Millions of Lines of Code(informationisbeautiful.net)
informationisbeautiful.net
Millions of Lines of Code
http://www.informationisbeautiful.net/visualizations/million-lines-of-code
8 comments
Interesting graph if it were accurate. The healthcare.gov stuff seems like jokes, right? That makes it seem less accurate. What else in the graph is a joke? E.g. can car software really have more lines of code than a boeing 787, operating systems, and the LHC?
It's not a joke, it's Obamacode. Allegedly there are 500 million lines of code running healthcare.gov. That figure is, of course, wildly inaccurate, not unlike most everything else concerning Obamacare. But their PR people passed this number to journalists in an effort to blame the sheer complexity of the project rather than run-of-the-mill incompetence.
I'm not from the US, so don't really know why there is so much focus on the website rather than the healthcare itself in the US news. But now I'm wondering, what language is healthcare.gov written in and did they put every keyword on a new line?
The focus on the failure of the website is actually mainly being pushed by Democrats and liberal US media organizations. The reality is that enrollment figures have been staggeringly low, and the program at this time appears to be headed to fall far short of the 7 million enrollees it needs to be financially viable. They can choose to either have stories about how Americans appear to be roundly rejecting the program itself, or they can talk about website failures. This being Obama's signature piece of legislation, Democrats and most of the left-leaning media have chosen to focus on the website issues while praying that Americans somehow come around and start signing up. There was actually a recent Associated Press article saying that the low enrollment numbers "only show the vast number of Americans eagerly waiting to sign up". Other, similar nonsense has been published extensively by left-leaning media hoping to spur public interest in signing up.
As for the language/line count, I don't know what language it is but this 500 million number had to have been picked out of thin air. Perhaps they counted the number of lines of actual code and multiplied it by the number of servers it is running on.
As for the language/line count, I don't know what language it is but this 500 million number had to have been picked out of thin air. Perhaps they counted the number of lines of actual code and multiplied it by the number of servers it is running on.
I think that number includes the LOC of every system they interact with, and every library they use... I recall reading there were over 850 "major" libraries used and it interacted with hundreds of other systems which had to be customized (state and local).
Just adding in boost and using like 1 function from it adds 25M lines of "code" (dumbly counted including comments and such).
Just adding in boost and using like 1 function from it adds 25M lines of "code" (dumbly counted including comments and such).
The line count (wildly inaccurate or not) for "healthcare.gov" includes all of the actual insurance exchanges and related infrastructure as well, not just the actual website healthcare.gov. They should have labelled it differently.
The article they got the figure from also states the typical bank's software is 100 million lines of code. I'm going to guess someone was off by a magnitude or two. And to claim it would take 5 million lines to fix it...
The figure originates from this Wired article, where it is uncited: http://www.wired.com/autopia/2012/12/automotive-os-war/
I think we can take it with a large helping of skepticism.
I think we can take it with a large helping of skepticism.
Knowledge is beautiful, but for it to be knowledge it needs more details and sources..
The chart would be more helpful if it compared lines of assembly and not lines of original code (which is what I suspect it's comparing).
The most surprising item in the list to me is Visual Studio 2012 with 50M LoC (healthcare.gov might considered ridiculous, but after all we observed, it is not surprising at all) .
This is by far, more than I can imagine needed for the most sophisticated IDE one will ever need.
It might includes all compilers for all supported languages though.
Also note they provide many of their infographic works as printed posters at http://store.informationisbeautiful.net/
This is by far, more than I can imagine needed for the most sophisticated IDE one will ever need.
It might includes all compilers for all supported languages though.
Also note they provide many of their infographic works as printed posters at http://store.informationisbeautiful.net/
About everything we use consume far more lines of source code it could (at least if we rewrite it with current knowledge). See the VPRI and their almost full featured OS (with compilers and applications) in less than 20.000 lines.
(Manifesto) http://www.vpri.org/html/work/ifnct.htm
(Latest progress report) http://www.vpri.org/pdf/tr2011004_steps11.pdf
(Manifesto) http://www.vpri.org/html/work/ifnct.htm
(Latest progress report) http://www.vpri.org/pdf/tr2011004_steps11.pdf
Something that might have been interesting (information content of life forms vs. programs) has been wasted by equating base pairs with lines of code. This is totally arbitrary.
Unfortunately I don't know of a specific better measure, but it might consider the used and potential information content in the structures encoded. Anybody know of such a metric for source code complexity?
Unfortunately I don't know of a specific better measure, but it might consider the used and potential information content in the structures encoded. Anybody know of such a metric for source code complexity?
Ah, lines of code... the most useless metric in software development known to man.
wow look at the symbian its double the android
What do the big semicircles with %s mean?
Ok figured it out. Wow that is unclear, its totally illogical too, the size of the circle seems proportional to the increase in LOC in the next random iteration he has chosen, but the diameter is also the distance between them vertically which is based on the order they are in. I reckon there was a lot of hacking and including / not including stuff in order to get those circles to match up and also present stuff in ever increasing order.
Also size of circles is arbitrary 100% size cirle is just whatever
+organism?? really? That is not helpful
% changes between related items (later versions of the same OS or software package etc)
Change from one version to the next.