From Daniel Brockman
<brockman@pacificseries.com>

I have worked with computers professionally for 22 years, first as a programmer, and the last eight years as a systems administrator. I like systems administration. The work suits me well. I have a consulting business with satisfied clients and a promising future.

I have concerns about what the SCSA (SAGE Certified Systems Administrator) movement might do to that future. I write this in hopes others will consider these concerns. I have a fear that SCSA threatens aspects of my work that I presently enjoy.

SCSA would, I think, put a fence around the work. I wonder whether I will be on the inside or on the outside. Being outside and not easily able to get in could threaten my future income or force a career change. Being inside could limit me to a narrow range of tasks that I'd have to perform to the letter of a "professional" standard, and that could truncate my latitude, my success, and my pleasure in solving the ill-defined problems that come my way. SCSA could discourage technically creative solutions. SCSA could impose a legal liability on me if I deviate from the standard.

There are social aspects as well. Certification could stratify USENIX, so that only the SCSAs could write articles or letters for [the SAGE section of] ;login:. As an SCSA, my clients might discount my perspective on business issues as impractically technical.

The uncertain rigors of the certification process concern me. In our general society, there are many certificates. A driver's license is a certificate almost any adult can get without too much work. This kind of certificate is not intended to exclude people, but to assert that they demonstrated enough knowledge about vehicles to operate one with some degree of safety. Other certifications, such as for CPAs or actuaries, have a strong economic barrier-to-entry aspect. Whatever the justice or the injustice in it, these certifications link up with legal requirements. Only the few who are certified have the privilege of carrying out particular tasks and the privilege of collecting the fee made larger by the difficulty of finding a certificate holder.

Proponents of certification say it assures the buyer of competence. But I did not choose my suppliers based on their certificates. My lawyer, my doctor, my dentist, my chiropractor, my carpenter, my realtor, and my mortgage lender all came through references. My own clients rely on my references. I don't think SCSA would change that.

I suppose there would be a test or tests for SCSA. I wonder what questions or demonstrations might appear on those tests. When I think on what I have done in my career, and the moments I would rather not relive, and how I accomplished what I did, and what I hope to do in the future, I think my success derives not so much from things I know but from my talent for figuring out things I don't know. I wonder whether a certification test would recognize that talent. I wonder how a test would show my ability to forge productive agreement among contending users. I wonder how a test would evaluate my responsiveness to changing conditions.

More likely, the test would survey my acquired arcana, and I would feel like Gulliver assaulted by Lilliputians who might even prevail. For example, my adventures have rarely led me to encounters with syslog. So I don't know much about syslog, and I've probably forgotten most of what I ever knew. However, on the two occasions when I had to make syslog work for me, I got it working as needed, and when I have to get syslog to do something in the future, I expect I will figure it out. For another example, there is emacs. I worked with emacs about 13 years ago for an extended period, and it worked for me on that non-UNIX system. I don't think I have used it since then. I have installed emacs a few times and I have gotten programmers started with it, but I use vi. I don't know much about emacs, and I am certain I will figure out whatever emacs I might want to know when I think it useful and important to do so. Will the SCSA test include questions on particular characteristics of emacs and syslog? If it does, then I can choose to study hard these topics which will add but marginally to my expertise, or I can choose to go uncertified, or I can study hard on syslog and emacs and get blindsided not knowing what service corresponds to port 17.

Naturally, some bureaucracy will arise around the certification process. There will be a committee to set the hoops through which we must all jump. We will fill out forms, and wait for cyclical processes, and pay fees, and correct the misspellings of our names, and hope it's all done for noble cause.

SCSA might cast me as a servant, not a role to which I aspire. Many people think of systems administration as a service. I prefer to think of it as an arbitration of contending desires, with me as the arbiter. I am nobody's servant. Will SCSA require me to swear an oath devoting myself to everlasting selfless service to the vice presidents of my client organization? Maybe that is a look in the wrong direction. The certification committee will consist of people who think SCSA is a good thing, otherwise they would not be on the committee. How many of them will see in it a fulfilling way to impose their own version of order? Though their intentions are probably most benign, will my certification serve to promote their fulfillment at my expense?

In summary, SCSA affords me no clear and obvious benefit. If I fail certification, it could force me to change careers. If I pass certification, it could encumber my creativity. SCSA could stratify USENIX, and fulfill the careers of the certification committee at my expense. It could limit my work to strictly technical issues and cast me in a servile role. SCSA may overlook my professional strengths and focus instead on relative minutiae. SCSA could create an economic barrier-to-entry and impose an administrative overhead, without affording practical marketable assurance of competence. These are my concerns.

