Why getting the Obamacare exchanges to work was difficult
washingtonpost.com
washingtonpost.com
Especially considering how the lowest bidder has worked so well in the past[4].
[1] http://www.pcmag.com/article2/0,2817,2404481,00.asp [2] http://arstechnica.com/gaming/2013/03/clogged-streets-simcit... [3] http://gamerant.com/final-fantasy-14-launch-issues-servers-s... [4] http://www.boston.com/news/specials/big_dig_problems/
There are many problems with the form too (e.g., can't sign up if your family has more than 4 people). To make matters worse, the form is down for maintenance. Apparently they couldn't figure out how to scale a single form.
It's hard being a supporter of Obamacare when the execution is being done so badly.
Did they have to build a new data center? Hire specially trained astronauts to run it from it's location on the moon? Sitting in special chairs made from exotic materials they reverse engineered from UFOs? I mean JFC where did that money go? It makes the dotcom days look like an era of fiscal responsibility.
It wasn't simply the website - it's the entire exchange including call centers, advertising, etc.
I would argue that collecting and data mining all the worlds data is an easier problem than integrating (U.S.) healthcare systems.
Finally, the NSA projects have suffered a most spectacular project management failure. They were revealed.
If only Amazon provided some sort of "Web" services, or Google offered some sort of application "Engine". Imagine, if new projects could rent computing resources from these companies, and make use of their scaling expertise?
(Actually, AWS is the default for federal contracts now.)
You're a citizen, hence you're insured and you pay your premium through your taxes. Period. It doesn't have to be any more complicated than that.
Granted, most Western countries these days also have very convoluted systems due to ideologically driven shifts to privatization (few if any of which have made healthcare any cheaper or more efficient btw). But most of them are not as convoluted as Obamacare, and all of them have been the result of evolutionary change, not a big bang attempt to catch up with decades of Western civilization.
Because for browsing, it doesn't seem to be much data. My state has about a dozen regions, and the only other variables are age and smoking status. You could do the whole thing with a static site, using a little javascript at the beginning to direct to the right folder.
You could handle huge traffic with a few nginx servers. Put it on AWS with Cloudfront and you're not even hitting the servers for a lot of requests.
Then send people to a database-backed webapp when they're actually signing up.
Three things.
It's impossible to solve a problem before knowing what the problem is. How can anyone here even begin to consider a solution before we know the problem? It's ridiculous. Maybe they're not allowed to use your favored solution. The solution is probably political, not technical. You can't fix an outdated/flawed policy with nginx.
It's just...weird. Half the posts in any politics thread at HN is people complaining about politicians coming up with pet theories and pet solutions for problems before understanding the problem, but they're so eager to do the same when the roles are reversed.
But the problem right now is, people are very curious about what plans are available. It's public information that doesn't require personal data. The whole point of the exchanges is to give the public clear data on the available plans. Distributing public information is a problem that has been quite thoroughly solved.
The technical people didn't seem to realize that they had two problems to solve, not one. And my frustration right now is that part of Congress is trying very hard to defund the whole program before people are able to gain access to that information, and it should be too late for that.
If someone has screwed up the regulations so badly that this information intended for the public is not actually public, then of course that's where the problem is, but that seems unlikely to me. It would make for a very interesting news story, if true.
Healthcare.gov is covered in HN-popular libraries, software, and toolkits (like Bootstrap, jquery, and Drupal[1]), so the problem is apparently not that they just don't know any better.
It appears they went as far as they could with the latest and greatest tech. Someone somewhere along the chain of command failed to include a "provide any kind of useful information on the plans" provision in the design document. All signs point to this being either politics or negligence, not technology.
[1]I'm assuming on this one since WhiteHouse.gov has been covered in Drupal in the HTML and JS since the Obama administration took it over.
Certainly the design is the root of the problem. You have to "apply" before you get any information. When I say that the team screwed up, I definitely include the people who designed it, not just the programmers who implemented it.
Poorly-implemented technology was at the core, but it happened because of a flawed and entirely political process.
You'd want to adjust prices to account for subsidy but that's a simple function and you could do it clientside with javascript.
really guys, we all know technology cannot fix foolish requirements.
(#) www.oss4gov.org
1) The RFQ/bidding process isn't as logical as you'd imagine.
2) There are layers on top of layers on top of layers of bureaucracy. The government is its own regulator, not the market. There are no shareholders or VCs that can curtail the government's power. Elections and recalls are slower signals by orders of magnitude.
3) government projects fail all the time. You just don't hear about it because then heads would roll, which is faux past in the public sector.
1. We F.d up 2. Were sorry 3. Heres what were doing to fix it
Aieeee! Brain explosion.
If there's a huge layer of requirements, review, and sign-off, to the point that it's impacting value creation, then they're not using agile methods.
Agile doesn't have a unique, firm definition. Best I can come up with is "best practices around iterative and incremental development"
There's no best practice I know of that requires spending tens of billions in new IT infrastructure to service 350 million Americans with some fairly simple data needs. Sure, they may be using _some_ agile methods, like having nice chairs that have those cool ergonomic options, but they're missed the forest for the trees. Reminds me of the old joke "Aside from that, Mrs. Lincoln, how'd you like the play?"
I found the article most uninformative. Nice example: "...These projects so big is that there is a very rigorous security oversight involved and layers of audit, layers of rules. The kind of thing that small start-up companies who are just winging it [don't deal with]..."
No, lots of startups deal with high-security situations. These are known problems and have hundreds of existing players with known solutions. The real issue here is the way the environment for the work was structured, not the nature of the work itself. Such comments are misleading at best, duplicitous at worst.
The problem here is that this is a reporter talking to some industry schmuck all too willing to make apologies for "that's just how things are" instead of actually answering the questions. Since the reporter doesn't know any better, there's no context or follow-up.
My money says the reason there were problems is directly related to a large political system dividing up the spoils of new work and insulating itself from being blamed for any disasters. Since lots of folks made money, and nobody in particular is to blame for any of the issues, looks like the system performed as it should.
The real lesson here for startups -- and it is an important one -- is that you can't use people to solve problems that are so far inside the existing system as to have created lots of excuses and barriers for why things have to be done a certain way. Identifying risk is fine, giving up and saying that there's no other way to accomplish things is not.
The larger problem here is using reporters and MSM outlets to analyze how large groups of people perform complex tasks in nuanced environments. Most of the time they're the wrong tool for the job. They're great for telling stories, how one person perceives the way things are, but not so much at analysis. Most of the time what passes for analysis is simply an assemblage of stories, either on deep background or assembled piecemeal from phone interviews.