No, HealthCare.gov Doesn't Require 500 Million Lines of Code

itwbennett writes: "Half a billion lines of code for a transactional website — more than five times as much code as that behind OS X — just didn't pass the sniff test. But just how many lines of code does it take to generate HealthCare.gov? This question came up on Reddit again last week and it appears that we may now have an answer. One commenter who claimed to have worked on HealthCare.gov as part of the post launch clean-up crew at the end of 2013, provided counts of the lines of code behind HealthCare.gov, broken down by programming/markup language."

Comments

646 lines of Perl?

[angularbanjo]

That much Perl?
That's probably the whole app there, with each line being around 10,000 characters of obfuscated self-referencing goodness.

The rest is just quotes from Tolkien.

So now we're trusting blogs face value?

[Whatsmynickname]

Has this been verified? Is this poster even supposed to be posting data like this? Main news channels now repeat blogs true or not as facts, et tu Slashdot

Re:So now we're trusting blogs face value?

[neokushan]

In fairness, it's no more unreliable than the 500million+ lines of code, claim. And somehow much more believable.

Re:So now we're trusting blogs face value?

[Ksevio]

Of course not! We're trusting blogs that cite reddit comments. Since the comment got "Reddit gold" it must be trustworthy.

Re: So now we're trusting blogs face value?

[Wdomburg]

In other words, it's not only on the internet, but it's been vouched for by anonymous sources. It clearly must be true.

47449 lines of Maven (pom.xml)

[Tailhook]

Hmm.

Could we, perhaps, use some of the techniques that people have speculated about for deflecting space rocks and, instead, guide one into Earth deliberately?

Re:WOW

[thedonger]

Forget about the number of lines of code. I work for a U.S. company that builds healthcare.gov type web sites and the reporting back end for large companies. The estimated price tag of the front end ($150 million or so) is about 20 times what the tax payers should have paid. Add in the back end reporting to the insurance companies and billing, throw a call center in at least two different time zones, main and backup datacenters and instead of the full price tag ($600 mil?), let's say at the high end $20 million for the whole thing. Ongoing administration costs maybe in the 7 digits per year. The whole thing was a sham to get votes and fill the coffers of some cronies.

Count faster

[Impy the Impiuos Imp]

I don't see how they could have reported 500 million lines of code in the first place. The Congressional authorization to spend $30 million to study the best way to count lines isn't even out of the House committee yet.

2.4M Lines of Java?

[hondo77]

To be fair, most of that is probably getters and setters.

Re:WOW

[ArcherB]

Finance guys are so cute.
I was an IT guy so....

For example a retail bank needs two tables in it's accounts database. One for the account, a second to record the transactions.
The DB needs a customer table (name, address, phone, address, ect), transaction table, account type table, account table, interest rate table, payee table, payroll tables (complete with more account data from other banks, employee names, etc) etc. There's a LOT of data involved, and this still doesn't include the cutesie stuff banks throw in like customer preferences.

The database may be queried by other databases (ie: the guy approving loans), but it is not actually a part of those databases.
Actually, different systems maintain different databases. For example the Internet Banking side will maintain it's own database. the ATM side will have it's own side. Then there's the credit card system, ACH systems, wire systems, the core system itself and others. All of these systems must interact with eachother. For example, the a customer may log into the Internet banking side, which will have to hit the core to get the current balance, EOD balance from yesterday, unprocessed transactions, processed transactions, interest rates, any messages from the bank, and so on. It also has to be able to inject transactions such as payroll into the core system, wires into the wire system and so on.
Of course, all of these systems are different. The ACH system uses a flat text file. The core is usually an UNIX based system with a terminal interface. The Internet Banking is probably an Apache Tomcat connecting to a MSSQL system. Then, there is the bank end that is comprised of DB front-ends, screen scrapers, batch files, transaction injectors and so on.

You could probably convince a bunch of PHB-English Majors your database is more complicated because you have six different, totally unrelated databases in the same file, but don't try that shit in front of engineers.
Not just different DB's but completely different architectures. And, of course, different states have different laws. For example, all states that take income taxes have a different method to pay them. Then their are business taxes, both federal and for all 50 states, loan laws, interest rate laws etc.

And there is much much more, but this is getting out of hand. Suffice to say that you have no friggin' clue as to what you are talking about when it comes to everything a bank does, much less when it comes to tying all those systems together.

Compare that to the ACA system which involves user data, finance data, what companies are available per state, what plans available per company, and an interface system to communicate between the handful of ACA authorized insurance companies per state and the back-office system. Many states run their own system. The government has claimed that their system doesn't even keep the data!