Some of the arguments against licensing for sys admins applies against
licensing for software developers. I.e., The systems are just too
diverse, too complex, too domain-specific for there to be any real
rules that we can all agree on as "minimum knowledge".
(I consider this further evidence that system admins and programmers
are basically the same, just programming with different constructs --
Some build systems with compilers, classes and variables,
Others build systems with config files, servers, and cables,
But we're all just toolsmiths.
)
It's also interesting to watch a field move from an era when licensing
was important, to an era when licensing is irrelevant. I'vwe been told
that in certain states, in certain telephone companies, you're
required to have a licensed Electrical Engineer supervising the
installation of "telecommunications equipment". In the old days, that
always meant purpose-built, -48 VDC equipment with special grounding
connections. That equipment was connected to a global network, and
malfunctions could interrupt telephone service for thousands of
people. But with telcos doing VoIP, "telecommunications equipment"
could mean ordinary servers and ordinary Ethernet cables.
If the risk of interrupting service was really the problem, it's too
bad they don't require licensing for the guys who configure BGP. *duck*
On Mar 15, 2009, at 12:18 PM, Richard Chycoski wrote:
> Mark R. Lindsey wrote:
>> So licensing might be beneficial if there are things about system
>> administration where it's critical that we all know that we all
>> know something. For example, all electricians know that the green
>> wire is NEVER the hot wire. And all drivers know that the person
>> turning right has the right-of-way (unless you're at a red light in
>> Philadelphia, where all the rules are different).
> Computing is a global exercise these days. Do you know that 'green'
> means something totally different under the British wiring standards?
>
> And where does the person turning right have the right-of-way? In
> the two jurisdictions where I have taken driving exams, it is the
> person *on* the right who has the right-of-way, when two vehicles
> arrive at stop signs at the intersection at the same time.
>>
>> For example, do we all know that a default route is supposed to
>> point toward the Internet?
> Not if you want to keep your traffic internal, for instance, in many
> labs.
>>
>> Less trivially, do we all know that the root password shouldn't be
>> 'root'? Can we all depend on everybody else knowing that?
> Does the field of systems administration only cover *nix platforms?
> What would a Windows admin be expected to know from 'root'?
>>
>> Would it be valuable for us to have some way of establishing --
>> among ourselves -- what we all know that we know?
> I'm not trying to be flip with the above responses. The examples
> that you bring were both location- and technology- centric. Should
> licenced 'sysadmins' also be expected to be able to write JCL and
> boot and confiigure zOS (aka MVS aka OS/MVT aka DOS) on mainframes?
> Believe it or not, there are still a lot of mainframes out there,
> with IBM selling more and more of them all the time! (And yes,
> although I haven't done it for more than 20 years, I can still write
> basic JCL. :-)
>
> Cars are required to have some common configurations (steering
> wheel, brake pedal, accelerator, ignition) and some optional
> features (gear shift, clutch pedal) that people are required to
> learn to use, and the road systems are required to have certain
> features (lanes, traffic signs and signals, etc) that are configured
> more-or-less the same within a given jurisdiction.
>
> Computers (the roads) and operating systems (the cars) have no such
> regulated environment, even within a given 'jurisdiction'. Computers
> vary considerably by manufacturer, and - believe it or not - Unix
> and it's near brethren are not the only OSes in town. Are we
> suggesting that licencing be required *only* to operate a *nix
> environment? What about Windows, zOS, and any other OS out there?
> Does a new distribution of Linux automatically make it an
> environment for which licencing is required?
>
> In a technology that is completely unregulated and continuously
> changing, the 'common knowledge' changes continuously as well. Even
> the example of not setting the root password to 'root' - what if the
> box is set up using a different security model, where the ID 'root'
> has all of the wonderful privileges of 'nobody' - and no more? How
> dangerous is it for the password of 'root' to be 'root' then?
>
> There are 'best practices' for a given environment, but are we
> prepared to set up licencing boards for every possible environment?
> What is 'obvious' at your $WORK may be anathema at mine - based on
> business decisions, different architectural decisions, different
> mixes of hardware and operating systems, and even different
> personalities.
>
> These kinds of issues are much better handled through training than
> through regulation and licencing. Until computing resolves into a
> relatively fixed set of hardware design rules, operating system
> platforms, and architectural design 'laws', systems administration
> can only become bogged down or made irrelevant by introducing
> licencing to the profession.
>
> - Richard
>
>
_______________________________________________
Discuss mailing list
[email protected]
http://lopsa.org/cgi-bin/mailman/listinfo/discuss
This list provided by the League of Professional System Administrators
http://lopsa.org/
