If you consider a best practice must be such in all cases to be best (an
absolute), then yes I would agree with you. I just would not put it on
my resume that I never use best practices, ;-), because such absolutes
do not exist. I sense that you have discomfort over the use of the word
"best" and see it as an absolute quality (black vs white) rather than as
a relative quality (more white than black in a gray world).
However, common practices rarely become such without good reasons. The
primary reason being around the service you are delivering. Uncommon
practices tend to have higher failure rates and longer recovery times
for the very reasons you have identified. And even when for good
technical reasons supported by business justification they are used,
will still have these effects. That moves the discussion to one of risk
and risk is about financial loss, more specifically, the probability of
financial loss. For an enterprise, that is the real driver of what
makes a practice best for that organization or not in the long term.
I certainly have used just about every trick in the book in the more
than three decades I have been involved in networking, so by advocating
best practices I am not advocating that uncommon practice should never
be used, an absolute position that I was not taking. That doesn't mean
the organizations with which I was associated did not experience the
effects of doing so. Not everyone in an organization will understand
the impact of change when uncommon practices are used and I have had to
live with the effects on occasion, too. Our technical decisions have
business impact - both positive and negative.
And that gets back to my original reason for objecting to uncommon
practice. Uncommon practices tend to be similar to giving a loaded
weapon to a child to play with. I hope the list will forgive the
hyperbole, but it makes the point. Tragedy doesn't always happen, but
certainly can. That doesn't mean that there aren't professionals who
are perfectly qualified to use loaded weapons and know when to use
them. For those seeking answers to questions here, being networking or
any other category of information, they are many times not in the latter
category. That's why they are asking. Those who seek answers here are
seeking the answers that have positive business impact for their
organizations (and their careers). Advocacy of "common practice" or
"best practice," whichever you prefer to call it, is the responsible
position for those who do know to take.
We are on same page even if the words on it are slightly different.
Harold Grovesteen
Rob van der Heij wrote:
From the knitpicking gallery: rather than "best practice" I would call
it "common practice"
I see some similarity with a recent discussion about allocation of
cylinder 0 as page. We know that CP does not mind and there should be
no issue. Still, for several good reasons people tend to avoid doing
this. And one of the reasons is that it makes people frown and ask
questions each time when there's a problem. You lose valuable time
explaining it (or even change it) each time before you can focus on
the real cause of the problem.
But there may be very good reasons to deviate from common practice.
Some of my systems at home have a different subnet mask for the same
LAN segment; it's not common practice, but in my case best practice
;-)
-Rob (an now back to my cage again)
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
