My Two Census

Formerly the non-partisan watchdog of the 2010 US Census, and currently an opinion blog that covers all things political, media, foreign policy, globalization, and culture…but sometimes returning to its census/demographics roots.

Dr. Groves calls a software change from two years ago a “late change” in operations strategy. MyTwoCensus says this is nonsense.

Earlier today, the Associated Press released a short article (below) that discusses the Census Bureau’s repeated paper-based operations control system failures. In response to Census Bureau Director Robert M. Groves’ comment that, “the problems stem from a late change from a handheld device system to the paper-based system” I can simply point to a response I just received from Michael Cook, Chief of the Decennial Media Relations Branch at the Census Bureau’s Public Information Office. Cook wrote me, “The change to a paper-based NRFU operation ordered two years ago by then-Commerce Secretary Gutierrez, required us to develop PBOCS in a compressed time-frame.” Now, this makes no sense. TWO YEARS IS NOT A COMPRESSED TIME FRAME. It is a ridiculously long amount of time to use engineers to tweak and test a system to make it perfect. Once again, the Census Bureau’s IT failures are pathetic and unaccepptable in the year 2010.

LOS ANGELES — The U.S. Government Accountability Office says a computer system needed to finish the 2010 census may not be up to the job.

GAO Strategic Issues Director Robert Goldenkoff said Friday before a congressional hearing in Los Angeles that the Paper Based Operations Control System hasn’t demonstrated the ability to meet peak requirements of the census as it seeks to count residents who did not return forms by mail.

Census Director Robert Groves says in a statement that the problems stem from a late change from a handheld device system to the paper-based system.

He says the system has worked well so far but that the agency is not out of the woods yet.

The GAO says the Census Bureau is otherwise well-positioned to finish the door-to-door count, which begins Saturday.

Tags: , , , , , , , , , , , , , , , , ,

15 Responses to “Dr. Groves calls a software change from two years ago a “late change” in operations strategy. MyTwoCensus says this is nonsense.”

  1. HermHollerith Says:

    Stephen,

    Two years may be a compressed time frame if much of it is consumed by negotiations and other communications with IT contractors. Very often the people who manage the developers are computer idiots.
    When a computer idiot govt project manager talks to a computer idiot contractor manager, very little work may get done right.

  2. QC CL Says:

    It is interesting – the PBOCS developers seem like they know how to fix problems but not how to avoid them in the first place. It goes down for several hours and comes back up running better. They’re always doing performance improvements etc. I’d be interested in knowing if they add more server capacity or something like that, or are tweaking the code or whatever. Not that those answers make any difference in the actual operation.

  3. Trying2MakeSenseof2010Census Says:

    Why would they move from a hand-held computer system to a paper system in 2010???!!! The decennial census does not know what it is doing. They are wasting millions of tax dollars in order to ask for millions of dollars for communities? It is completely illogical. Furthermore, in census training as an enumerator a woman from the local census office proudly yelled (they are fond of yelling and screaming) to our class that:

    We [census bureau] already have the answers to the questiionnaires. We [census bureau] just want to see what answers you [enumerators for 2010 decennial census] come up with.

    WTF?

    This same woman who has been employed by the census bureau for quite some time and was an enumerator in the 2000 census bragged to our class about her “style” of enumerating. She quipped:

    I was, well I’ll just say it as it was, a stalker when it came to getting questionnaires completed as interviews. I would wait outside people’s homes after they had refused to take part in the census and would learn their schedule; when and where they took their kids to school; when and where they worked; what time they returned home; where they did their grocery shopping. I would return to a NRFU address as often as ten times in order to get them to complete the interview with me. I refused to take “no” as an answer. You (2010 decennial census enumerators) have to do whatever it takes to get your interview questionnaires completed in full. I’m telling you – whatever it takes people! Just don’t think you can make things up. If you do, I’ll be happy to fire you. I like firing people. Just ask my staff.

    These are the types of unstable, unprofessional, completely unqualified people who run the DC East census office. Our crew leader (an even more unstable, unprofessional and unqualified census bureau employee) if from Chicago. I cannot help but wonder if she’s a little Acorn. At least then it might make a lot more sense as to how she got hired in the first place.

  4. asoftwareguy Says:

    “tweak and test” …like installing a patch from Microsoft or something, right? From what I have read, NRFU sounds like a pretty big operation being conducted in a small amount of time. I’m pretty sure a major assumption change in operations would drive a significant amount of systems change and risk. Yeah they took a risk and yeah they have some system issues and yeah the enumerators are just now hitting the street, but clearly it is not bringing the Census to a grind yet. I’d say they are having some start up issues that will work themselves out rather quickly. I’m reserving judgment until later.

  5. Michael C. Cook Says:

    Stephen,

    Again, here is the Census Bureau’s complete response to your questions about PBOCS–that I provided you last night. For you and your reader’s edification here it is:

    ” The change to a paper-based NRFU operation ordered two years ago by then-Commerce Secretary Gutierrez, required us to develop PBOCS in a compressed time-frame. Outside of a war-time deployment, there is no comparable scaling up of an operational control software system to be employed for a one time use, with limited time for testing. This required a strategy for developing the system in phases and conducting testing while implementing.

    The heavy use of this system began last week, when we produced millions of pages of worker assignment materials in time for the May 1 start of our door-to-door operations.

