ewencp

7 ms·

1. 31.▲
   by ewencp 14y ago

   What you're quoting refers to the scaling of the axis after all the values are computed. This affects the numbers on the Y-axis, but not the shape of the graph. The issue wting was raising is the definition of search volume. If search volum
2. 32.▲
   Twitter may leak people you visit but don't follow in 'Who to follow'(bmistree.com)

   2 points by ewencp 14y ago|0 comments
3. 33.▲
   by ewencp 14y ago

   At least on the Twitter web site, they already have the unwrapped URLs -- see the data-expanded-url and data-ultimate-url attributes on links (and, of course, at least one unwrapping is used for displaying the link). They probably don't wan
4. 34.▲
   by ewencp 14y ago

   Right, it isn't that surprising. I was originally motivated by the general question of how many URL shorteners you have to hop through to get to sites from Twitter. What this doesn't show is as interesting as what it does: other URL short
5. 35.▲
   by ewencp 14y ago

   I think there are other reasons as well (e.g. analytics, as mentioned in the post, when links are consumed through third party clients), but I don't buy this one. Wouldn't this only be useful in breaking spam links when they are spread beyo
6. 36.▲
   by ewencp 15y ago

   Not exactly what you're suggesting, but https://github.com/jterrace/js.js is in a similar vein: SpiderMonkey -> LLVM -> emscripten -> JS to run a JS within JS.