[Discuss] Fwd: Re: Programming vs Engineering

[Bill Horne]

On 1/22/2012 3:05 AM, Matthew Gillen wrote:
> Licensing is about having well-established / well-known ways of solving
> problems. The problem-space for software is still expanding.  I don't
> see how you could come up with licensing until your problem set is
> stable (unless you take a very small subset: e.g., JBoss development, or
> Win32 driver development).

In computer-related fields, certification has been a substitute for 
licensing in many areas. The problem, however, is that the technology 
has always changed too quickly for any single certification to provide a 
potential employer more than a general indication that an employee is 
able to acquire  a lot of task-oriented knowledge.  The issue, in a 
nutshell, is that there is no consensus about the professional standards 
that should apply to software or computer engineers, and therefore 
employers have found themselves settled for applicants who are more 
likely to do /some/ things well, even if their knowledge locks them and 
their employer into a single vendor's architecture.

I am a Certified NetWare Engineer - for version 3.12 of Novell NetWare, 
which was being shouldered aside by Windows NT at the very time that I 
was taking the seven exams that qualified me to write this sentence. I'm 
also an Microsoft Certified something-or-other: a credential that 
arrived unannounced in the mail after I took my first two exams for 
Windows 2000 certification, just as XP was taking over desktops and 
Windows 2003 (or was it 2002?) was being moved into server rooms. The 
certifications of today have the same problem: employers have learned 
that they are an indicator of skills and training which are likely to be 
obsolete before their holders finish filling out their W-9 forms, and 
job-seekers are more and more leery of being locked into a single 
technology silo that can be shoved aside as quickly as Novell was in 1996.

The difficulty of licensing is that there is no agreement about what 
constitutes "competency", for the same reason that certifications remain 
popular: employers have found out that "experts" are easy to find, but 
that "best practices" are not a substitute for the task-oriented 
knowledge needed to make things actually work. Doctors and lawyers have 
an advantage: the glacial pace of human evolution in the first case, and 
the slow pace of legal change in the later. With fields of study  that 
change so slowly, task-oriented knowledge /is/ generalized expertise, 
and I don't feel it will be possible to agree on licensing standards for 
software engineers unless, and until, the pace of change slows to one 
which spans multiple generations of practitioners.

Bill Horne

-- 
Bill Horne
339-364-8487