80 karma · joined November 19, 2013
Discreet industries have quite specific legal terminology (in my experience) and machine learning is probably better suited to weeding out mistakes in documents rather than interpreting the contents within them. It's the interpretation that earns the big money. The example you've given is useful - but useful to an audience with specific prior knowledge and an existing familiarity with the licence landscape. Expanding beyond this is probably going to be an exercise in unintended consequences.
The idea has a lot of merit and would be truly disruptive it it could be achieved, but if lawyer jokes have taught us anything, there is no shortage of motivation to reduce the worlds reliance on lawyers, but nobody has yet come up with a good solution (to my knowledge).
Also, I can see how someone who's paying this guy a day rate might not take too kindly to an article written by said day-rater that looks rather like "don't worry about how long you'll be paying for this project". He's asking them to pay that day rate for an undetermined length of time for undetermined outcome.
Edited for grammer
1] http://web.archive.org/web/20121028212422/http://code.google...
2] http://www.google.com/trends/explore#q=usb%20stick%2C%20clou...
Of couse, add "dropbox" to that trends search and USB sticks look like their losing serious relevance from mid 2010 - until you factor in the fact that you're looking at a self selecting sample: People on the internet looking for internet connected things. Which USB sticks route around for a lot of users.
yes - anything that signposts card data clearly to an onlooker is problematic, but any malicious shoulder surfer will probably be able to spot a payment page a mile off in any case. This might however, have a small possibility of encouraging opportunistic theft.
I think the bigger threat to this probably comes from the networks (MC, VISA, AMEX etc) who get VERY possessive about mock-ups / card image facsimiles that use their logos. The brand logo protection that they enforce is pretty strict and I would imagine this approach would not be welcomed. I'd like to be wrong, but my experience with the networks makes me pessimistic.