paul

54 ms·

541.
by paul 17y ago
The Tornado part of the test is mostly useless (and misleading) since he's running it with select instead of epoll.
542.
by paul 17y ago
5% isn't as modest as you think once you take into account inflation and taxes, and the fact that the stock market isn't going to magically go up 10% every year (as many people erroneously assume). Living off of 2% of your assets is a safer
543.
by paul 17y ago
You should be able to get near-instant updates by using Twitter's streaming API: http://apiwiki.twitter.com/Streaming-API-Documentation#statu... I just updated Tornado's httpclient to include a streaming_callback parameter which allows yo
544.
by paul 17y ago
Exactly. Stuffing everything into one process is a whole lot of complication for very little benefit.
545.
by paul 17y ago
Ideally one could "realize that large chunks of the human experience can't be conveyed through language", etc without getting scammed and joining a cult, which is why it's good to you both parts of the brain :)
546.
by paul 17y ago
Not having to deal with Apache is its own reward :) Seriously, having a self-contained server is often nice, and the async support would be useful if you need to do long-polling or make http requests to other sites (e.g. making an API call
547.
by paul 17y ago
Only in the sense that Python is written in C. Tornado has a C module to expose the epoll system call on older versions of Python, but it's not needed at all with Python 2.6.
548.
Paul Buchheit: Left brain, Right brain, and the other half of the story(paulbuchheit.blogspot.com)
54 points by paul 17y ago|12 comments
549.
by paul 17y ago
It would also be fun to see how the raw Tornado httpserver ( http://github.com/facebook/tornado/blob/master/tornado/https... ) performs without the Tornado web.py framework on top of it.
550.
by paul 17y ago
Your point is fair, but I think you may be dramatically overestimating the amount of time that was spent writing and debugging Tornado. Overall, I'd estimate that less than 1% of our time was spent on Tornado and other similar "reinventions
551.
by paul 17y ago
Yes, opportunity cost is super-important and is probably under-appreciated. In the startup example, you should definitely be considering the worst case (and best case!) opportunity cost since working on one startup (or other job) largely pr
552.
by paul 17y ago
I think it's a c++ library (based on scanning the docs -- I could be wrong), so it's not too surprising to see it outperforming pure Python solutions.
553.
by paul 17y ago
Yeah, people don't actually make decisions logically for the most part, but to the extent that there is any about of reasoning (or rationalizing), there is some technique to it. I updated the post :)
554.
Paul Buchheit: Evaluating risk and opportunity (as a human)(paulbuchheit.blogspot.com)
94 points by paul 17y ago|27 comments
555.
by paul 17y ago
Tornado relies on a small number of well defined system calls (such as epoll). For me, the difference between reading (and debugging) a few hundred lines of code down to the system call level, and reading (and debugging) a big library like
556.
by paul 17y ago
No, not that I'm aware of. It simply means that Python 2.6 makes the epoll system call available directly without need for a custom module.
557.
by paul 17y ago
The great thing about open source is that people are free to share ideas and code. If the Twisted team likes parts of Tornado, they are free to replicate those parts back into Twisted while ignoring the parts that they don't like. Meanwhile
558.
by paul 17y ago
It will also use Python 2.6's native epoll (see line 326 of http://github.com/facebook/tornado/blob/9a8bd2fb6fd6279be16d... )
559.
by paul 17y ago
Fortunately, I don't care. The broad concepts of REST are good, but this pedantic stuff is so silly. There's a reason that most popular APIs will fail this test -- the people who built them were thinking about more important issues than ach
560.
by paul 17y ago
It started as Lucene, but is quite different by now as Jim ( http://friendfeed.com/jim ) had to create all of the real-time indexing and standing query support (at any moment, it is simultaneously executing thousands of queries).
561.
by paul 17y ago
I'm not sure what you mean by "tight polling", but we basically keep a connection open to the server at all times (often known as "long polling"). This is similar to how IM works. You can pause it by pressing the pause button on the top-rig
562.
FriendFeed Makes Its Search Results Real-Time Too(techcrunch.com)
2 points by paul 17y ago|0 comments
563.
VC Vinod Khosla: Vision is 'bumbling' around(vator.tv)
4 points by paul 17y ago|0 comments
564.
by paul 17y ago
Yes, that's roughly my thinking. Everyone has some part of the puzzle, but nobody has been able to put it together. My goal is to find a way to combine all the parts of the puzzle.
565.
by paul 17y ago
It's also interesting that faulty research from almost 100 years ago is still affecting the way we eat. Humans are not rabbits -- they have very different dietary requirements. Apparently nobody has managed to show a link between dietary ch
566.
by paul 17y ago
That's why I'm looking for commentary from knowledgeable people, not just simple vote counts (which reflect enthusiasm, but not effectiveness).
567.
by paul 17y ago
Here's the TechCrunch story: http://www.techcrunch.com/2009/06/24/friendfeed-adds-file-sh...
568.
by paul 17y ago
My point is not that they are the same, but that details matter a lot, and that just because there are parallels between two things, it does not make them the same thing. My .plan example was not made up -- that was an actual criticism from
569.
by paul 17y ago
And blogging is basically just a .plan file...
570.
by paul 17y ago
Let me know if you have a YC startup and I missed your blog - http://friendfeed.com/yc-startups

More ›