Notes and Translations for the New York Times coverage of healthcare.gov | Hacker News Reader

Notes and Translations for the New York Times coverage of healthcare.gov

7fff.com

39 pointstuke12 comments

So 55 separate goverment contractors trying to integrate incredibly nitpicky set of data and it does not work? I wonder why. Maybe first they should have fixed the process. You can't build nationwide websites same way you build physical things.

You can't "fix the process". The website has to talk to hundreds of corporations, and dozens of federal and state agencies.

They could tell/encourage all the agencies to follow one schema/API

They'd balk if you did that. But they could have been told "Create a wrapper API that adheres to this standard, or your product will not be available on the new national Healthcare site."

Most of the insurance companies would have scrambled to not be left out, because if they missed the Oct 1 date they could have lost out on a huge pool of new business for a year.

This wouldn't work for credit agencies and State systems, but that's a much smaller pool of systems to integrate with.

Yes, that's what I meant. Push the work to the insurance companies.

This is a good rebuttal. That NYT article smelled bad immediately.

The other thing that bugs me is when people complain that Obama should have just pulled in his campaign people. Integrating with undocumented, legacy APIS, and strange law-driven requirements is a completely different challenge than the type of still very difficult but free-wheeling and modern integration the campaign people had the freedom to play with.

Could anyone have made this project succeed? Is any IT project of this size destined to fail?

What private enterprise projects could compare in scope and size? Google's infrastructure? Facebook? Could they have succeeded if they hadn't started small?

I think the size and scope of this project compares pretty well with yodlee.com, which pulls financial records from all of your banks and credit cards and acts as a service for other sites to build upon (mint).

Pick pretty much any private enterprise and you'd have a winner.

Amazon is a good example. If someone in France buys a "making bombs for dummies" book and then logs on a month after buying that book on the US site, guess what pops up? A subtle reminder like, "Here's some recommendations from our Terrorism for Dummies" series you might be interested in." And as far as I can remember, it always been like that.

However, If I were a terrorist and tried to cross the border near say Emerson North Dakota and they stopped me crossing, I could simply drive over to hwy 59 about 60 miles to the East and try again and their computers would have no idea I was just stopped from entering within the last hour.

This is the difference between Public companies and things the Government does. Government NEVER EVER gets it right. How about that huge project called the Office of Homeland Security? Even 10 years later, we still have agencies who can't easily share information they have in their databases. Even though this was the primary reason the office was even created in the first place.

Government NEVER EVER gets it right.

While this too often seems to be true, I happened to see this the other day, which describes a few federal IT systems that apparently were implemented well:

http://www.volokh.com/2013/10/20/big-federal-web-deployments...

Now the world runs on software you'd think the Times could find somebody to intelligently write about it. Or not assume the reader knows nothing -- they don't do that for cars.

I just finished 'managing' the integration of a service API with a web front-end. It was not at all easy to manage and everyone working was in-house.

Healthcare.gov has 50(!) contractors supervised by HHS?

To fix it they are bringing in 'specialists?' Someone should send everyone in HHS a copy of the Mythical Man Month.

I believe the probability of this thing EVER working are pretty close to zero.

Reply on news.ycombinator.com