Three 20-Year-Olds Make Their Own WORKING Obamacare Site – In Just Three Days

Demoblican
Some of these comments show that most people have no clue what goes into a website of this proportion. This is not Facebook or Twitter where users gradually increased over the years. This is a site where when it opened on day one, it was hit by over 1million visitors! No company in history can say the same, as when FB and twitter first started they had merely 100s of users at the most. The first time something is done, there will always be complications.

Just know, if you do not have years of an extensive background in web programming and design, you DO NOT know what you are talking about regarding this.

rubbish.

it's 2013, if you build a site that cant handle a million clicks when you've been told it will have a million + clicks on the first day you don't deserve the job. having a buggy functionality is to be expected, but not handling 1% of Facebooks total userbase? HA! Give me a break!

Loads of game devs have had successful launches with 10 million people logging into a resource hog known as a video game at midnight. They learned from the stories of people who messed up and planned ahead then scaled back once they knew what they needed.

Same thing with a website. Except you don't need millions to do it, like a video game does.

I wondered myself about why there were so many problems with the website. I am a web developer and programmer myself, and I understand the amount of scripting that would need to be done, but I cannot wrap my head around why this is so difficult for whoever is doing it. There are going to be many different scripts that need to talk to each other, but from what I know of the site, there is nothing all that complicated about it. A website can only be so complicated. There is so much info though that it could take a while, but I mean come on. I would be leery of believing that the site these two guys created is on par with what the government is wanting as well.

reply to post by christoph


I don't think you can compare developing games that network 10 million people to developing a webpage that has to pull private information from 1000s of insurance companies, private information from millions of Americans from multiple databases, and make all these different instances talk to each other seamlessly and securely. Games are based on a platform that has a beginning and an end, and is no where near as secure as you may think. These games get high jacked all the time by jr hackers.

The webpage isn't the problem, it's the entire infrastructure of the website's functions. In a gaming environment, millions of users connect to 1 area, the game's server. In the ACA scheme, the site has to pull information from 1000s of different servers who only permit it to access certain files (permissions and certificate nightmare) and send that information to the ACA server then the ACA has to push this information back to the end user with the utmost accuracy and security.

Not as easy as you may think. However, I do agree that they knew what had to be done and could have planned better. But how do you plan for a project in which the technical aspects are the 1st of its kind??

Okay Okay, I felt the uncontrollable need to jump into this thread to explain that a site like this -IS NOT- simple... at all...

I AM a professional Web Application Developer and I know every trick in the book. I have worked for multiple fortune 500 companies on their web sites and systems, and let me tell you, the system they are building is not simple at all. I have also worked in the Health Insurance IT area, and let me tell you, most of those health insurance companies data warehousing and underwriting processor companies are running on COBOL based batch systems decades old.

So let me start with the cost. While true, it is a lot of money for a website, that is -not- where the majority of the money goes. The majority of the projects expenditure is in the requirements gathering phase and in initial planning phases.

You need to know "WHAT" to build, before "ACTUALLY" building it. Requirements gathering is by far the biggest problem and expenditure in most projects I have worked on. What will go into the site? How will this part work? Where will we get that info from?

It is THESE failures that cause the screw-ups you see in the site. BAD Requirements lead to BAD implementations!

So lets just get that out of the way now.

Now on to these 3 guys. While its cool and fun to say they have made a better website in however long, the truth is, they didn't do #. They simply tied into some REST API's or services provided by the real beast that is healthcare.gov.

Anyone can make the front end of a website look pretty, be functional, be optimized, and be clean looking. But where the real power comes from is the BACKEND SERVICES!

Lets make sure that is very clear. Backend services are what make or break a web application. If your backend is slow, you're screwed. If your backend isn't properly cached, you're screwed. If your front end isn't properly cached (think Akamai or edge server includes for you techies) you're screwed!

The backend is what drives the system and what is most likly the pieces that are failing right now at healthcare.gov.

The worst part? Chances are, the healthcare.gov backend services are probably relying on other backend services from health insurance companies. Having worked in that area, I can promise you, those health insurance backend services are probably slow themselves! So chances are its not even the fault of the healthcare.gov services, but rather the services they are hitting.

So lets get that cleared up right now.

