I did one too, a few years ago: http://tirl.org/software/sudoku/ I hadn't seen Norvig's article yet then, but I had just taken a course taught out of Russel & Norvig's book -- and I arrived at a similar solution (depth-first search).
They take the coded sports results, and automatically generate narratives using statistical speech models. They have a startup that is doing it too, don't recall the name of it....
But do we need one (I mean social graph)? I believe our code (ie. github) or advices (ie. SO) speaks more than social graph when it comes to programmers, doesn't it?
One thing the new readability does that I love: when paging down with the space bar, it puts a spot (•) in the margin where the previous bottom of the page was, and then smooth scrolls. Together, these make it much easier to keep your place on the page when scrolling; a bit enhancement to easy readability.
It would be even better if you could bookmark your place in a long document and return to it later...
Noted. Count on finding both of those in a future release.
The first one I was actually doing before Readability -- the old version of Readable did it (and Readability wasn't), but I haven't yet reimplemented the feature in the new version.
The "contribution" percentages are a little funny. They certainly aren't based on lines of code, or number of commits. How is that calculated? Feels weird when you own a project and have done 99% of the work on it, and it says your contribution is 50%.
Can you give me a link to it to the repo in question? It should be the number of commits. Was the initial master branch you used what you uploaded to github? I'm curious as well ....
It appears that it just hadn't finished acquiring the data for the profile. It grabbed a minor, outdated fork rather than the latest version. After some time, it's caught up to the more central fork and more believable percentages.
Once the "incomplete profile" marker disappears the numbers should be correct. If they are still incorrect, could you do me a favor and send me a link to the repo in question? I'd greatly appreciate it!
Thanks for the feedback. We're testing confirmation-less registration now and will likely roll it out later today. It's very important to us that people find it easy to comment (though that's checked by concern for keeping a handle on abuse, given the politics of the site's content).
Thanks very much for the reply. Topix has an analysis at http://blog.topix.com/2008/01/anonymous-comments-by-the-numb... of the effects of allowing "anonymous" (no registration, but still with an alias) comments. The results were the percent of abuse didn't change either way since trolls and the like have lots of free time to grab a public email account, but no-registration greatly increased the amount of contribution, and possibly the quality since professionals with limited time and high quality contributions are perhaps less likely to deal with the rigamarole of registration.
Topix does still have a lot of bad comments, but it's more the nature of a small town political board and so attracts a lot of gossip and low IQ people.
I think a better design is HN/Reddit where one has to register but email (with confirmation if presented) is optional in the registration and all registering does is validate the user with a captcha and group their posts so they can be found and maintained (edited/deleted) by the user. I think that topix would be well served in fact to make the slight adjustment to switch to hn/reddit style.
Of course both reddit and HN are notable for having virtually no noticeable spam and being relatively free from abuse especially compared to boards with registration like yahoo news discussions.
Reddit HN and topix all do have a comment feedback mechanism though so that may provide social feedback regarding appropriate comments. These mechanisms do have a weakness though in that they are often used to punish unpopular points of view even when reasonably presented.
OK, so that's on the topic of the social effects of registration methods. I also want to talk about the value and tech implementation of anonymity on such a site but I'll start that in a new comment so as to segregate these.
I've been thinking about your site over the last day. Although there are many sites with good ideas announced on HN, this is the first site announced here that I think has the potential to change society. Obviously the United States is well known to have an unusual take on imprisonment and punishment that is not shared by any other societies in the modern era. Having a place to put together more of the story and follow up is going to be extremely valuable especially since it will have the one voice that was missing, the side of the con. Whether truthful, delusional or outright fabricated, his words will be read and analyzed, often by parties with personal information about the crime, its effects and aftermath.
I see that families of the con, families of the victims, concerned citizens, the many people fascinated by crime, real lawyers, amateur lawyers, judges, police, all will contribute points of view. No doubt some of these will want to make demands that some points of view be removed and this will lead to conflicts.
But overall this will lead to a fruitful discussion and insights about both crime but the way it is handled in our society.
Regarding anonymity, consider the case of a gang member or other member of an organized crime unit in prison for murder is proclaiming his innocence. But a bystander who saw the shooting posts about what he saw. The defense attorney, working on appeal, may also want to know the person's identity and will subpoena the site, unveiling the identity and making it public in court documents. This information would be useful to the gang or organized crime syndicate which would then want to track this person down and kill or intimidate them. If the person's identity is found out he may be killed or harmed. Because of this, it would be best when subpoenaed then to truthfully say "we do not store ip address or browser fingerprint in our logs". DuckDuckGo does it this way (http://www.gabrielweinberg.com/blog/2010/05/duck-duck-go-sea...), and mailinator.com works similarly by keeping everything in RAM and trashing logs that are more than a few hours old. You may wish to implement something similar, and if not, make clear on the site usage policy to contributors that their identities can theoretically be traced.
