[Discuss] an "Enterprise" Linux for desktop

[invalid at pizzashack.org (Derek Martin)]

On Sat, Nov 26, 2011 at 09:00:18AM -0500, Edward Ned Harvey wrote:
> In any OS, it's best not to get deeply entrenched into what you're
> calling "advanced" features of the UI.  I'm going to call them
> details.  Whether it's linux, mac, windows, solaris or anything, all
> this stuff is prone to change.  For better and worse.  Learn not to
> be dependent on specific details.

I have a hard time with this argument.  It depends on what you call
"details" and there's always the question of where to draw the line,
AND you need to consider your target audience.  Ages ago when gnome
replaced its window manager with Metacity, there were a number of
features missing that made it really hard to use for me...  It was
lacking in its window manager focus policy options, and didn't provide
tool tips to tell you what the dimensions of your console-based
windows (but GUI windows had them) when you resized them, just as a
couple of examples.  These features have been present in every
reasonable window manager since the beginning of window managers; yet
metacity developers felt it was perfectly fine to leave them out.
Like it or not, still to this day, the majority of Linux users are
Unix nerds who are accustomed to and want that stuff.

> I wouldn't say removing these particular details was a step back

I would.  Forcing users to change the way they work by taking away
a longstanding feature is asinine.  It's one thing if it's just not
implemented yet; it's entirely another if it's intentionally removed.
The gnome developers have a long history of doing this, to the
detriment of their users' productivity and/or sanity.

> Every change that is ever made to anything benefits some people at
> the expense of some others.  

That statement is probably literally true, but not very compelling.
Removing the window manager features I described made the code
simpler, which benefits the code maintainers; but it benefitted the
users basically zero.  Those who don't use it are unaffected, and
those who did lose.  The benefit is small, and the cost is large.


On Sun, Nov 27, 2011 at 10:32:16AM -0500, markw at mohawksoft.com wrote:
> If a feature comes and goes, what does that mean? Why was it introduced if
> it was not understood to be useful? If it was useful, why did it go away?

Excellent point.


On Sun, Nov 27, 2011 at 11:05:10AM -0500, Edward Ned Harvey wrote:
> > From: markw at mohawksoft.com [mailto:markw at mohawksoft.com]
> > Changing existing paradigms is usually a bad practice unless there
> > is sufficient evidence a new paradigm is better.
> 
> The same is true of everything that is being developed by anyone anywhere.

Mark's statement was already stated universally.  :)

> Even if you invented a car that consumes negative thoughts and emits love
> and happiness as the waste product, 

I seriously doubt anyone who could be considered "reasonable" would
object to this, in and of itself.  If you want a definition of
"reasonable" in this context, I'll refer you to the legal concept of
the reasonable person test.  However many good changes can have side
effects that are not so good.  It's usually those to which people
object.

> or even if you were abolishing slavery, or promoting voting rights
> for women or african americans in the U.S. a few decades ago, people
> who are entrenched in the existing paradigm will oppose your
> invention, and you might end up murdered if it's sufficiently
> important or whatever.

These all have obvious negative consequences for a large class of
people.  Yes, all of them.  An altruistic person can argue that they
are "obviously good" and I'd agree, but people are motivated by self
interest, and a large group of self-interested people have a stake in
all of these changes.  Removing a feature from a window manager can
only positively affect the guy who would've had to maintain it.  The
only way it could positively affect anyone else is if that guy was too
incompetent or too lazy to get it right.  [I acknowledge that some
computer science problems are hard to solve and maintain, but given
that many similar programs have implemented said features for years,
it seems obvious that this is not the case here.]

> As you mentioned, if a feature makes it into a product, was there ever
> justification to put it in?  

If more than a handful of users notice its absence, that's
justification.

> And if it's later removed, was there justification to take it out?  

Virtually anything can be justified.  The question is whether those
justifications outweigh the justifications for leaving it in.
Computers are a tool; their very purpose is to make things easier for
the user.  If the question is one of make the code a little harder to
write/maintain vs. make the user's job harder, the user wins in my
mind every time.

> Perfection is the enemy of progress.  Plus it's subjective.

Absolutely, but that is not a compelling argument for making the
user's life harder.  At best it justifies making a cost-benefit
analysis on a case-by-case basis, which is always the smart thing to
do anyway.