Also, since you can pretty much cache the entire site these guys made in Akamai, you can take an incredible load before tipping the machines, because the machines are never actually getting hit! But when it comes to having to deal with realtime user data, and realtime un-cachable quotes, you get hit, and the healthcare.gov site got hit hard.

Now, don't get me wrong. Heathcare.gov SHOULD have done a load test before going live. They should have used Keynote, or HP Loadrunner, or at the very least some internal AB loading to see how far they can push the systems. This was a failure on their part. Also, they should have locked the requirements for phase one and stuffed new requirements into another iteration. But they let them push new requirements last minute into scope and that is fail two.

I would go so far as to say that it isn't even the fault of the programmers themselves. Most times it is the fault of the project managers, program managers, product managers, and QA for letting crap get through.

It really pisses me off when I hear people who have no idea what it takes for a true enterprise system to work correctly, and suggest that two guys can build a system better in three days.

YOU HAVE NO IDEA WHAT YOU ARE TALKING ABOUT!


/rant

Wow. And here, at home, I thought we were the only ones who would misappropriate money towards something so simple:
R140m for gov't website

(^^fully understanding that all govs misappropriate taxes. Oh, and don't get me wrong, I state a website is 'simple', understanding the complexities, though not partial to the knowledge.)

I understand the argument that 3 people making a website as is needed for this health care service is over-simplified, but being informed on the amounts spent raises eyebrows.... unfortunately those brows are already distracted by iWhatsits and t.g.i.f.'s.
...I mean really, that's 6 billion South African Rands, for a website!! I don't think the typical keyboard scrawl, "aifgjlagjaglawoiivgaihasg" could emphasize or articulate the w.t.f.-ness that stirs within me.

Where are the people asking questions? Petitioning, parading, protesting; about this issue and others - asking WHY, demanding action.. It is so frustrating how these bugs in office scamper behind dark laws and taunt society.

arg!! ranty rant rant - let's start a country

Oh, out of curiosity, is there an analysis or estimate on how much the obamacare site, or one like it should cost? I dunno much, but screw government tenders and contracts to big business (in most cases)... there are people needing employment who could surely benefit from jobs such as this, and not that they're worth 'less', but wouldn't expect 100,000's to millions to complete.

The student financial aid site pulls from most of the same government systems, and guess what. It works.
This web site is nothing new or innovative. Name one technology it uses that is not used by a hundred other large sites.

The discussion is getting sidetracked from the fact that it is way over priced and over complicated. It also proves, yet again, that governments can't run much of anything correctly.

They had three years to work on this thing and it sucked so bad they are going to have to re-write large portions of it.

PlatinumShatinum
The student financial aid site pulls from most of the same government systems, and guess what. It works.
This web site is nothing new or innovative. Name one technology it uses that is not used by a hundred other large sites.

The discussion is getting sidetracked from the fact that it is way over priced and over complicated.

Was it?

In my previous posting I said it was ridiculous to say that 'the crummy website has cost the American citizens almost 700 million dollars'. Even the OP's own quoted sources believe it's more like 90 M$ - quite a difference.

But that's still a lot of money, I agree.

But the 90M$ did not buy Americans just a crummy website - it bought them an entire system, including all costs made to ensure all was done according to laws and regulations, hiring staff, buying and housing hosts and other infrastructure, building and buying and connecting a whole stack of applications etc. etc.

As a number of my fellow IT savvy companions have underpinned, you can't compare the complexity of the system with the wrapper on exisiting functionality that these teenagers built. The healthcare.gov site is only a small (but admittedly important) piece of a HUGE system that consists of many interconnected servers over many networks, running many applications. To get an idea of that system (and an opinion on what went wrong, and not a bad one in my professional judgement), read for example this article.

That being said, I share the opinion that it is a lot of money and that it is time that a more detailed specificaton of what costs were made for what would be in order. And as the authors of the quoted article state: a postmortem of the failed launch.

And of course, it should have NO CONSEQUENCE whatsoever for the ACA itself. This quote for the article says it all, IMO:

For example, some have claimed that problems with the website indicate that the Affordable Care Act is a bad idea and won’t work, but that’s a radical distortion of logic. This claim is like saying that a problem with an automatic sliding door or a broken cash register at a grocery store indicates that the grocery store (and better access to food) is a bad idea and won’t work.

PlatinumShatinum
The student financial aid site pulls from most of the same government systems, and guess what. It works.