    PBOCS has also been central to our ability to ensure that we are getting responses from every household, and its management functions are supporting our operations. For example, PBOCS was effectively used to manage the work of census workers as they counted people living in Remote Alaska and updated our address list while leaving a questionnaire for households in rural areas during Update/Leave. All of these operations began on time and were managed successfully with PBOCS,

    We have been completely transparent about PBOCS, and proactively disclosed to oversight committees in Congress the status they system. We have also invited representatives of the Inspector General’s office of the Department of Commerce and the Government Accountability Office to be present daily in our meetings where we are managing PBOCS. These oversight agencies are receiving real time reporting about the status of PBOCS.”

  6. Jen Says:

    Mr. Cook:

    Three months of NRFU work suddenly changed to two weeks?

  7. FactChecker Says:

    So SRM, based upon your knowledge of IT systems, how long should this have taken?

  8. Stephen Robert Morse Says:

    @FactChecker – Never said I was an IT expert, but if I had a multi-billion dollar budget, and I had a deadline of two years, it would take me one year tops to create the software so I could then test and implement the system during the second year.

  9. GS-X Says:

    Michael C. Cook,
    Please invite Stephen Robert Morse to the same meetings the GAO and the IG attend.
    Swear him in, embed him, conference him in, put him in FedTraveler, whatever it takes.

  10. Lerrin Says:

    I worked AdCan last year and worked in a variety of roles this year (now happily retired, though they wanted me to transfer to the office from the field- NO thanks!).
    Used the handhelds for AdCan- which worked GREAT last year.
    Only problem EVER incurred was if the weather was a bit overcast – the GPS locator might put you on the other side of the street, but you could easily rectify it so as to put in the correct data point.
    It was WONDERFUL to just travel with the HHC (hand held computer) & your bag – you could send in your E308 (electronic time card), etc.- easy and efficient!
    THIS year’s tenure involved BOXES of information which took up significant personal space (half a backseat of my car and my entire dining room table), with significant bicep development as a result- with NO concerted effort on my part to do so!
    Not using the handhelds is so ill-advised!
    However, there were many ‘less-computer-savvy’ persons in my AdCan crew who found them challenging-and I really do believe that it was a case of ‘cockpit’/1D10T errors rather than problems with the hardware/software.
    Then again, I suspect that there are significant contracts in place to provide employment opportunities with document shredder companies, paper companies, office supply companies, FedX, etc.- all which become redundant/unneeded if things go digital.
    It’s consistent with my observations that the Census tends to overhire people, and overstate actual tenure times for work. Take whatever they tell you, divide by two, and you MAY work that many weeks, but likely NOT the length of time originally quoted.

  11. techy Says:

    “@FactChecker – Never said I was an IT expert, but if I had a multi-billion dollar budget, and I had a deadline of two years, it would take me one year tops to create the software so I could then test and implement the system during the second year.”

    Spoken like a person who’s never actually tried creating a system like this from scratch.
    You can’t just throw money at a programming issue. Google “mythical man month”
    Putting more people on an issue like this solves nothing and wastes more money.
    What PBOCS needed was time. Time it never had. To be frank, the fact that the kludge is working as good as it is currently is astounding.

  12. HermHollerith Says:

    asoftwareguy, We look forward to hearing your assessments of PBOCS.

  13. MyBadgeExpired Says:

    Hey, look at the bright side! Thanks to PBOCS, I have had plenty of overtime and have been able to purchase new shoes, new socks, a new bluetooth, new headphones for my ipod, and much more. This is amazing vs. the amount of money I was making before without working my life away! In our region, office clerks make $9 per hour. I couldn’t even afford my rent. Thanks PBOCS for making my life a little cushier!

  14. sg Says:

    “Never said I was an IT expert, but if I had a multi-billion dollar budget, and I had a deadline of two years, it would take me one year tops to create the software so I could then test and implement the system during the second year.”

    Echoing “techy”, above: Morse would be embarrassed to utter the line he did, if he had **any** idea of what he was talking about.

    Census has certainly embarrassed itself with most of its software acquisitions. Either DAPPS **or** the fabulous FDCA/HHC fiasco alone — let alone together — should have spurred many, many upper-level “managers” to “spend more time with the family”. It’s hard to understand why, in the year 2010, a distributed system can be brought to its knees by a few **thousand** workers who are entering text and running canned queries. To choose just one example, Netflix manages to stream gigabytes of video to perhaps **millions** of users with barely a hiccup.

    PBOCS, however, is different: It’s an ad hoc attempt to use software to glue together a whole gaggle of procedures and data sources that don’t seem to cohere well. Under the circumstances, two years from requirements to specs negotiation to implementation isn’t unreasonable at all — especially for the federal government. Hell, when I was in industry, the team I was on had a stand-alone project with pretty well-defined specs stretch from an expected six months with six engineers, to three years with twice as many developers. (Just to be perfectly clear: I’m not in any way associated with PBOCS development.)

    So sure, PBOCS is a terrible bottleneck. Under the current circumstances, Census employees are almost like involuntary PBOCS beta testers. But that’s more a consequence of other bad decisions involving other systems, than PBOCS itself.

  15. fedup Says:

    I mailed my form back and have been hounded by census ever since. their regional manager was a byoch when I talked to her on the phone and she threatened me after I told her that I should not have to duplicate my work. That Mr. Groves should be fired or penalized for the inept way the census is being done. Basically they told me in a phone call today that although I mailed it in, they demanded that I speak to a worker. I have witnesses who saw me fill it out and put it in the mail. Why doesn’t the census have any accountability? This bureau is not a good representative of our government or our nation. Our president should not have such lackadaisical people on his staff.