sgk284

47 ms·

1. 451.▲
   by sgk284 17y ago

   I love this idea, but it should really be initially implemented as a kind of A/B test. Simply choose half of the UIDs at random and always show them the comment score, and don't show it to the other half. Track which comments are upvoted mo
2. 452.▲
   The Pareto Principle and Google Code Jam(stevekrenzel.com)

   2 points by sgk284 17y ago|1 comments
3. 453.▲
   by sgk284 17y ago

   Really, you believe that? The two challenges that immediately come to mind are scaling and data mining, but there are a number of others. The use cases presented by the web have pushed distributed computing further than any other single thi
4. 454.▲
   by sgk284 17y ago

   Damn near every LCD is 6-bits per channel, people are getting upset over nothing.
5. 455.▲
   by sgk284 17y ago

   There is quite a difference. Einstein swore that his very own equations dealing with quantum mechanics had to be wrong because it described a phenomenon known as quantum entanglement, which goes against all logic and seems impossible. This
6. 456.▲
   by sgk284 17y ago

   They did have gambling back then. The roulette wheel came about in the 1800's, or they could have shuffled playing cards and drew some out of the deck, or thrown dice. These aren't "natural" as in "from the earth", but they are trivial ways
7. 457.▲
   by sgk284 17y ago

   The best solution I know of is: a, i = 11, 0 while a > 0: a, i = a & a-1, i + 1 If there are k bits set out of n bits, it only does k iterations.
8. 458.▲
   by sgk284 17y ago

   Do you like using Blackboard though?
9. 459.▲
   by sgk284 17y ago

   Thanks for the feedback. Yea, the two things we're really missing on the front page are "Why should you use this?" and "How do you use this?". We're talking about adding a video and some concise copy like you recommended.
10. 460.▲
    by sgk284 17y ago

    Haha nice suggestion :) Thats one of the perks of having assignments viewed in the browser, everyone sees the same thing and no one can claim they didn't know it was corrupt. Thanks for the compliment too.
11. 461.▲
    by sgk284 17y ago

    Thanks for the feedback. We're trying hard to get all perspectives on this issue since we know we're tackling two major problems at once: designing an awesome web app and competing in an entrenched industry. Again, the feedback is really ap
12. 462.▲
    by sgk284 17y ago

    That's okay. We're looking for a challenge. And we have a couple ideas on how to tackle this.
13. 463.▲
    by sgk284 17y ago

    Thanks for the feedback!
14. 464.▲
    by sgk284 17y ago

    With regard to "It's kinda hard to see the advantages of a one off system like this in that context." We see this as a major advantage. Most current systems provide (at best) mediocre solutions to the problems their users face. Instead of
15. 465.▲
    by sgk284 17y ago

    One of the ways we're tackling this is by letting teachers use Classlet without needing their institution to support it. Any teacher can hop on and start using classlet immediately, which is great because I've never met a teacher who likes
16. 466.▲
    by sgk284 17y ago

    Does my reply to icey help? http://news.ycombinator.com/item?id=679018 Thanks!
17. 467.▲
    by sgk284 17y ago

    Good point. Wasn't sure if it was better to see if we designed the site well enough that people could figure it out without us telling them. Classlet makes handling assignments easier for teachers and students. We've got a whole bunch of fe
18. 468.▲
    Ask HN: Review our startup - Classlet.com

    30 points by sgk284 17y ago|49 comments
19. 469.▲
    by sgk284 17y ago

    I don't. I've given it a shot twice, but it simply wasn't appealing. I guess the better question is, why should I? How does it benefit me?
20. 470.▲
    by sgk284 17y ago

    Thanks jganetsk, words of encouragement are always appreciated.
21. 471.▲
    by sgk284 17y ago

    Ah, that would be specific to the job. If your doing a sum, maybe 0, or if you're building a list, it'd be an empty list. I suspect that a good default would just be whatever the first value generated by the mapper is, in which case the fir
22. 472.▲
    by sgk284 17y ago

    Completely agree. I should have elaborated more on that, thanks for bringing it up. In many cases you can have a significantly lower memory profile. In a worst case it's the same as MapReduce (as you said). I think having the additional fle
23. 473.▲
    by sgk284 17y ago

    Think of the hashtable as a defaultdict or the fold as having an initial value for when the key doesn't have a value yet.
24. 474.▲
    by sgk284 17y ago

    Here ( http://files.getdropbox.com/u/196807/peaktraffic.tgz ) is a proof of concept of HashFold that I used for my solution for the facebook puzzle ( http://www.facebook.com/careers/puzzles.php?puzzle_id=8 ). It's written in python. You
25. 475.▲
    by sgk284 17y ago

    Does my response to jganetsk help at all?
26. 476.▲
    by sgk284 17y ago

    Both. First it does a local fold on local inputs, then it does a global fold across all of the local fold results. i.e. Assume we have three nodes and one key, "our_key", and values for "our_key" are distributed across the 3 nodes: node1 :
27. 477.▲
    by sgk284 17y ago

    Hey guys, I've been playing with this concept for a little now and thought this might be a good forum for discussing the idea. I've got a small prototype that I used to solve a few problems (most notably: http://www.facebook.com/careers/pu
28. 478.▲
    Improving MapReduce with HashFold(stevekrenzel.com)

    41 points by sgk284 17y ago|18 comments
29. 479.▲
    by sgk284 17y ago

    Using heuristics, it attempts to form structure from unstructured data. It's a pretty good first attempt.
30. 480.▲
    by sgk284 17y ago

    Not only is it voice recognition, but the microphone array determines the direction of the sound and correlates it to a body... so it knows exactly who is talking and where they are talking from. Pretty bad ass.