Is there anything we can do as a group about SCO ?

[Daniel Wolstenholme]

>Message: 3
>Date: 23 May 2003 08:17:02 -0700
>From: "Ted Gould" <ted@gould.cx>
>To: plug-discuss@lists.plug.phoenix.az.us
>Subject: Re: Is there anything we can do as a group about SCO ?

Sorry, but as an electrical engineer who writes software, you're
way off.  You've basically regurgitated the same line of crap
they feed engineering majors in college trying to get them to
take the EIT.  I fell for that crap when I was in college too,
but when I got into industry I realized I just wasted my time
and money taking that test.

>I would have to disagree with you here.  Licensing of software
>developers will provide some quality in the industry.  And is
>rarely about government control.  I would also argue that it
>would probably help free software in a round about way.
>
>Remember, that almost all 'engineers' are licensed through the
>state. And, in some states it is actually illegal to call
>yourself and engineer without a license.  Here is the
>organization for AZ:
>http://www.btr.state.az.us/

What idiotic states would these be?  I've never heard of any
specific ones, though I've heard this legend before.  I'm an
electrical engineer, and I don't have a PE, and I sure as hell
don't need one to call myself an "engineer".  I have a degree
and I work in that profession; that's all I need.

It is illegal in most states to have a company with
"engineering" in the name unless there is a PE on the staff,
however.  Maybe this is what you were thinking of.  Most such
companies are small engineering consulting firms, who do things
like structural engineering, where one engineer looks at your
building plans and makes sure it won't collapse.  This at least
makes some sense from the licensing standpoint, but it's nothing
like the way software is developed, or even how other
engineering fields work.

>To start off, I should say that I am officially an 'Engineer
>in Training' and have the option to get licensed in a couple
>years - so perhaps, I'm a bit bias.
>
>Licensing of engineers provides alot of things to the public
>at large, because in reality an engineer needs to have a level
>of trust with the public.  You need to have faith that some
>hacker didn't build the bridge your driving on or the building
>your in.  You want someone with certifiable credentials
>certifying those projects.

Most engineers don't build bridges.  Professional licensing of
engineers came about because of bridges actually, but that was
in the early part of the century.  Times and technology are
different now.  If a processor fails, and someone somehow dies,
who are you going to blame?  There were THOUSANDS of engineers
working on the Pentium 4.  How are you going to get Intel to
figure out which particular engineer screwed up?  With a project
that large, it's impossible to lay blame on any one person. 
Also with a project that large, it's impossible to make it
error-free.  Any individual engineer makes all kinds of errors
in their work; the organization they're a part of eliminates
most of those errors because of all the checking and validation
that happens.  Do you think there was one head engineer for the
P4 that looked over the entire design and signed off on it?  Do
you have any idea how complex semiconductor design and
manufacturing is?  No one person could possibly even have the
expertise to know all the different fields involved.

>But yet, the pacemaker that you get doesn't have any
>requirements on the software developers.  Heck, the thing
>could run on WinCE and if it failed, all the people working on
>it would be protected by the corporate shell.  Licensed
>engineers don't have this luxury.  They are legally
>responsible for projects they sign off on, and for protection
>of the public.

Ok, what are you going to do when you find the one PE that
signed off on the failed pacemaker?  File a class-action
lawsuit?  Whoopie, you've spent tens of thousands of dollars in
legal fees, just to make the guy declare bankruptcy.  You're not
going to get a penny out of him.  That's why liability is
assigned to the corporations that manufacture and promote these
products.  They have the financial backing, and insurance, to
make it worthwhile to sue them if they screw up (and hence, with
that amount of money involved, it gives you a measure of trust
that they bothered to engineer the product correctly since they
would obviously want to avoid a damaging lawsuit).
This is also why the concept of a limited-liability corporation
is a good one, and has been so successful for the past 300+
years.  Instead of one guy having to worry about losing his
whole life over a mistake, and never bothering to take on a
large project, groups of people get together to achieve large
things, and as long as they don't commit fraud, they only have
to worry about losing their investment or their job if something
bad happens.  Of course, license or no license, anyone who does
something criminally negligent can still go to jail.

