Advice to mainframe programmers: stick with it, and specialize within mainframe

The scuttlebutt on the job market related to mainframe now comes across to me as this:

Ten years ago, if you have worked mainly as a mainframe programmer, particularly in the old-fashioned world of nightly cycles, JCL, on-call support, COBOL, CICS, S0C7’s, AbendAid, etc (don’t we remember this), it was more prudent to grow your background in the mainframe area where you had focused expertise, than to move into client-server, where it’s hard to pick things up in non-linear learning curve mode. Why? For a moment, think about how you work at home. You use packages. You get proficient in what Blogger, Facebook, etc can do in detail, but you almost never need to write Javascript any more, let alone perl script. (Although, if you have to make an executive decision on how to redeploy your own website, it helps to know some things in detail.)

It seems like the contractor market seems pretty strong in focused areas like DB2, especially stored procedures; MQ programming, and certain kinds of applications, especially Medicaid MMIS and state welfare systems. The practical problem is that a programmer needs to be focused in one or two of these areas (not too many areas, because it gets diffused) and build a “reputation” over years – even an “online reputation” – on these areas. That may seem like asking for a lot of commitment from an employer base that acts utilitarian and not too willing to return the appreciation with long term career stability.

Be careful: you can actually quit your job in Twitter

Well, you really have to be careful about what you say on your little microblog tweets, when you can resign from a job via Twitter. That’s what Jonathan Schwartz of Sun Microsystems did a midnight last night, according to this MSN story by Elizabeth Strott, link

Gone, it seems, are the hardcopy resignation letters from the company laser printer.

What if somebody created a fake Twitter account and resigned for you?

The story also reports that sales jobs will increase at Oracle. It’s possible that sales jobs will increase among a number of software companies. For example Pitney-Bowes (including Group-1 software, dealing with address standardization and USPS interfaces like MoveForward) has a page explaining its sales program here. Again we have the question, can techies sell? Should they?

Census jobs: I may bite, but I must pay particular attention to usual confidentiality concerns and rules

Some time in the next several days or weeks, I may consider working for the 2010 Census. I want to set the record straight on some heightened concerns about strict Census confidentiality policies.

The Census Bureau, given its mission, has an unusual need to protect not only specific data items reported by individuals but also certain aggregate date in order to prevent any possibility that a hostile party could use aggregate date to identify individuals in unpopular groups. Because the data is used only in a statistical purpose for public policy, none of it can be made available to other government agencies, including law enforcement.

The Disclosure Avoidance Procedures are explained in the following web link. They may need several readings to be understood.

It’s also good to review the questions on the 2010 census, here to get a feel for how various kinds of demographic data could be summarized by computer programs. Additional “details” as intermediate values could be computed from the responses on each form, following well known concepts in “procedural programming” of computers (common in older languages like COBOL).

The September 13, 2009 entry on my GLBT blog explains, with a Washington Post reference, how Census will count same-sex couples. This will be a new responsibility for the 2010 Census.

However, Census would have to protect data carefully, with suppression techniques like those mentioned in the above reference, in communities reporting only a small number of same-sex couples, to preclude any possibility that hostile parties could try to misuse the information. Similar concerns could exist for many other groups.

The same-sex couple problem illustrates how there could develop a heightened concern over accidental employee disclosure away from the job, in comparison to many other jobs involving data, where there is nearly always an expectation or legal requirement for confidentiality.

The legal basis for Census confidentiality requirements resides in Sections 9 and 214 of Title 13, United States Code, link here.

Every employee takes an Oath of Non-Diisclosure (link here) which is legally effective for life. Criminal penalties can be assessed for violation, even after leaving employment. In many private corporate jobs, similar legal requirements to protect stakeholder confidentiality after leaving employment exist, but often only with civil penalties. The Census Privacy principles are here.

Since my blogs cover such a range of issues (due to my “connect the dots” philosophy), I would be concerned about possible prospective concerns from an employer about inadvertent disclosure, particularly of sensitive aggregate information, at an unspecified time in the future, using a “propensity” model of thinking (following a legal model ironically known from the military’s “don’t ask don’t tell” law).

However, I have always had to practice discretion in what I publish in my blogs, which, as I have discussed before, are unsupervised. There are always some ancillary personal situations that should not be disclosed, and there are always some specific matters related to previous employment that cannot be disclosed because of specific confidentiality agreements in the past. I have always honored these. Furthermore, I have been in situations before where someone could imagine scenarios where I might have a political motive to mine available data for some aggregate finding of potential use elsewhere. However, I have never used employer data for my own purposes.

In writing without third party due diligence to an “everyone” market, I would indeed have to practice care, and perhaps more care than ever before. For example, a comment that I believed there were X same-sex couples in some neighborhood of city Y could actually be a legal violation of confidentiality, even though the comment might seem to have political value. However, in principle there is no reason why it would be wrong to continue to blog (non-confidential and non-work-related) other subject matter off the job publicly while holding such a job. I see no direct problem with the Federal Statute (Title 13) or the lifetime oath. If I have direct reports, that could be another matter, and that may be covered in another post later.

