US CIO/CTO: Idea of Hiring COBOL Coders Laughable
theodp writes "If you're a COBOL programmer, you're apparently persona non grata in the eyes of the nation's Chief Information and Chief Technology Officers. Discussing new government technology initiatives at the TechCrunch Disrupt Conference, Federal CIO Steven VanRoekel quipped, 'I'm recruiting COBOL developers, any out there?,' sending Federal CTO Todd Park into fits of laughter (video). Lest anyone think he was serious about hiring the old fogies, VanRoekel added: 'Trust me, we still have it in the Federal government, which is quite, quite scary.' So what are VanRoekel and Park looking for? 'Bad a** innovators — the baddest a** of the bad a**es out there,' Park explained (video), 'to design, create, and kick a** for America.' Within 24 hours of VanRoekel's and Park's announcement, 600 people had applied to be Presidential Innovation Fellows."
You mean ass. No need for silly regular expressions.
I'm sorry to re-post the same comment from another story, but in this case it seems very apropos:
Agreed. As someone who's worked for the U.S. federal government, the amount of effort required to comply with various directives, even to accomplish the most basic of tasks, is maddening.
For example, suppose you needed to order some laptops for your developers, and some compilers as well. Private sector: 4 hours to shop around, and you'd have the order fulfilled in about 3 weeks. Most of that delay would be for custom builds of the laptops by Dell, HP, etc.
In the government: 20 man-hours gathering competitive bids from 3 vendors who agree to work under the pricing schedule your agency requires. 4 man-hours / 2 calendar days ensuring the order complies with Clinger-Cohen and Section 508 regulations. 20 man-hours / 2 calendar weeks getting permission to place the order from one approving authority. Another month going back-and-forth with another approving authority. Then the order gets placed.
The opportunity costs and labor costs associated with the effort and delays in getting s**t done in the federal government is mind-numbing. When feds get bashed for having, in some cases, more costly compensation packages than the private sector, there's one factor that rarely comes up in conversation: any competent software developer will demand a pay premium in exchange for putting up with this soul-sucking crap on a daily basis.
COBOL is still around because the systems that use it only get rebooted every 10 years or so. People don't realize how much business and legal knowledge is locked up in these programs. In many cases it's more efficient to "screen scrape" than even attempt to get 15 years of collected business intelligence and regulation compliance exactly correct... And all that stuff is MOVING pieces that have to be adjusted every year because laws change.
This is why company ERP conversions fail so spectacularly. Many company systems have a great deal of "tribal" knowledge from long-retired employees hard-coded by long-retired programmers.
I agree with the deadwood issue, but there are also some dynamics that favor having work done by government. The big one is that there's essentially no profit motive. In a well-functioning federal agency, all of the staff are encouraged to "do the right thing" for the people they serve, rather than maximize profit.
Secondly, because it's harder to fire someone from the U.S. federal government than from a U.S. private company, employees may be more willing to report illegal activity, because there may be less fear of effective retribution. Although my confidence in this has been eroded in recent years by seeing less whistle-blower protection than I would have expected.
You people can joke all you like about old languages.. I'm getting paid to use, maintain and write FORTRAN code.
In the past, I have written FOSS in FORTRAN and put it in the public domain. People still download it on a weekly basis.
FORTRAN has gone through 10 updates and code that was written on cardboard in the sixties can work together with OO code from last week.
FORTRAN is the back-end for the NumPy and SciPy numerical libraries. Python is just a fancy way of writing FORTRAN.
And, no, I'm not an old fart (yet), but I can chase you off my lawn nevertheless.
Now go away, or I shall taunt you a second time...