I think SCSA may come inevitably. Many people want it. I suppose they think it will improve their lives and their future incomes. So, I suppose they will keep at it until SCSA is a reality. I am optimistic about my own future, and I don't see that SCSA could improve it. When SCSA comes, how will it change my life?

Three to the SAGE Editor

From Bruce W. Mohler
<bruce.w.mohler@saic.com>

I wanted to drop you a note to say thanks for letting me get my article on system summaries published in ;login: (April 1999). To date, 51 people from all over the world have expressed interest in helping with the project and I'm hoping to package up and send out the source code today.

Again, thanks!!!

From Robert Yoder
<ryoder@tci.com>

I must take issue with your statement in the June 1999 ;login: "A degree is not required to be a Professional Engineer."

Doing some research, I found, to my surprise, that this statement is technically correct, in California (see <ftp://leginfo.public.ca.gov/pub/code/bpc/06001-07000/6750-6766>).

But for most of the nation, it is false (see <http://www.ncees.org/>; choose "Engineers & Land Surveyors," bottom). Here's an excerpt:

Licensed professionals have satisfied a number of important requirements on their way to professional licensure. While the requirements for licensure do vary somewhat from state to state, there is a great deal of similarity as well. In most cases, a successful candidate has completed a minimum of an undergraduate education and has successfully passed the Fundamentals examination. Most jurisdictions then require that the candidate has four years of experience working under the supervision of a licensed professional, followed by the candidate taking and passing the appropriate Principles and Practice examination. It's very important for you to know the requirements for licensure in your state or jurisdiction. Please contact your local licensing board for these requirements.

To look up info on individual state laws, see <http://www.ncees.org/engineers/licenseboard.html>.

Choosing Texas as an example (<http://www.main.org/peboard/law.htm>), here's another excerpt:

Section 12. GENERAL REQUIREMENTS FOR LICENSURE. (a) The following shall be considered as minimum evidence satisfactory to the Board that the applicant is qualified for licensure as a professional engineer:

(1) graduation from an approved curriculum in engineering that is approved by the Board as of satisfactory standing, passage of the examination requirements prescribed by the Board, and a specific record of an additional four (4) years or more of active practice in engineering work, of a character satisfactory to the Board, indicating that the applicant is competent to be placed in responsible charge of such work; or

(2) graduation from an engineering or related science curriculum at a recognized institution of higher education, other than a curriculum approved by the Board under Subdivision (1) of this subsection, passage of the examination requirements prescribed by the Board, and a specific record of at least eight (8) years of active practice in engineering work of a character satisfactory to the Board and indicating that the applicant is competent to be placed in responsible charge of such work. . . .

From Bob Moneymaker

<Moneymaker@aiso.com>

What a refreshing point of view you expressed in your column [of June 1999]. To me your point made an awful lot of sense. It was interesting that you contacted the Board of Registration. I spent 13 years as a member of the surveying and engineering profession in California and the surveyors were going through the same wailing and gnashing of teeth that we are engaged in in the profession of system administration. How that debate came out or if it was even finished I don't know.

I do know that I hope more and more reasonable voices such as yours continue to surface so that we, as system admins, can continue to progress in being recognized as a separate group worthy of distinction in the IT industry.

Thanks again for your sensible comments.

Flavors of *BSD

From Tim Daneliuk

<tundra@tundraware.com>

In the latest issue of ;login: Rik Farrow asked for comparison/contrast of the *BSD variants. Silly question — everyone knows God created the world using BSD and edited the relevant scripts with emacs. All other systems are infidel by nature and lesser by definition. The BSD variants are merely minor theological schisms along the path of The One True Way.

(rabid, er, I mean, avid FreeBSD user ;)

Rik Farrow replies:

Tim:

Actually, I am interested in the minor differences between variants, and why they are there. Infighting between different UNIX factions is one of the reasons for Microsoft's success with inferior operating systems. Ideally, there are differences, and those differences lead to best-of-breed feature sets in *nix. But what I would rather not see are incompatibilities introduced mainly to make it difficult to port software, device drivers, and system administrators between the various *nices.

So, I am really looking for people who can explain to me some of the history behind there being four different versions of BSD today (net, open, free, and incorporated), how they are different, and how this helps us (or if this helps us).

Anyway, thanks for your letter.