So, the back ends DO work then? But .. aren't they part of the system too? If a couple of 20 year old guys can build a working front end for those systems in just a few days, I think that is proof the Government didn't do such a bad job after all...

This web site is nothing new or innovative. Name one technology it uses that is not used by a hundred other large sites.

That in itself only proves that the designers made sensible choices, by not risking the use of 'innovative' but untested techniques. So, what's wrong with that?

The discussion is getting sidetracked from the fact that it is way over priced and over complicated. It also proves, yet again, that governments can't run much of anything correctly.

They had three years to work on this thing and it sucked so bad they are going to have to re-write large portions of it.

I do agree that the quality of some parts of the system seems to be (very) flawed. I do agree that it seems to be an expensive system. I do agree that you folks over there in the US of A are entitled to a full explanation, and public post-mortem. However, I don't agree that it is a reason to doubt the validity of an act that the Americans choose to have. And I feel that stating that "the crummy website costs us 700 M$" is just plain and simple outright lying.

Americans wanted this Act. Well, at least more than half of them, such is democracy. The poor democrats had to live through those crazy Bushridden years too, and they survived. You will too. And might you occasionally suffer from stress, rest assured that your doctor now is affordable :-P..

reply to post by ForteanOrg


---

I'll do another foray into this discussion by offering up the LIKELIHOOD of
the CGI Federal developers NOT referring to or ASKING about system loads
experienced my Massive Multi-Player Online Games such as World of Warcraft
where user logins, system preferences and on-going user data are changed
on an almost DAILY basis AND where logins at specific times of day can
get into the 400,000 per minute on MULTIPLE SERVERS AND MORE!!!!

That's over 6600+ logins and Game initialization streams PER SECOND -- PER SERVER!

You need really FAST Oracle or IBM DB2 Back End Servers in MULTIPLE locales
THEN you MUST have MULTIPLE OC-48 communications lines or FASTER just
to handle the logins and game play and THEN you need failover systems
to handle the inevitable power outages, hardware failures and
rolling/continuous disk image and user file backup processes.

The main user interface is the EASY PART!

It's the Back-End communications to, from and between MULTIPLE LARGE-SCALE
DATABASES which SHOULD HAVE BEEN FORCED TO BE CENTRALIZED in maybe FOUR
locations (NorthEast, NorthWest, SouthWest, SouthEast) in the USA.

SO this is PARTLY a lack of POLITICAL CONTROL which means FORCED Centralization
of MAJOR databases which can THEN to be cached and replicated to regional or
state-specific edge servers via high-speed LEASED fibre optic lines.

Each query to a separate insurance company's database results in another 2 or 5 second delay
and for results that can range up to 100 items in some states, you're looking possibly up to
a 5 minute up to 15 or 20 minute delay for a user which is unbearable for an online user's
multi-company insurance quote query.

The Insurance companies SHOULD HAVE BEEN FORCED to send and update their
Master Insurance Illustration tables to the four master regional servers which THEN
can get queried by the state-level edge servers which will present the data to only
residents within the edge-server catchement area. This would save time by
PREVENTING recursive and out-of-area queries from applicants.

If I was CGI Federal, I would have PAID some of that 88 million bucks to Blizzard (i.e. Warcraft Online)
for a few days of systems analyst and programmer workshops and tips and tricks sessions on HOW they
can handle sometimes up to 400,000 users PER SERVER -- PER MINUTE world-wide and applied that
info to the Obamacare system.

Sometimes you have to just BITE your tongue and pride and ASK AND PAY FOR HELP from
someone WHO HAS DONE AN EQUIVALENT TYPE of web infrastructure rollout !!!!!!

That's MY 3 cents!

good points above but Im sure they could of easiy just tapped into existing networks or hell used all that NSA crap.

with that said

$300 million for a website even with fancy battery backups, raid arrays and some extra sophisticated php lol you have got to be kidding me. 300 million is beyond ridiculous.

Good rule of thumb

When you hire someone 1st thing you do is check their work and track record and if your project is complex and mission critical make sure they have actually pulled off similar work.

Incompetence really?

ps didn't obama call kanye a jackass one time years back. Look who's talking boss. All those teleprompters and script writers cant always save your ass. I got a feeling this one is going to finally bite you.