In my not-so-humble opinion, the entire concept of licensing
engineers is absurd for most engineering fields.  Almost no
modern engineering project is small enough to be checked over by
one engineer, and no engineers on large projects work alone. 
This is the reason people form companies and corporations: no
one person has enough money or enough productivity to do
everything himself when it comes to significant projects. 
Accordingly, no one person should be held responsible. 
Engineering isn't like the medical field, where usually only one
doctor cuts you open and does surgery on you, and is responsible
for doing everything right.  It's also unlike the medical field
in that engineers don't have malpractice insurance, and sure as
hell aren't paid enough for any such thing.

Companies are already scrambling to move all their engineering
development to India and China because they can hire engineers
there for less than $10k/year.  Why are you in favor of
accelerating this trend by making engineers unnecessarily more
expensive?

>Now, does this mean that all software production will be
>illegal without a licensed software engineer?  Unlikely.  Just
>like it is not illegal to build a bridge in your backyard
>without a licensed engineer.  I don't know what your insurance
>company would think about it, but that's different.  But what
>will likely happen is that 'critical' software buyers (medial,
>defense, nuclear power plants, etc.) would make a sign-off
>from a licensed software engineer a requirement.

Sorry, it is illegal to build a bridge in your backyard without
a licensed engineer.  Ever heard of building codes and building
inspectors?  Even if you wanted to build your own house by
yourself, you have to get your design signed off by a structural
engineer, and the building inspected for building code
compliance.

>Does this kill free software in those fields?  No, not really.
>It provides more of a market for companies like RedHat.
>Remember that the engineer building the bridge doesn't have to
>do all the work, he just has to be aware of all of it and
>certify it.  The same would go for RedHat hiring licensed
>engineers that will certify the packages in the new 'RedHat
>medical edition'.

You think some engineer is going to sit there and read through
the millions of lines of kernel code, or the additional millions
of lines of code in the common userland programs, glibc, etc.,
that are present in Linux?  You're kidding, right?  What exactly
would this achieve?  This isn't a bridge, it's a software
product.  You can't look at it and verify that it'll work
correctly.  That can only be done through testing and
validation.  As a validation engineer I can tell you that no
amount of validation will eliminate all bugs; it's basically a
statistics game.  You do enough validation to reach a confidence
level.  After that you have diminishing returns; it takes
exponentially more validation effort to gain less and less
additional confidence.

And what about all the countless developers that wrote all that
code?  Aren't they supposed to be licensed too?  Most of them
aren't even in the US, lots of them are still in college, and
none of them are going to have any interest in paying some
stupid government agency thousands of dollars per year to
maintain their license so they can write software for free.  The
Free Software movement is about empowering users and developers
to have control over the software they use.  It most certainly
is not about creating an ivory tower where only the elite
licensed software engineers are able to write code.  The Free
Software movement certainly didn't build up so much code so
quickly by artificially limiting who they accepted code from.

>Lastly, another offshoot of licensure would be keeping more
>jobs in the United States.  Because the licenses are managed
>by the states it is nearly impossible for someone overseas to
>become a licensed engineer in the states.

Huh?  Companies are already shipping engineering jobs overseas,
and compulsory licensing would only accelerate this.  Is it
illegal to buy assembled circuit boards from overseas
contractors?  Of course not.  How could they make it illegal to
buy complete software packages from overseas?  That's basically
the way it works for many places: they contract with a company
in India, which has a development staff, to produce a software
product needed by the US company.  How exactly do you think
licensing would stop this?


__________________________________
Do you Yahoo!?
The New Yahoo! Search - Faster. Easier. Bingo.
http://search.yahoo.com