[clue-tech] Interesting sidux/smxi news
Nate Duehr
nate at natetech.com
Sun Sep 21 01:29:05 MDT 2008
On Sep 20, 2008, at 7:44 PM, Collins Richey wrote:
> On Fri, Sep 19, 2008 at 10:38 PM, Nate Duehr <nate at natetech.com>
> wrote:
>
> I'll try to sort through the torturous logic this once, and then I'm
> signing off.
Maybe we should back up here a bit, Colins. You and I are coming at
Linux with RADICALLY different goals in mind, and your posting turned
a light bulb on for me.
You're out to PLAY with Linux. I'm out to be a professional sysadmin
and EVALUATE Linux along with other options out there for doing
productive work with computers.
Sidux is getting in the way here... back up and view this big-picture
from the 40,000 foot overview level. If my comments about Sidux seem
too harsh, they're not JUST directed at Sidux. Sidux is just one of
the MULTITUDE of distros out there doing what Linux has always done...
mess around.
I'm shooting for a higher understanding of why Linux never makes any
serious headway anywhere other than the server room, when it comes to
real business and getting things done.
For years and years, I've been listening to the Linux community rant
and rave that it's all about Microsoft and Apple and "evil" commercial
company business practices, and that Linux is "getting better all the
time", and "you have the source, you can fix it".
Then reality set in. After years and years of running commercial Unix
and Linux side-by-side, and watching literally HUNDREDS of normal
folks using Windows on the desktop, I realized that the real problem
is that Linux devs simply AREN'T INTERESTED in making it truly better
so that it completely fits what USERS want on the desktop.
I see new distros as just another extension of this... more busywork,
more "stuff going on" but little effort to reach the users or even the
pro admins! MSFT may have been "evil" but they do listen to what
their customers want. AAPL too. Linux, as a whole, just doesn't care
enough to clean house and go after it.
Distros changing release cycles, security patches being dropped
because "that stuff you're running is too old", whole distros popping
up that attempt to take another distros KNOWN TO BE BROKEN stuff and
turn it into a usable "product"... all just seem like busywork that
distracts from any real goals, to me personally.
I spent hours and hours tinkering with Linux too, but these days as I
get older -- my time and effort are limited. I still LIKE LINUX, I
really do... but watching hundreds of folks in another project I'm
involved with get just utterly HAMMERED by RedHat's change to Fedora
and fast release cycles, and seeing how that group of people who LOVED
Linux at RH 8/9 get utterly PISSED OFF at being forced into hardware
upgrades, so they could keep getting reasonable security patches...
just boggled my mind that the "altruistic" folks I thought Linux was
meant to represent in the world... just didn't care.
Watching the mess the kernel devs are allowing something as seemingly
simple as SOUND support become, and then ever distro under the Sun
changing sound modules and daemons like they change underwear, is
shockingly saddening.
I guess what I'm saying is this... Linux's hype engine and the beliefs
of many are this: That many hands make light work. In reality, many
hands just make MORE work. More stuff to sift through for the gold,
more distros all the time specializing in this or that, so much
"diversity" that no one can possibly figure it all out.
In commercial software, developers sometimes are forced by business to
simply get something working enough to cover 95% of what the customers
want, and get it shipped. The constant churn in Linux seems to get it
only to about the 85% level of what most folks expect out of a
desktop, and it hovers there. Millions of lines of code, and hundreds
or thousands of devs... all CAPABLE of working together to go after
the 95% solution -- but can't be "bothered" with that.
The response for over a decade has been the same as yours... "But you
don't UNDERSTAND what we were trying to build!" Yes I most certainly
do. What I'm saying out loud lately, and it's obviously unpopular in
the Linux community is this: What you were trying to build still
didn't move Linux toward the 95% solution! After two decades, Linux
isn't beating ANYTHING commercial on the desktop.
And I'm searching for the reason why. On the surface, I don't CARE if
Sidux or any other distro exists as a "specialty" thing. What I care
about deeply is Linux not having 300+ distros, and devs somehow
learning to work together -- which really SHOULD be more effective if
they're focusing on the BIG distros -- so it can eventually get
somewhere in the real-world marketplace.
I would LOVE to see Linux on my corporate desktop. It's not going to
get there with devs off creating Sidux and playing around with it.
But Sidux isn't the problem and I don't "hate" it, any more than I
"hate" badly designed components on my car. What I think the devs
forget is that changes of core features, changes of distro, changes of
policy, change change change... doesn't lead to anything STABLE that
can really be a rock to be built upon.
Companies have "change management" for a reason. Too much change
knocks the foundation out from under things. Sidux is just "more
change" in the distro "market" to me. And you're just the poor guy
talking about it quite a bit.
