neilc

45 ms·

421.
by neilc 17y ago
Did anyone read through to the article NYT review? ( http://www.nybooks.com/articles/23231 ) The yottabyte quote is: "As the sensors associated with the various surveillance missions improve, the data volumes are increasing with a projectio
422.
by neilc 17y ago
Yeah -- the quoted comment reflects a remarkable ignorance of the intellectual content of Rand's novels.
423.
Does it make sense to replace all disks with SSDs?(perspectives.mvdirona.com)
11 points by neilc 17y ago|3 comments
424.
by neilc 17y ago
Most surprisingly good speaker: Chris Anderson Really? I'd be curious to hear what you got out of his talk. I thought most of the discussion of Freemium business models was pretty rudimentary.
425.
by neilc 17y ago
I would also like a pony.
426.
by neilc 17y ago
Dribbling out an interview over a series of days is annoying (especially if you don't bother to edit the transcript for basic grammatical correctness). Better to post the entire interview series and then post it as a single article.
427.
by neilc 17y ago
By exporting this data first into HDFS, I can now run Map/Reduce jobs more quickly and not hang up DB connections Right; the typical configuration is to have one database that does transaction processing, and another separate database (a
428.
by neilc 17y ago
Yes, MapReduce is very scalable -- that is well understood. There is more to building a system than pure scalability, though: for example, energy efficiency. If Hadoop was less inefficient, it would be significantly more green. As energy gr
429.
by neilc 17y ago
There's no reason you can't use a relational database for "huge batch-processing jobs." You might not be able to use MySQL for such a task, but there are a large collection of DBs that are designed for large-data analytics: Greenplum, Aster
430.
by neilc 17y ago
this is completely back of the envelope vaporware Sure -- the linked-to paper is a position paper that basically just explains the problem they are trying to solve, not the specific details of their solution (the project is just beginning
431.
by neilc 17y ago
http://www.feyrer.de/NetBSD/gmcgarry/ suggests that the cost to acquire an uncontested mutex on NetBSD in 2005 was ~37 nsec, so ~25 nsec doesn't seem too unreasonable (although strangely it was more expensive on FreeBSD).
432.
by neilc 17y ago
even with a statically typed database, type matching errors storing data are only reported at runtime! (That is, our java compiler doesn’t check our MySQL schema.) That isn't necessary the case: for example, PG'Ocaml allows Ocaml program
433.
Numbers Everyone Should Know(everythingisdata.wordpress.com)
22 points by neilc 17y ago|10 comments
434.
John Ousterhout: Startup Company Culture(stanford.edu)
5 points by neilc 17y ago|0 comments
435.
by neilc 17y ago
leveraging shared memory and other operating system tools, you can turn a fork(2) implementation into a thread alternative -- but then, you'd have ... threads There's an important difference: in the threading model, all memory is shared b
436.
by neilc 17y ago
Cool stuff. Bryan O'Sullivan is a remarkably productive hacker.
437.
by neilc 17y ago
You will definitely get a mail if you get accepted. If you get no mail, you probably didn't get accepted. I think that is precisely wrong. The FAQ says that all applicants should expect a response, whether they were accepted or not: Q: I
438.
by neilc 17y ago
You should just read the (very readable) MapReduce paper, rather than content-free blog posts like this one.
439.
by neilc 17y ago
Given that Hennessy is the President of Stanford and that he founded a successful company (MIPS Inc, when went public and was then sold to SGI), Hennessy is surely pretty wealthy.
440.
by neilc 17y ago
PostgreSQL has IMHO a nicer solution: abstract data types to represent IPv4, IPv6, and MAC addresses, along with the functions over those types that you'd expect. http://www.postgresql.org/docs/8.4/static/datatype-net-types... http://www
441.
by neilc 17y ago
Compatibility with IPv6 does not mean you need to store addresses as strings -- most IP address ADTs support both IPv4 and IPv6 in some way. But in reality, IPv6 is still not worth worrying about for the vast majority of software (YAGNI).
442.
by neilc 17y ago
Yeah, the TOR website is blocked by GFW, I believe. Basic idea: 1. Email bridges@torproject.org ; the body of the email should contain the line "get bridges". 2. In Vidalia Network Settings, click "My ISP blocks connections to the Tor netwo
443.
by neilc 17y ago
TOR still works fine -- you just need to configure it to use bridges. https://www.torproject.org/bridges
444.
by neilc 17y ago
Folks interested in using logic programming for building distributed systems might be interested in some work my research group is doing: we've built Hadoop, HDFS, and Paxos in a high-level distributed variant of Datalog, which is another l
445.
by neilc 17y ago
Title is obvious linkbait.
446.
by neilc 17y ago
I wouldn't say it is "obviously" a problem: READ COMMITTED are perfectly reasonable semantics, and are preferable to many applications (e.g. many apps aren't written to retry transactions that abort due to serialization conflicts). Which is
447.
by neilc 17y ago
get your configuration variables correct, or it wont work out-of-the-box (kernel.shmmax, autovacuum, shared_buffers) The only thing you really need to worry about is making sure the kernel SysV limits are appropriate (kernel.shmmax); auto
448.
The Blurring of Layer 2 and Layer 3(idleprocess.wordpress.com)
14 points by neilc 17y ago|1 comments
449.
by neilc 17y ago
More information on the GCD FreeBSD port is here: http://lists.macosforge.org/pipermail/libdispatch-dev/2009-S...
450.
by neilc 17y ago
Waterloo has a very strong CS program, albeit it can be a lot of work, and the campus life isn't great. If you want an intense undergrad CS experience, I think U of T and Waterloo are the best two choices. That said, I think you should also

More ›