[ale] ranting about new Ubuntu UI

Michael B. Trausch mike at trausch.us
Mon Jun 20 19:17:10 EDT 2011


On Mon, 2011-06-20 at 17:44 -0400, Ron Frazier wrote:
> Here are some other things I hate at first glance.
> 
> * There is a huge bar of icons on the left.  It seems to be a quick 
> launch bar and a task bar combined.  I'm OK with that concept.  I've 
> been keeping my task bar vertical anyway.  What I hate is that the icons 
> are enormous.  They take up a huge amount of vertical space.  Therefore, 
> if you have 5 quick launch icons open and 5 apps open, then the task bar 
> is full.  I don't know what happens after that.  It may be possible to 
> customize it, but there is no obvious way, like right clicking it.  The 
> other thing I don't like is that it's not obvious what are quick launch 
> icons and what are running applications.  The system seems to put a 
> small arrow next to applications, but I don't think it's very effective.

I hate it.  GNOME Shell has a much better way for seeing which
applications are active and even browsing the applications list.  You
can click the retarded little Ubuntu logo at the top corner to get
something that resembles something useful, but AFAICT it is a
half-assed, brain-dead attempt to free up real-estate.  They could have
just put GNOME 3 in 11.04.  But, Canonical seems to suffer from NIHS and
so they couldn't be bothered.

> * The traditional Gnome "panels" seem to be gone.  Or, if they're there, 
> there's not obvious way to access or customize them, including the top 
> panel that appears by default.

Yup, half-assed.  Granted, there isn't a real easy way to do that with
GNOME Shell in GNOME 3 either, but I think that G3 makes up for it in
other ways.

> * The APPLICATION menu bar (for the text editor, for example) appears in 
> the top of the screen panel OF THE OS when you roll your mouse up there 
> and disappears when you roll your mouse away.  It does not appear in the 
> window for the application.  Why the @$$%%$$# do the designers think 
> that I want my APPLICATION MENU to be married to the OS DISPLAY?!  And, 
> even if I did, why the $%#$$##$ do the designers think I want it 
> vanishing every time I move away from it?!  Stupid.  Stupid.  Stupid.

Why they didn't copy the concept from the 30 years of Macintosh user
interface design is beyond me.  I for one welcome the menu not being
contained in the application window.  But the design?  You said it:
"Stupid. Stupid. Stupid."  The menu should display itself persistently,
and the menu should be visible.  Why the hell would someone make a
design that makes you do extra work to accomplish the same thing‽  It is
senseless.

If the menu were consistently displayed at the top of the screen, I
would be much happier.  For that matter, if it weren't buggy as shit, I
would be happier.  But again we have NIHS as the underlying cause, and
they didn't bother to work with upstream to make this a seamless thing.
(I don't want to hear about how they have posted things to the mailing
lists upstream and been universally rejected; they've been universally
rejected because shit of their design honestly belongs nowhere close to
GNOME's core, and the GNOME people are smart enough to see that.)

> * Finally, there is no scroll bar on long display items and no up and 
> down arrows.  If you roll your mouse over where those features should 
> be, a scroll device which can be dragged or clicked magically appears.  
> Stupid.  Stupid.  Stupid.

There is a tiny orange bar.  If you hover over it, you get a floating
scroll bar.  Now that, I *do* like.  It does take some getting used to,
and I think that when you login to the environment for the first time it
should play a little video or something to show you how that all works,
but it is a nice method for improving the real-estate available to an
application window.

I fully disagree with the fact that they did this without any clue to
the normal user that this is what they did.  Half-assed!

> Again, no offense intended.  Others may love these features just as 
> passionately as I hate them.  Anyway, I'm definitely NOT installing 
> Ubuntu 11.04 or Windows 8 (when it's available) unless I extensively 
> test in a VM first.  Probably, I'll just stick with Ubuntu 10.04 and 
> Windows Vista or Windows 7 until they stop patching them.  That should 
> be about 2 more years for Ubuntu and 7 more years for Windows.

I've been using Ubuntu 11.04 long enough to give it a try and become
familiar with the stupid-assedness that Canonical is saying will be the
only UI available in the next LTS and beyond.  I know this much is true:
I won't be using it for serious production.  Count me out, there.

I think that it is time for a new leader in the distribution world.  And
I think that that new leader will have some properties that set it
worlds apart from any existing system.  Alas, I don't see it ever
happening:

  * No RPM based distributions.  No dpkg based distributions.  A new,
    community-oriented, distributed package management system with a
    robust and stable API.  It should involve the use of the BitTorrent
    protocol.  It should be possible to perform most management tasks
    using either a GUI or a command line, but never in more than one or
    two clicks or commands.  It should be visible to the end-user, and
    it should be amenable to the developer, and it should make all
    involved parties happy as pigs in shit.

  * It should have a rolling release, at least alongside of periodic
    releases that the normal everyday user would be expected to install.

  * It should embrace upstream for all software projects, and it should
    make a serious effort to encourage that upstream (even when it is
    stubborn, like GNOME can be) accepts changes.  The people who work
    for the distributor should realize that the distributor's job is to
    take software issued from the upstream and put it into the hands of
    users in a friendly, transparent manner.  The people who are working
    on the distribution should encourage upstream to accept patches that
    increase usability, robustness, and reliability.

  * It should embrace solid platform documentation.  It should be the
    job of the distributor to draw in the users and the developers so
    as to ignite the passion of people to seriously want to help to
    improve things by lowering the barriers of difficulty to doing so.

  * It should embrace the opinions of its users and it should be able
    to communicate those opinions to the developers of the software that
    is within the distribution.

  * It should not be run by a single person or a closely-held group of
    people!  On the other hand, it shouldn't suffer from problems that
    step from DBC type issues, either.

  * It should be a nice system for servers, for workstations, and for
    everyday end-users.  It should have people who will give it the TLC
    that is required in each of those genres, while coordinating
    to be sure that none of them are taking anything away from the
    others.

If I had the resources to do it, I'd hire people.  I have ideas that I
am relatively sure would work.  I'd put up my own money (if only I had
any!) in order to at least test the idea out.

There is a veritable gold mine in the technologies that are available
today as part of the kernel, the core system libraries and utilities,
and the user interfaces that we use.  There is no single distribution
that rocks the socks off of every area of use---at least, not one that I
have seen.  There are many general-purpose distributions that aren't
quite so general-purpose, and there are many highly specialized
distributions.  There needs to be a distribution that is good at being a
general-purpose server (and damn it all to hell, it should be possible
for someone like me to deploy servers with it in only a few minutes time
when I know exactly what I am after!), a general-purpose workstation,
and a general-purpose end-user system.  (Alright, so those should be
sub-distributions, but you get the picture.)  And there really need not
be anything magical to it, other than the fact that the upstream
components should be glued together in a way that is pragmatic and
elegant.  And it should be possible to deploy a network of 5, 10, even
50 systems without having to jump through many hurdles to figure out how
to alter the configuration of 351 packages in order to fit the needs of
your network.  Maybe for corner cases, but not for the average network.

	--- Mike



More information about the Ale mailing list