us a lot of time.

2) There seems to be a general trend in UI design towards netbook/tablet optimized UI with a distinctly Mac flavor, which I think is a bone stupid idea for a whole host of reasons, and for which Unity and Gnome 3 are the standard bearers.

3) Nobody designing these things seems to give a fig about anyone who's been using computers for more than a year.

I've posted in some detail about my problems with macish UI before, so I won't go into it too deeply, but:

- window-docked menus suck for:
-- large screens
-- multihead
-- floating focus
-- programs (ie: gimp) that have lots of windows
-- use cases (ie: programming, monitoring, art) that use lots of windows

- maximize by default sucks for anyone trying to use a computer as a general purpose machine, rather than a word processing or web browsing appliance

I do like some of the things I'm hearing about quicksilver-style menus and the like; text is a far more rich form of interaction, and I'd like to see what a modern command-line aided interface would look like. It's one of the few places where it seems like some effort is being spent on long-term users and the technically competent, rather than people who have never used a computer before.

What I'd really like to see is an interface that grows with the user, and that is something the mainstream desktops (all of them, including the commercial ones) fail at horribly.

Apple (well, NeXT, really) had some interesting plumbing ideas with services, which Microsoft seems to be sort of building on with the pluggable pipeline apps for win8, but this is the sort of thing that we ought to have already in Linux. DBus + something quicksilver-esque, perhaps. Something that lets us move beyond the point and cry interface.

I call it the "point and cry" interface because your vocabulary is reduced to that of a baby. You point the mouse at things, hit the cry button, and hope mommy (the machine) knows what you want. Ok, we have right click for context now, but that ups the number of verbs to something you can still count on one hand most of the time. So you have the vocabulary of a two year old at the best of times.

And I guess ultimately that's the problem; I think the innovation effort is being spent at the wrong end, in the wrong direction. For new users, xfce or KDE or Gnome2 are just fine; the user is effectively a baby in a new world, learning how to survive. I've yet to see a windows user who couldn't sit down with any of those and start getting useful work done immediately. What we should be working on is places for those users to go once they get experienced. We should be providing the force-multiplying tools for experts that the other systems haven't even thought of, or don't care enough to implement. For all UI people curse the command line, you can do things in a few tens of characters that would take weeks to do in the file manager with a mouse.

Children grow up. We need to cater to the adults too. Nobody else is.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 17, 2012 22:46 UTC (Fri) by dlang (guest, #313) [Link] (10 responses)

something I heard a long time ago

a GUI interface is like training wheels, the nice thing about Unix compared to Windows is that Unix lets you take the training wheels off (use the command line)

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 2:04 UTC (Sat) by tshow (subscriber, #6411) [Link] (9 responses)

I agree with that, but I think we've got complacent. We can do better than bash, and when I say that I don't just mean some other direct descendent of sh. There have been some tentative steps in the right direction, but the command line has been largely stagnant for decades.

I've got a boil-the-ocean project to fix this, but I haven't had the time to get it to the point where it's worth putting out there.

To give you a sense of the sort of thing I'm thinking of, consider a standard terminal session. You type a command, it executes, produces some output, you get control back. Standard REP loop stuff.

Now, imagine each cycle of the REP loop is in a bubble. A bubble you can tear off and put somewhere else on the screen. Or name, and pipe things to/from. Imagine if when you launch a command, it "backgrounds" immediately, but to its bubble, so it can keep outputting to the bubble without messing up the current command, or intermixing destructively with other commands.

Imagine if you could hook up DBus services trivially as pipeline stages.

And the terminal has a richer vocabulary, so it's capable of generating (and later updating) things like progress bars, or other simple widgets. Or can say "make this mouse-active and launch this program with that as an argument" so filenames and urls and the like can be mouse-active. Or "this bit is block text, so typeset it nicely in a proportional serif font", while having monospace text for all the interactive stuff. Or nesting levels, so you can navigate program output hierarchically.

If programs using the shell had some standard calling convention (like, say, --describe-args) that would output a machine parseable argument tree, so the shell autocomplete would know that -f expects the name of an existing file. And which would provide data the shell could typeset into proper, modern manual pages.

Imagine a robust scripting language and support for interactive function editing, so you can build up a library of useful functions and tweak them as you need. Being able to type simple math on the command line and have it evaluate without having to jump through syntactic hoops.

Imagine being able to highlight something and run a command on it which substitutes the result for the highlighted area; like interactive backquotes.

And have the shell backed by a daemon, so you can attach/detach as with screen, even if the session is on a different system.

That's just a fraction of the plan, really. The whole idea takes pages. The idea isn't to replace the existing desktop, it's to provide a powerful tool for controlling whatever desktop you choose to use.