I now understand that your goals for computers are FAR different than
mine. I don't want to live AT the computer... I want the computer to
do things FOR me so I don't have to care and feed it on a daily,
weekly or monthly basis unless I choose to buy or build a new
APPLICATION. That underlying change in my attitude toward computers
has been slow in coming, but it's where I'm at today.
Changing something as core as the OS continually, seems counter-
productive to me now. Running twelve different distros just to see
how they behave is best limited to a yearly or biennial weekend or two
"jaunt" for me.
I'm a sysadmin. I get paid to evaluate and eliminate risk and keep
systems running. Linux systems CHANGE too much compared to the others
on the market, but they sometimes DO more... so in the server farms,
they're useful. On the desktop, I don't need/want CHANGE -- I need
the darn thing to fire up, act the same way it did the last time I sat
down at it to learn how it worked, and run the applications necessary
to manage the server farm. The servers make money, the desktop is
overhead and any changes to overhead in business, means expensive time
or money spent to re-test it and certify it for use.
Maybe you'll "get" that I'm truly NOT picking on you, or even really
Sidux... other than that I simply have no need for "yet another"
distro anymore in my life. Wake me up when the distro does something
an order of magnitude BETTER than another.
It's not you, and it's not Sidux that's "annoying" me, it's the lack
of serious increase in quality or features in quite some time, that's
troubling. Linux in the early days "got better" so fast, you had to
try new distros. Now they're all just copycats of the most successful
distros created way back when. Sidux is just a copy of Debian's
unstable tree that kinda works ahead of schedule. Okay, I get it.
That's cool if you want to play on the bleeding edge, but I'm just not
"there" anymore. I don't get paid for trying the bleeding edge, and I
don't get as excited about it anymore.
Part of what "ruined" me has been a multi-year on and off again
relationship with Solaris. It's just rock solid and always works...
Linux on the other hand, has this nagging tendency to release
something every once in a while (SSH on Debian) that brings all
serious work to a halt while you go deal with the sloppiness. It's
"okay" the first time, "annoying" the second time, and downright "this
is beginning to piss me off" the third time...
So I delve into the question: Why? And the answers are
enlightening... because they are: We the linux devs were NEVER
interested in being your partner, Mr. Sysadmin. We don't care if you
like what we release or not. Really. We're here to play.
Okay... that's fine. But I don't have to LIKE it, even on a Linux
list, do I? Really? If I speak my mind, maybe some devs start to
think there's a real "market" for a nice solid stable distro that
DOESN'T allow strange UI changes for no reason, doesn't allow support
to be cut off after relatively short numbers of YEARS, you know...
stuff like that.
If I'm letting my frustration with Linux as a whole be focused too
much on Sidux, or you personally -- I'm sorry. It's not Sidux. It's
the attitude that created Sidux. And all the other "also-ran"
distros. The culture of "let's just change this" without any warning.
I think Linux culture eats some of the good stuff people try to do in
Linux alive. Do you know of a single linux dev, other than the PAID
ones who have to answer to a boss, who WOULD answer to bosses that
said, "let's cut the fat, stop all this work on 300+ distros, focus
our efforts, and beat the commercial software that's out there!".
I for one, don't realistically ever see that happening. But I'd LIKE
it to, as a sysadmin/user. Think that there aren't hundreds of
thousands of other pro admins who don't feel the same way?
What are we to do about it? Put a bug-report in a bug-tracker saying,
"stop changing things!" You say devs listen, but that one would
SURELY be ignored in ANY packages upstream bugtracker.
A more relevant example for DESKTOP Linux might go something like this:
Developer moves a button from the left side of the screen to the right
because he/she likes it there better. No pre-warning, no asking users
where they want the button, it just moves.
Enter sysadmin/support guy -- a very BOLD one -- who files a bug
report: "300 users have to be re-trained here because you moved the
button, could you move it back?"
It won't get done. Snowball that into the real-world numbers of
desktops... 3000, or 30,000 -- or maybe 300,000 users.... affected.
Will the bug get closed with "Won't Fix" and the dev move on? Almost
a guarantee. CHANGE is a devs life, especially a volunteer dev.
Those actually MANAGING the needed resources in training, retraining,
and communicating the changes are completely and utterly overwhelmed
by this, once even say... ten highly used packages do it, twice or
more a year.
That just doesn't scale. That's all I'm getting at. Spinning off new
distros constantly from stalwarts like Debian, just makes that worse,
in my opinion.
I'm just Scully (skeptical maybe to a fault) and you're Mulder (you
want to believe, maybe to a fault). Doesn't mean we can't work
together to catch the bad guys, eh?
--
Nate Duehr
nate at natetech.com
More information about the clue-tech
mailing list