Python Library for Google Sets | Hacker News Reader

Python Library for Google Sets

catonmat.net

23 pointspkrumins15 comments

Just a heads up for guys that are new to this: seeing that this is scraping - you can get your IP blocked from Google.

That's true. But you can always use anonymous proxies.

Sorry for nitpicking, and although I'm not sure about Latvia, (where Peter lives), in the US "no license whatsoever" certainly != "you are free to do absolutely anything you wish with the code".

Thanks for the comment.

What I have in mind is that there is no license, none, it's public, free code. Doesn't belong to anyone. Anyonce can take it and use it any way they wish.

What do you suggest I write instead to make this point correctly?

I am not a lawyer, nor have I ever published code under any license, but I believe that, to let anyone do whatever they like with your code, you say that it is in the public domain.

"Dedicated to the public domain" creativecommons.org/licenses/publicdomain/

This is probably the best option, but the FSF-approved WTFPL (Do What The Fuck You Want To Public License) may be clearer.

[Previously discussed on HN here: http://news.ycombinator.com/item?id=652183]

Honestly, pick a proper open source license like apache to protect yourself (no-warranty clause) and your users (explicit patent and copyright grant).

Also, as you are latvian, under Wipo, you are essentially unable to give something away without a license due to the moral rights clause. I mean you can, but no one has a right to use it unless you explicitly grant it. You cannot disclaim rights and place it into the true public domain.

Also, scraping is bad, m'kay.

scraping is bad, m'kay

To the contrary, I think scraping is part of the original smalltalk OOP vision, where objects are entities universally accessible, without limitations on who/what/how they are being seen/edited.

I think he outranks you when talking about scraping google contents: http://en.wikipedia.org/wiki/Chris_DiBona

Thanks for the suggestion. I'll choose a better license! I agree that scraping is bad, but there was no other easy way to do what I wanted (the english improvement program that I talk about in the article), so I had to write this code.

This is cool but is there a function for finding hidden topics? i.e. given red, yellow, blue -> colors

It doesn't have, and I can't think of an easy way to group them this way. Perhaps I could cook something up, like given "red, yellow, blue", do a Google Search (with my library) for each "red", "yellow", "blue" and find the surrounding words. Then count the frequencies of surrounding words, do another search at Google Sets for "red, yellow, blue", remove all those items from frequency list, and then output the most frequent word. It might be "color" or "colors". Perhaps also de-pluralize the words so that "colors == color". Just a quick idea.

mmm... might actually try it out, if it works, will write an article about it.

That may work but there's going to be so much noise in the population how do you filter it to get "colors?"

There's a google vid on this subject: I never did get through it but it might make more sense to you: http://www.youtube.com/watch?v=vgqWMGT9haY

Thanks, will watch that one!

Reply on news.ycombinator.com