This comes after the intense backlash following the squashing of the project's entire contributor/commit history.
I was wondering why the issue complaining about this didn't have any recent updates.
---
EDIT: It looks like some people may have used the Issues system to launch asinine political bromides against Obamacare proper - that were and are completely impertinent to the GitHub project. This could have been part of the reasoning for closing off the repo, but I've never dealt with users basically defacing my Issues system, so I don't know what your options are as repo owner in a situation like that.
> It looks like some people may have used the Issues system to launch asinine political bromides against Obamacare proper - that were and are completely impertinent to the GitHub project.
Seems a fair reason to turn off the issues, at least for the time being. However, it is interesting that a similar abuse of issues [1] is left undisturbed by administrators of the WhiteHouse account when the asinine political bromides are favorable to the current administration.
There's also the aspect of giving contributors their due credit and thanks. As time and furloughs allow, to be sure, but I'm only talking about two paragraphs' worth of giving people who've poured hours into the project an explanation of what's gone on.
I honestly don't care one iota about the concept of being entitled to the FOSS source code; I'm just slightly miffed about the seeming lack of thanks for the people who chipped in. If you will, it's about what's owed to contributors, not tax-payers.
A 'bromide' is the verbal equivalent of patent medicine, a meaningless nostrum or 'thought-terminating cliché' in the modern vernacular; I think you meant 'broadside', which is literally a ship of war firing her entire main battery at once, or metaphorically a full-throated and fiery verbal attack. The More You Know(tm).
This was the first and only commit. So it didn't reflect the actual website anyway. There are bugs in the website such as incorrect email address validation that are not present in this code dump.
Since this is public domain software, I will put a copy of the git repo on Archive.org when I get a free hour.
I've never used a buggier website before. Half the links were broken, the application process just wouldn't proceed halfway through (until I tried again today), I get giant error boxes every other page, and it just plain wouldn't do anything in Chrome.
It says I have to "Review my application details" before I can go further, and sends me to:
Yes, that's the actual link. The variable name "%=applicationID%" is directly in there, and so it just sends me to a broken link. This is one of many broken links I've come across, this one just happens to be an essential step in the application process.
If anyone's interested, I've been working on an issue tracker & locally runnable version of the marketplace in a new repo[1] that was spawned from the former issue tracker on the removed repository.
It has some issues with hardcoded redirects coming back from their API calls, but it's a start.
I originally made it just to simply demonstrate how a few small changes could make a large performance difference, but I've kept it up due to encouragement from some friends and hopes that somebody inside the project will use some of the code or make changes based on it.
It's way below the quality you'd expect from a modern front-end. Poor standards, undocumented code, monolithic sources, hard-coded data all around. Despite using Backbone and other modern frameworks, it looks like something a novice JS developer would concoct, not a project in the millions...
DevelopmentSeed (the guys behind MapBox) designed & developed a frontend site and then handed it over to HHS. That's where the good stuff comes from. Clearly however, whoever received the project from that point onwards doesn't understand things like asset minification or packaging.
My completely random conjecture is that whoever inherited DevelopmentSeed's code does not know how this stuff should work.
Yep, this is 100% real code, taken from the current working site. The initial commit is exactly the code that you would download every time you visited the site (and re-download, and re-download, as they failed to set caching headers).
My hope is to reorganize it into a sane set of Backbone views & rich models so it makes a lot more sense. There seems to be a lot of dead code so pruning that is the first priority. Every function is in the global scope so it's actually somewhat easy to detect what is used and what is not... not that I'd ever advocate throwing everything in the global scope.
I agree that it is much less than you'd think that amount of money would buy. I hope that by bringing attention to it, I can make it clear to taxpayers and government decision-makers that we are not getting anything close to what we are paying for. The system for winning government contracts needs to be completely revamped, at least for the technical sector. I would love to see a competition-style bidding process, where a detailed set of specs are presented and all submissions that meet the spec are paid, and the winning submission is paid more. $92M buys you a lot of runway. I think we could have seen some much better code out of this if small businesses & independent groups were allowed to tackle it.
I've not paid much attention to this, but your comment just hit me right in the face. We spent $650M on this thing—and the three-orders-of-magnitude-too-high price can only be explained by a nonsensical favoritism for firms that can navigate our minefield of regulations and the selection process... and we ended up selecting a Canadian firm?
I am not especially patriotic, though I also don't really think patriotism is a bad thing. Nevertheless, learning that the contract was awarded to a Canadian firm makes me reel. I mean no slight to Canadians, but if we are going to overspend by 100x, let's at least send the money to our cronies in the United States! Just for appearances.
No, that's right wing spin. We spent $94M on this thing. Still absurd, but don't be misled by the spin campaign.
Also your math is off. A $10M price tag wouldn't be totally ridiculous. Consider the cost of a team of 50, including QA, design, product guys, engineers, management, lawyers, analysts, data entry, etc for a year. Add 30% to cover risk.
So, it's somewhere in the 10x too expensive region.
No it's not "right wing spin". The bigger number is total awards to CGI as a company and the $94M is what they got for doing their part of the healthcare system. People are just confusing the two.
It absolutely is right wing spin. Yes, the bigger number is the total awards to CGI. But the story being pushed by the right wing is "we spent $650M for a website that doesn't work", which has confused a lot of people.
$94M for a website is absurd, and reeks of incompetence and bureaucracy.
$650M for a website would be downright fraudulent. It's not a coincidence that it's the latter number that's being talked about.
