Free Markets, Free People

Competence, incompetence and the Obama administration

USA Today nails it:

President Obama’s chief technology adviser, Todd Park, blames the unexpectedly large numbers of people who flocked to and state websites. “Take away the volume and it works,” he told USA TODAY’s Tim Mullaney.

That’s like saying that except for the torrential rain, it’s a really nice day. Was Park not listening to the administration’s daily weather report predicting Obamacare’s popularity?

Park said the administration expected 50,000 to 60,000 simultaneous users. It got 250,000. Compare that with the similarly rocky debut seven years ago of exchanges to obtain Medicare drug coverage. The Bush administration projected 20,000 simultaneous users and built capacity for 150,000.

That’s the difference between competence and incompetence.

Yup … and all we’ve seen, for years with this current administration, is exercises in incompetence.

I remember when the word of the day for Democrats during the Bush years was “incompetence”.  They had to work very hard to try to sell it.

Well, the sales job now would be a walk in the park.  Except the parks are closed.

Incompetence – look in the dictionary and you’ll see this administration depicted.

Worst. Governance. Ever.

And that covers a lot of territory.


Tweet about this on TwitterShare on FacebookShare on Google+Share on TumblrShare on StumbleUponShare on RedditPin on PinterestEmail this to someone

25 Responses to Competence, incompetence and the Obama administration

  • So much FAIL…
    Were this the private sector, would anyone associated with this still have a job?
    If a company-grade officer turned in this kind of performance, what would happen to their career?
    Imagine the crash-and-burn scenario of a young military pilot doing something analogous.

  • “Take away the volume and it works,”
    Bullskirt.  Plain and simple.   Volume isn’t what failed the dropdown lists.  This is elementary.
    What Congress needs to do now is back-trail the expenses on this and see what got spent where.
    A competent small programming company could have easily handled what I’ve seen so far, and would have had it field ready in a lot less than the time they’ve had.
    (facilities…different story, though I’d like to know where these servers are based…)
    Someones along the way got a whole pantload of money for pushing papers.  I’d bet the bulk of the software expenditure is in pushed paper.
    I’ll even wager that people who are incensed to find it outsourced to a company in Canada while we’re in the midst of an economic slump will be further incensed when they find
    the coding was offshored to Eastern Europe or India.

    • As you say, dropdown lists and the like won’t fail because of load. So, how long before someone claims that some random workers with a grudge or ideological bias deliberately crippled the system. Maybe the same guys that rig the programming of voting machines?

      • No doubt you are right!  Low level tea party zealot interns caused this problem!
        Canadian tea party racists in Toronto?!!!!!  or Polish tea party racists in Krakow?  or maybe even Hindi tea party racists in Mumbai?

      • Ultimately they won’t bother with a reasonable excuse.   They’re not afraid of us.
        They clearly lied about what they were going to deliver.   They lied about what it would do, how it would work and who and what it would affect.
        The lies weren’t even credible in the first place.  They proposed to create a new system that would insure more people, would cost less, and wouldn’t reduce the type/quantity and quality of care for anyone.  It was all going to be good, green grass, blue skies, crystal clear waters…There’s a lake of stew and of whiskey too, you can paddle all around them in a big canoe.
        A guy who can throw out the remark that premium costs for companies would drop 3000% and not end up with a dunce cap photo on the cover of the NY Times the next day isn’t afraid of anything we’ve got.
        After all those lies, this will be small potatoes.   They’re going to tell us it’s Constitutional, it’s the law, and we’re evil bastards for even thinking of denying old granny and young Sally and her poor baby health insurance.

  • Watching Obama right now is surreal. He calls Boehner to say I will not negotiate. Then calls press conference to say I will not negotiate
    — Newt Gingrich (@newtgingrich) October 8, 2013

    Maybe he’s reading the pom-pom girl’s blog…

    • Who is the “I” Gingrich refers to?