On Sun, Nov 27, 2011 at 11:28:45AM -0500, markw at mohawksoft.com wrote:
> Think about the QWERTY keyboard. The Dvor?k proponents made their
> case and lost. QWERTY is no better or worse than Dvor?k and thus,
> the prevailing paradigm won. Rightly so. That's the competition of
> ideas.

Betamax.  'Nuff said. =8^)

> With user interface, we are not getting a "vote" or a choice. They
> are changing it and we have to learn a new one. 

This is one area where commercial software actually favors the user,
ironically...  If the users don't like the UI, they won't buy your
product and you'll have to make changes to keep the users happy.  With
OSS software, the developer's opinion always wins, cuz his only
motivation is (or at least might be) his own self-satisfaction.

> > It's impossible to get everyone to agree that any new change is positive.
> 
> Provide a choice and see which wins.

Hurrah for reasonability.  Of course, there still IS a choice... you
need not run Gnome's default window manager, even if you want to stick
with gnome.  I keep meaning to look at the current state of FVWM in
detail... I have a number of coworkers who use it (some much younger
than myself, to my surprise), and I always did like that WM.  One of
these days, I may just replace my window manager with it.  That won't
completely solve my issues with the gnome devs, but I can probably
live with the rest. :)

There are still barriers to choosing, though.  I used to know what I
needed to do to change my window manager (within the gnome
framework)... I no longer do.

> > Perfection is the enemy of progress.  Plus it's subjective.

Progress is sometimes the enemy of productivity...

> I'm all for experimentation, it is, in the end, how we learn. However,
> "serious" work creates knowledge. Knowledge lasts and contributes to the
> whole. Haphazard work doesn't last and doesn't contribute, its a
> distraction.

I think it's also worth pointing out that what we had when Gnome and
KDE first arrived on the scene were already the culmination of well
over 20 years of development (if you go all the way back to Doug
Englebert's developments in the 1960's) in the space.  What we arrived
at was a common set of functionality which were found useful and
desirable by some of the world's smartest people at the time.  There's
a reason the concept of "conventional wisdom" exists...


On Sun, Nov 27, 2011 at 12:27:43PM -0500, Richard Pieri wrote:
> I've reverted the netbook to Windows 7 for one simple reason: it's
> more usable than any of of the Linux desktops.  

It's sad to say but I've done this even on my home desktop.  I don't
even dual-boot anymore; I just run Linux in a VM.  I still greatly
prefer Linux for getting real work done; but for the everyday
compunting needs, I find Windows 7 is a bit less of a hastle.


On Sun, Nov 27, 2011 at 02:01:21PM -0500, Matthew Gillen wrote:
> There is a point worth considering here regarding exploring new
> directions.  Sometimes when you're trying something new you need
> time to develop it before it can really be judged on its merits the
> same way you might do with 'legacy' systems.  

While I agree with this 100% in principle, what we've seen from the
Gnome devs is a long history of changing things in ways that are not
better (and mostly not worse), just different.  Meanwhile many old
bugs never get fixed, and new ones are delivered due to all the
changes.

> The other point is that forks are crucial part of F/OSS; they are in
> essence your voting mechanism.  Don't like the cabal that is ruining
> your favorite project?  Take the good parts and start your own
> project.

There's a big difference between forking a mail client, and an entire
desktop UI environment.  The task is monumental.  The amount of effort
involved, just to fix a handful of (major!) annoyances with something
that otherwise works fine is unquestionably not worth the effort.
Even if you have the technical skills required...

> There was a long standing feature request in openoffice that was
> enough to make me not use it or advocate using it
> (https://issues.apache.org/ooo/show_bug.cgi?id=43937; my comment is
> #14).  The people in charge didn't like the feature, so they refused
> to implement it.  When LibreOffice forked, implementing this feature
> was one of the first things I noticed different (besides the splash
> screen).  So sometimes forking is the best way to get progress...

But note that this feature had nothing to do with why the fork
happened.  That fork was entirely political in nature, and there was
already a body of developers eager to continue work on the software.
Maintaining such a large project is not a one-man job.  The cost is
great and the benefit is small.

-- 
Derek D. Martin    http://www.pizzashack.org/   GPG Key ID: 0xDFBEAD02
-=-=-=-=-
This message is posted from an invalid address.  Replying to it will result in
undeliverable mail due to spam prevention.  Sorry for the inconvenience.