3 ms·
He acknowledges that MIT doesn't prevent this in the conclusion: > As my code is MIT-licensed, there’s nothing legally preventing the use of my code this way.
by ohitsdom 11y ago
He acknowledges that MIT doesn't prevent this in the conclusion:
> As my code is MIT-licensed, there’s nothing legally preventing the use of my code this way. However, I’m not going to change my workflow; I will still open-source my data visualization code, and it will have the same licensing. It wouldn’t be fair to punish others who have made very good use of my code.
> But outright using the code, without any attribution, and claiming it as your own when it blatantly isn’t, is a jerk move.
- DannyBee 11y ago"> As my code is MIT-licensed, there’s nothing legally preventing the use of my code this way. " He's legally wrong, however ;-) Let me clarify: The MIT license does not prevent someone using the code to do a visualization without crediting him. It does, however, prevent taking those snippets without reproducing the copyright notice For anyone who wants to argue about the copyrightability of it, i'll point out A. Google is being sued for 2.6 billion dollars over 9 lines of range check code B. While copyright has some notion of de-minimis, it's not a bright line in the code space. Copyright is more about originality. You can't really say, less than ten lines, do what you want. 10 lines of java does nothing. 10 lines of perl could be an entire OS.
- PebblesHD 11y agoThat perl comment is possibly my favourite thing on this site so far.
- brazzledazzle 11y ago>A. Google is being sued for 2.6 billion dollars over 9 lines of range check code If I were a Google exec I'd strongly consider dumping a few million in sponsoring some developers to make postgres's clustering options feature competitive with RAC. You'd never unseat it in some places but a long term commitment to support development from a heavy hitter like Google might go a long way.