    • Well, he has to keep it in the head lines with these various phony conferences.   Last week his no-negotiation photo-op, this week his no-negotiation press conference.
      Like any pouting 14 year old girl, he’d call Boehner directly to tell him he wasn’t talking to him except we’d never hear about it unless we worked for the NSA

    • It’s like a bad Bond film … President NO in the Spite House

  • There are two different problems to look at in the software side of 404Care. One is simple scalability. Many sites, such as Twitter, struggled with that problem, and early users remember the “fail whale” that came up when capacity was strained. Startups often have this problem, because they slammed something out without worrying about scalability – they only wanted to know if the concept would be successful.
    Scalability in such fairly simple applications can be solved by throwing more hardware at the system, or by identifying bottlenecks in processing and removing them. There are decent tools for finding the bottlenecks, presuming the architecture is based on standard, mature platforms.
    The second problem is that 404Care requires a complex, rules-driven architecture. Such applications are really hard to design and write. My team specializes in such applications because the healthcare world is full of them. The clinical side is bad enough; the business side has constantly changing regulations and contractual terms for claims.
    There are two approaches to writing a rules-driven application. The most commonly used one is brute force. Just encode all the rules in code, with programmers constantly tweaking the code as rules change.
    Above a certain, fairly minimal level of complexity, that approach breaks down. The code running the rules gets too convoluted for anyone to understand. Putting in a new rule then has the risk of inadvertently screwing up other rules, and the combinatorics make testing every potential rule combination impossible.
    The result is buggy software that exhibits seemingly random problems that arise when some new combination of circumstances hits the system. Those problems are often excruciatingly hard to find and fix. You can’t just throw bodies at fixing them. It takes a person a long span of time to be able to change a single line of code because of the potential for interactions. That person doesn’t just need to be a good developer, and understand the architecture of the system – he/she must also understand most or all of the rules, which means acquiring a ton of domain knowledge.
    If 404Care is implemented that way, the rate of problems will decay very slowly.
    The other technique is to use a rules engine. The rules then come out of the database. With that approach, the programmers are just responsible for the generic application, and someone with domain expertise is tasked with creating and maintaining the rules.
    One large pitfall in that case is to ensure that all the rules made it into the database. In my experience, they usually don’t. That’s because there is no way my team has ever found to test and find out whether all the rules are there. Also, the domain experts putting in the rules have to learn to “think” like computers. At first, they don’t, so they leave out obvious things.
    But a well-written rules-driven application makes it possible to quickly add new rules, and to change rule precedence. Testing for rule coherency and validity is usually built in. In such a system, the rate of problems can decay pretty fast.
    The rules engine can be an off-the-shelf engine, or a custom-built engine. Off the shelf is much faster to implement, but because it runs highly generic rules, it’s usually slow in execution speed, and getting rule completeness and precedence is harder. Fixing scalability problems with such an engine mostly comes down to throwing hardware at it, but that usually does work if you have money to burn.
    Custom-built rules engines are really hard to design. Interestingly, they’re not that hard to write – it’s just hard to get the design complete and correct for the problem domain. Well written ones are pretty fast and can be scaled to high volumes, and have built-in testing for rules. So this is the best approach to create highly scalable complex applications.
    On the other hand, if a custom rules engine is badly designed or badly written, now you can’t fix scalability or reliability problems in any predictable time frames. It’s even possible to get to the point that any kind of long-term fix requires ripping the flawed engine out and replacing it – rewriting the rules engine from scratch, hopefully with a better design.
    The team doing 404Care might have a lot of experience in doing large, scalable web sites, and still be inexperienced at doing complex, rules-driven web applications. If the team is doing the first such application they’ve ever done, then they will probably be coping with their mistakes and bad decisions for a long, long time.

    • While I DO understand that many analogies are rather inapposite (i.e., Drudge gets as many hits on a good day as the ObamaDoggle sites got in the aggregate on week one), and allowing as how I am NOT a techie…
      IF I wanted to manage creation of this system…
      AND had the resources of teh Federal government…
      AND had YEARS to do it…
      I am certain I could have guided a team to a MUCH better result than we see.
      But this is just exactly what we have predicted…DMV health care.