In the big picture of the mess of the ACA from the bill writing to passage by way of reconciliation, push back of the states, twisted supreme court logic, the president deciding illegally what parts of the law to implement, and now this states' and federal IT debacle, I find someone getting dollar figure wrong as to how much the website frontend cost to be of no consequence.
Your attempt to silence debate by shaming those you disagree with has been noted.
Like I said, I'm not following this very closely, so thank you for the tip that $650M is not accurate. Doing a Google search, the first article I found confirmed the $93M figure. And it was from Fox News of all places. So I'm not sure if the $650M thing is right-wing spin so much as confusion.
In any event, I still think that's 100x the actual value of the web site work. The site strikes me as the kind of thing a decent consulting company could execute for just shy of $1M.
My point earlier, though, still stands. It's surprising to me that the "best" firm at navigating our regulatory maze is Canadian. A credit to them, I suppose.
94M for 50 absolute "rock stars" of course. At nearly a mil a pop for a years worth of sweat, swearing, and extremely undue stress </sarcasm>.
I do have a small bit of sympathy for the obvious peons that did the bulk of the work. If the comments above regarding how this functions is any indication, these are certainly not "rock stars" but more likely intern/severely junior developers. A consulting place like this obviously can't afford to put their best and brightest on this but they sure can take the money as if they did. If they're suggesting this was their best and brightest work, they need to seriously evaluate their "ability" to handle web development of this magnitude. 650M total for yet more projects of this level of trash? Cancel the contract now while you still can and save at least some face.
If the Canadian part bothers you (it bothers me too) note that we have been giving them millions for years since they also developed the online part of medicare.
But you should note, the STATE run systems, are not built by the same canadian company.
Only the states that decided they weren't going to run their own exchanges and refuse federal funds to help people who need healthcare caused the federal system to be created.
Apparently New York and California developed pretty good enrollment systems of their own.
The federal system was only supposed to be a fallback, but it because a primary because of mostly right-wing run states.
The front end was pretty decent using Jekyll and other open source items and I think that was what was all that was hosted in github. The back end was never public nor was it intended to be I think. The back end is the crappy thing. But people keep confusing the two even on HN.
The whole thing is so politicized I'm not surprised that truth and reality are the first to go. I wish HN could be the font of reason in a sea of moronity.
Funny enough, I have experience working with them and I knew they had no clue what they were doing. How did the government not figure it out? -__-
It's the kind of company that is years behind on programming practices and over-charges in the millions because their clients never have any clue of the difference between good/bad software.
They tend to target markets that have almost no real software servicing it so there is nothing to compare it to (eg. commercial insurance).
Government procurement is an incredibly coreographed ritual with a minefield of rules and regulations designed to make the process as human-proof as possible. Unless there was conclusive evidence about CGI's lack of clues, the process would continue with them.
The code that was on there has nothing to do with what people are clamoring about... It was just the informational side. The /marketplace stuff was not open source (nor would it be very useful considering it would be very dependent on the specific stack of private databases and APIs that it taps into).
I wonder if the newer commits would be along the lines of "Shit! XYZ had massive bug, we have done a hacky workaround for the moment" which might not do their company image much good.
That's easy to hide - just write in the commit "Fix issue 65438" and nobody's the wiser :) Of course, analyzing diffs can uncover the truth but if the fix already deployed that's not the worst position to be in.
That was my understanding as well... so if it's out of date, why bother taking it down, and changing their policies in the middle of a PR shitstorm. Seems to just add fuel to the fire...
Where do you get your news? No where near $650m was spent on Healthcare.gov (but, perhaps maybe it should considering the project grew considerably in scope when so many states refused to setup their own exchanges).
So as I thought, you were completely misinformed. The story for that link has been updated:
> Correction: We miscalculated the expenditures related to the healthcare exchanges established under the Affordable Care Act, and incorrectly attributed the total cost of these expenditures.
Another link in that thread (as noted below), contains the actual amount of money:
> The U.S. Department of Health and Human Services awarded CGI $55.7 million to launch Healthcare.gov, its central Obamacare health exchange website. Over the full five years of the contract, CGI could receive as much as $93.7 million.
$650m turns into $55m real quick and I'd bet a lot of fixed costs for things like hardware and software licenses made the actual figure even smaller. That was for two years, so $28M a year or $2.3M a month. Considering the cost of developers alone that doesn't seem like a ton, especially when developers are only a small part of the challenge.
I'm glad I didn't get the contract. I sure wouldn't want to build an exchange to handle many billions of dollars worth of transactions (and be responsible for millions of tax records!) on that kind of budget. Doubly so with talk radio , Fox News and the WSJ watching my every move.
If a startup came up with a way for any American to buy health insurance in one place it would be valued in the billions of dollars (even with a fail whale on demo day).
Yea that article caused a lot of confusion amongst folks.
Personally, I don't think that the $650m has to apply for the one contract (healthcare.gov) for the general act to be quested like the post above ('and we get THIS')?
As for your math - sounds reasonable, but wow my expectations of what $55m can yield are vastly different than how you rationalized it.
It's a lot of money in some ways and not a lot of money in others. It's an exchange that will be used by millions of people who will make decisions to spend thousands of dollars a year... The amount the exchange costs is a fraction of a rounding error in the scheme of things. I'm sure the insurance companies themselves have spent vastly more just getting their systems set up.
I was wondering why the issue complaining about this didn't have any recent updates.
---
EDIT: It looks like some people may have used the Issues system to launch asinine political bromides against Obamacare proper - that were and are completely impertinent to the GitHub project. This could have been part of the reasoning for closing off the repo, but I've never dealt with users basically defacing my Issues system, so I don't know what your options are as repo owner in a situation like that.