One unfortunate truth of the profession of system administration
is that we wouldn't enjoy it so much if it didn't reward its
practitioners for being lateral thinkers with a predilection for
high-intensity short-duration incident-based problem solving, or,
as we say in the trade, "Triggerhappy psycho-freakazoid gloryhound
powermongers with stress-induced facial tics."
It's well-known that we're a bit unruly. Which brings us to the
topic of management.
We've all heard the "herding cats" analogy with regard to managing
programmers. Managing sysadmins is like leading a neighborhood gang of
neurotic pumas on jet-powered hoverbikes with nasty smack habits and
opposable thumbs. Oh, and as a manager you're a neurotic junkie
puma too, only they cut your thumbs off and whereas all the other
pumas get to drive around on their badass hoverbikes and fire
chainguns at the marketing department, YOU have to drive a maroon
AMC Gremlin behind them and hand out Band-Aids and smile a lot,
when all you're REALLY thinking about is how to get one of them to
let you borrow his hoverbike for a few minutes so you can show those
fools how it's DONE. This is because managers are usually people who proved
that they were handy with a chaingun and were thus rewarded by
having their thumbs cut off and their weapons handed to some punk
I digress. Let us read from the book of Peter [Drucker], Chapter 31, Verse II:
[Note: HE'S ABOUT TO TELL YOU WHAT MANAGERS ARE SUPPOSED TO BE DOING. If
you ever wondered about that, wonder no more. Read on. Oh, and I turned
the original text into a prose poem to make it more Zen.]
A manager has two specific tasks:
[Imagine a fleet of steely-eyed pumas riding shoulder-to-shoulder on
hoverbikes with guns blazing instead of arguing over who should get one of
the new laptops. Not only does the manager have to be the organizing
influence in order to make that happen, but the manager also needs to have
decided to make it happen, even if nobody said it should, if it's the
case that the situation demands it. Which is a little implausible, so
I'm going to drop the Smack-Puma-Hoverbike model, because it's a little unwieldy for precision work. Back to
The first is creation of a true whole that is
Larger than the sum of its parts,
A productive entity that turns out more
Than the sum of the resources put into it.
One analogy is
The conductor of a symphony orchestra,
Through whose effort, vision, and leadership
Individual instrumental parts become
The living whole of a musical performance.
But the conductor has the composer's score;
HE is only interpreter.
The Manager is both conductor AND composer.
The second specific task of the manager is
[Note: This is funny. You may laugh. Drucker uses humor like Buddhist
monks use honey: when all you get is a teaspoonful once a month, you damn
well better appreciate it.]
To harmonize in every decision and action
The requirements of immediate and long-range future.
He cannot sacrifice either without endangering the enterprise.
He must, so to speak,
Keep his nose to the grindstone while
Lifting his eyes to the hills --
Which is quite an acrobatic feat.
Or, to vary the metaphor, he can afford to say neither
"We will cross this bridge when we come to it," nor
"It's the next hundred years that count."
He not only has to prepare for crossing distant bridges --
He has to build them long before he gets there.
And if he does not take care of the next hundred days,
There will be no next hundred years --
There may not even be a next five years.
[Amen. Three pounds of flax. Etc.]
To recap: ideally, managers create organizations to carry out their
plans, and they keep a watchful eye on their resources, especially the
most valuable resource, time. Given that, a few questions arise naturally,
and it is the specific responsibility of a manager to find out or figure
out answers to them:
Why are we here?
Where are we going?
What should we do?
Who should do what?
How do we balance decisions for both immediate and long-term success --
or even just survival?
Most of the time, managers' jobs are defined by the rules, processes
and implicit and explicit expectations of their management chain;
things like doing nigh-meaningless performance evaluations and
firing people who spend all their time surfing porn. And since the
managers are still skilled technical professionals at heart, they
also end up doing bits and pieces of their subordinates' jobs which
are either too hard for the subordinate or too much fun to resist
playing in. And since this takes all available time, nobody goes
looking for trouble in the form of the real work -- the work described
above, which is uniquely that of a manager and no one else. Why,
one might ask, don't more organizations actually coach managers on
the practice of management? (Hint: the managers' bosses are managers
It's the lack of integration and planning that results in the well-known
punctuated equilibrium of the organizational life cycle: A new manager
comes in, clears the decks, slags off the old manager, and gets really
busy. But without a clear set of goals or a sense of the future, work
assignments are made according to whoever's available to do whatever task
needs doing, and so strengths are wasted and weaknesses can become
bottlenecks. Worse, a failure to consider the future means that problems
are solved with brute force when encountered. If the manager were thinking
of the long-term considerations, the problems could be captured alive,
studied carefully, and then slaughtered and disemboweled, so that the
entrails can be studied for signs of what ills the future may hold. As
time passes, lack of planning eventually causes a pileup of problems; the
necessary amount of brute force necessary to keep up is unavailable; and
everything blows up with spectacular force, obliterating big chunks of
productivity, at which point a new manager comes in, clears the decks,
slags off the old manager...
So in summary, I guess one of the reasons we aren't good at system
administration yet is that we aren't good at management either, which can
set a practical upper limit on the amount of time we spend developing
beyond crisis management and Nerf wars. Fortunately, though, management
IS a well-documented discipline, and there exist a few good books to help
fill in that gap. My favorite is the one quoted above:
Management : Tasks Responsibilities Practices
by Peter Ferdinand Drucker
It's 861 pages. Not exactly a Reader's Digest favorite. But it delivers a
complete worldview within which to consider what it is managers do, and
that's nice to have. It helps you keep a sense of what ought to be going
on when everywhere you look it's nothing but pumas and hoverbikes. And, really, isn't that what it's all about?