    • Heh…they couldn’t get the account creation part right.  That doesn’t require a slew of rules (yet), it looked simple data entry based on my stab at, um, helping the volume on the first day.
      I worked on the legacy Texas NHIC system 13 years ago… was rules based (and  written in a 4GL so the non-techies could do rules changes).   It was ugly, and a former employer was scraping it all and re-writing.   They horror showed it, massaged test data, threw out difficult instances, etc, to make the bench marks for timings.  Last I heard back then the state was a might peeved and, um, took some folks to court over it all.
      If the 404Care (love that) insurance stuff wasn’t up to working quite right, I’d cut them slack, but you can’t even enter your personal information yet.   Inexcusable.
      I don’t buy their lie that have created accounts.  If they had, they’d be telling us about the numbers because they NEED some success and numbers can be spun to sound successful even if the numbers don’t mean what they imply.  They don’t even have numbers yet.

    • Good points.
      I’m not sure the Fed Gov has ever done a large-scale IT project in a reasonable amount of time for a reasonable amount of money with a reasonable amount of defects.  They do seem to be good at throwing a lot of money at a complex problem and having it turn out shitty on day 1.  I’ve been involved or in close proximity to such projects and they were all massive clusterfucks.
      The private sector has had plenty of such projects too, so it’s not unique to government, though they do excel at it.

    • I just tried to create an account, which is a really simple process (name, state, password, email) and it failed after I got the basic info entered.   That’s just shitty design.  It’s not 1995. There are enough enterprises with massively scaled systems that it should not be hard to get right, especially given that the act has supporters amongst some of the technical elites. All they’d have to do is ask for some guidelines.


    Told ya, pwns…

  • You are wrong, Bruce. Obama is very competent. But he’s competent only about winning elections.
    Thus all deadlines for Obamacare were carefully designed to happen in off years, etc.
    Every move shows extreme competence…in one area, and one area only.

  • Take away the volume and it works,

    >>>> Oh well, ok then. Carry on.
    I wonder if that logic would work in other contexts:
    “Take away the lack of money and my checking account works”
    “Sure I smashed his face with a crowbar your honor, but take away the scars and his good looks still work”
    “Take away the bacon-fat clogging my arteries and my heart works”
    The idiocy, it burns. In a just world, this dope would dare to utter that line of crap to the public and two seconds later an overripe tomato would smack into his face.

  • I was at a small business the other day and talking to a bunch of workers, and I said, you know, when you’re at the plant and you’re in the middle of your job, do you ever say to your boss, you know what, unless I get a raise right now and more vacation pay, I’m going to just shut down the plant; I’m not going to just walk off the job, I’m going to break the equipment — I said, how do you think that would go? They all thought they’d be fired.—Barracula

    WTF…!?!?—Lil’ Dick Trumka

    “The Congress isn’t on strike, dummy”—the rest of America.

  • Health and Human Services spokeswoman Joanne Peters said Tuesday the government omitted a window-shopping function because officials first wanted consumers to know the amount of the subsidy they might be eligible for. Those income-based tax credits can dramatically reduce premiums for people with modest incomes, and personal financial information is needed to calculate the subsidies.
    “Our process allows us to show consumers plans with prices that reflect what they will pay with the tax credit they may be eligible for,” Peters said. “Window shopping would not allow for this.”
    The result seemed strange to Maureen Bardusk of Galena, Ill., who’s hesitant about giving her personal information before she can shop.–politics.html
    Yes. It seems strange because it is a lie, front to back.
    There is no inherent need to design the system so that you have to divulge personal information to get any result.
    Matter of fact, you should be able to model several possible scenarios as a matter of course…just like you would in making any buying decision.

    • Well, the good news is their data gathering scheme fell over because they’re unable to gather anything at the moment.