Be careful if employers give personality tests when you're already on the job (and when applying, too)

On Sunday, Jan. 31, Lily Garcia ran an article in the Jobs Section of The Washington Post about personality tests at work. The article is titled “Skip the test if you don’t want your personality pegged at work”, link here.

The article fields a question from a visitor who faces a Myers-briggs survey at work (link), which HR promises will be used only in the aggregate. Obviously, she fears being pegged by her employer as a “type.”

Outplacement agencies give these tests to help give advice. But often employers give applicants 400-question true-false personality tests, both to detect deception but also to determine socialability, especially for customer service jobs. AT&T gave the test online to applicants, and would not let applicants who “failed” reapply for six months (back in 2002). TSA gives a test like this to applicants for screening positions.

Picture: Full Moon on a winter morning (no relation to post), when days start to get longer. "The days lengthen, the cold strengthens."

Security clearances and federal jobs, and contractor jobs

Derrick Dortch, of the Diversa Group (link) sets the record straight on how security clearances affect federal jobs in a column on p H1 of the “Jobs” section of the Washington Post today, Sunday, Jan. 31, 2010.

His column today is “You don’t need a security clearance to get a federal job,” link here

When a federal agency wants to hire you and you don’t yet have the necessary clearance, it will make a conditional offer and then submit the required background investigation. You can’t start work until the clearance is granted. In a few cases, as with the CIA, this could take a year or more (and that’s something that the administration should fix, because intelligence services need more brain power right now to connect the dots – badly). Military departments may take longer (but civilian employees are not held to the “don’t ask don’t tell” policy, even if they go overseas or live temporarily in military quarters or on ships).

However the “Beltway bandits” – the contractors – often require that candidates already have necessary clearances, and even hold special job fairs for them (Dice often sends me advice on these fairs by email – and Dice is a good source on these jobs). I’ve always wondered why polygraphs are OK for government clearances but not in court or for ordinary job investigations. (I wonder if future clearances will involve MRI scans for “thought reading”, like “No Lie MRI” ).

By the way, Dice and other similar databases can be fairly pricey for very small headhunting agencies. Economy of scale is definitely advantageous for placement companies using these databases.

How much has changed since 1969 (when we put Man on the Moon)

Remember how I.T. was in the days that we put Man on the Moon? We had the technology to control a mission to an object 240000 miles away, and could control it through space remotely.

Yet, in the late 1960s, for programmers, computing was a tedious and bureaucratic process. I worked for two different parts of the Navy: for the David Taylor Model Basin (or Naval Ship Research and Development Center) on the Potomac River near the Cabin John Beltway crossing in the summers of 1965 through 1967, working mostly in FORTRAN (“Formula Translation”, at one time a competitor of COBOL), and then in 1971 and 1972 at NAVCOSSACT, the Naval Command Systems Support Command at the Washington Navy Yard in Washington DC.
(An electric utility I worked at in 1972 with Univac still used FORTRAN, as did a health care consulting company in 1989, although then it was switching simulations to SAS. I don't know if anybody uses FORTRAN today.)

We would code programs on coding sheets, where specific columns were set aside for specific areas ofd statements (COBOL certification tests still ask about this), turn them in to keypunch (or sometimes keypunch ourselves), get them back, and submit compilation or “load and go” decks as “shots”. Two or three cycles of submission and output (printed and deck) a business day was good turnaround (or you could go to the EAM room sometimes and run it yourself).

People tended to become specialists on one little subroutine of some sort of simulation model. Whole jobs were budgeted based on expertise on one little bitty area – because for some things in space or military applications, the work had to be perfect, and there was not an efficient way to get it done.

Our culture has certainly changed direction since 1969 (“One step for mankind…”) Now people set up businesses on the Web with little programming skill, or whole publishing operations, or invent on their own whole facilities for social networking or gaming. How things have changed.

And then James Cameron gives us a movie that suggests that on other planets, Internets could grow as part of biology – natural “social networking” with bees and ants have already!

Second picture: both the Moon and Mars.

CNN outlines 22 best companies to work for, each with over 500 openings (also with Fortune)

Check out CNN’s Best 22 Companies to work for, with a total of 87000 openings, each company with over 500 openings, link here.

The companies vary from major consulting and accounting firms, to information technology, to retail. Many of them have a major presence in the Twin Cities (I lived in Minneapolis from 1997-2003 and it strikes me how many of these do have major operations in Minnesota.)

Each page includes a blurb from the company’s HR on what it looks for now.

Google was on the list, and I was particularly impressed with the idea that the company says, “Candidates should bring their whole selves to the interview and not try to fit some mold that they think Google wants” . It sounds like they want candidates who can “connect the dots”. I relish their finding this comment on my blog if I ever have an interview with them.