Ultimately, if I can build it, it replaces bash, man, info, xterm, screen, emacs and a whole host of other stuff. So it's a boil-the-ocean project in every sense. But I think it's worth building.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 3:44 UTC (Sat) by ovitters (guest, #27950) [Link] (2 responses)

Sounds like PowerShell.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 5:58 UTC (Sat) by raven667 (subscriber, #5198) [Link] (1 responses)

As far as the mouse interactivity, sounds like a LISP machine

http://www.cliki.net/Lisp%20Machine%20Videos

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 20:31 UTC (Sat) by tshow (subscriber, #6411) [Link]

I admit to being one of those people who wishes lisp was more prominent than it is. It's a good language family.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 5:54 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link] (3 responses)

Congratulations, you've just described PowerShell. Oh, there are differences from you description, but the end result is mostly the same.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 20:42 UTC (Sat) by tshow (subscriber, #6411) [Link] (2 responses)

Not really. There are some similarities with PowerShell, but (to me, at least) PowerShell doesn't go nearly far enough. From what I can see, it's just a REP loop front-end for the .Net framework, with some shell functionality worked in.

Which is better than cmd.exe was, but it's not nearly enough.

I should write the whole thing up properly and put it on a web page.

As an example, I don't believe you can use PowerShell to create widgets that sit inline in the terminal. It's .Net, so you can presumably create windows with things in them, but you can't (for example) create a scrollable list or a progress bar inline in the terminal text.

I don't believe PowerShell has screen-like functionality, though I could be wrong.

As far as I know, PowerShell has no equivalent to the named buffers scheme I'm proposing, beyond possibly being able to hamfist it through global variables.

As far as I've seen, PowerShell is still expecting to run on a terminal running a monospaced font, and has no way of breaking out of that mode.

I could go on. And I'm not saying PowerShell isn't a step forward in some ways. But I think we can do way better.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 18, 2012 21:40 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link]

>Not really. There are some similarities with PowerShell, but (to me, at least) PowerShell doesn't go nearly far enough. From what I can see, it's just a REP loop front-end for the .Net framework, with some shell functionality worked in.

And very powerful introspection capabilities.

>As an example, I don't believe you can use PowerShell to create widgets that sit inline in the terminal. It's .Net, so you can presumably create windows with things in them, but you can't (for example) create a scrollable list or a progress bar inline in the terminal text.

You can, by working with the terminal directly. But it'll be more like ncurses programs.

>I don't believe PowerShell has screen-like functionality, though I could be wrong.

I've seen utility which allows to detach and attach to a running PowerShell server. Essentially, the 'screen' utility but with some missing functionality.

>As far as I know, PowerShell has no equivalent to the named buffers scheme I'm proposing, beyond possibly being able to hamfist it through global variables.
?

>As far as I've seen, PowerShell is still expecting to run on a terminal running a monospaced font, and has no way of breaking out of that mode.
Not really, posh can be run just fine in complex GUI. It's mostly output-agnostic.

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 20, 2012 6:15 UTC (Mon) by alecs1 (guest, #46699) [Link]

Please do create that page. I would expect that companies that try to make money out of Linux should pay for the kind of work that could advance Linux to new grounds. Ubuntu has finally come up with something very interesting, HUD; but I'm expecting more from them :) (maybe something like you describe).

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 23, 2012 5:21 UTC (Thu) by blujay (guest, #39961) [Link]

That sounds neat. Reminds me of TermKit (google it if you haven't seen it--maybe you guys should collaborate).

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 24, 2012 16:45 UTC (Fri) by pspinler (subscriber, #2922) [Link]

This inspired me to start checking out what one can do with dbus from the command line. Kinda interesting. Thanks for the clue!

For anyone else interested, the tool 'qdbus' seems like a nifty and useful tool. It introspects objects by default, but also allows you to send messages via method calls and reports results.

-- Pat

Lea: The Unity design process (and how you can play a part in it)

Posted Feb 23, 2012 19:12 UTC (Thu) by sorpigal (subscriber, #36106) [Link]

I'm a mostly-neutral observer. I don't use Ubuntu, GNOME, KDE or Unity and haven't for years. I left when GNOME 2.2 made it clear that 1.4's features weren't coming back.

I dislike what GNOME Shell and Unity are doing because I do not feel that these things will appeal to the masses and I want the masses to want to use Linux because I want Linux to be the dominant operating system.

Spirit, fragmentation and superiority do not matter, just market success, and I don't care who has that success. I bought the argument that GNOME 2.x improved in GNOME 1.4 for the average user, even though not for me, and I believe I saw GNOME 2 succeeding, even from the first release. I do not buy in to the argument that says GNOME 3 is a further improvement on GNOME 2 and I do not believe I see it succeeding so far.

With Unity it's the same story with the difference that some of the things that I see as broken are not broken by design but, apparently, due to inattention. The case for mass appeal *now* is a little better and the case for improved mass appeal *over time* is better. Even so I do not believe that Unity improves on GNOME 2 and I do not believe it is succeeding in appealing to the masses, but I believe that eventually it may.

Given the above it's quite dismaying to see GNOME 2, once a bastion of making Linux work for the masses, be replaced by two competing and incompatible successors that do not appear to be capable of appealing even as broadly as their predecessors appealed.

I know my opinions don't count. I don't exist and I wasn't even an existing user, so no one is going to listen to me. I hope, however, that this post comes off as criticism free of irrational hatred.