jbellis

68 ms·

781.
by jbellis 16y ago
Pedantic note: MSFT started paying dividends in 2003, 1/3 of the way through the Ballmer period.
782.
by jbellis 16y ago
In-memory with periodic snapshots sure implies "not durable," though. (That is, you will lose whatever data has not yet been snapshotted on a power failure or crash.) Replication doesn't fix that, unless you're willing to impose the laten
783.
by jbellis 16y ago
Most use cases also don't have the entire data set equally active, and it's silly to pay to keep everything in RAM instead of just the hot parts.
784.
by jbellis 16y ago
... ruba will be closing shortly. /cynical
785.
by jbellis 16y ago
It's a little sad watching people who live in the Perl echo chamber convince themselves that Perl use hasn't declined in the last years.
786.
by jbellis 16y ago
This (and speed) is why Cassandra uses murmur for its bloom filters: http://spyced.blogspot.com/2009/01/all-you-ever-wanted-to-kn...
787.
by jbellis 16y ago
I would think it would be possible to find spending-per-pupil for catholic schools to compare apples to apples. Quick google says catholic schools are financially "twice as efficient spending less money per pupil" as public schools: http:/
788.
by jbellis 16y ago
It's based on the work discussed in http://cs-www.cs.yale.edu/homes/dna/papers/hstore-cc.pdf
789.
by jbellis 16y ago
I bet if I googled hard enough I could find one of these articles for every state in the union.
790.
by jbellis 16y ago
Thanks for the correction.
791.
by jbellis 16y ago
Except they can't, really. Dalvik is completely unrelated to Hotspot or any other JVM out there. (Remember that Android doesn't run Java class files natively, it's first compiled into dex.)
792.
by jbellis 16y ago
I understand why Google would want to use Java in their app platform. But if they didn't have the resources to get JIT working before now, wouldn't gcj (or hell, write your own java-to-machine-code compiler, clearly Google isn't scared of
793.
by jbellis 16y ago
Well, roughly speaking, you can group Cassandra deployments in two categories+: new products that are hoping they need Cassandra's scaling ability someday, and existing products moving to Cassandra from something else because the pain of sc
794.
by jbellis 16y ago
320k per machine per second sounds really high, even granting high-end machines, since the non-10gen ops/s numbers I've seen have been in the 2k to 10k per second range (e.g., http://prajwal-tuladhar.net.np/2009/11/15/500/mongodbs-perfo...
795.
by jbellis 16y ago
It's like how in postgresql, if a client runs "select * from lots", and you kill the client, the query keeps going even though there's nobody to hand the answer to. That said, there are ways we can mitigate this, primarily in https://issu
796.
by jbellis 16y ago
Incidently, on EC2 the extra cpu in an XL or HMXL makes all the difference between "oh shit, we ran out of capacity and bootstrapping more slows things to a crawl" and "we ran out of capacity but we can limp along during the bootstrap." YY
797.
by jbellis 16y ago
It wasn't that the resiliency safeguards didn't work, so much as reddit was simply underprovisioned. I have no idea what TFA meant by "at only three nodes we weren't able to take advantage of most of Cassandra's safeguards for ending up in
798.
by jbellis 16y ago
It doesn't say anything about performance. I checked find-in-page in case I somehow missed it. :)
799.
by jbellis 16y ago
Bet you didn't know there was a [relatively] new COBOL-2002, either. (If I were snarky, I would note that the next COBOL spec after that -- yes, there is one -- might well be released before perl 6.)
800.
by jbellis 16y ago
Similar project for Python (that I believe predates Toto, could be wrong): http://www.blogofile.com/
801.
by jbellis 16y ago
nbio isn't a magic wand; context switching is virtually free in a language without a GIL http://paultyma.blogspot.com/2008/03/writing-java-multithrea...
802.
by jbellis 16y ago
Turns out it's difficult to parallelize without also denormalizing...
803.
by jbellis 16y ago
To clarify, Reddit replaced memcachedb w/ cassandra entirely, but a lot of the site is still on postgresql. My understanding is they plan to move more to Cassnadra gradually.
804.
by jbellis 16y ago
Marc Andreesson wrote (about hiring suits, not engineers): "if you know what you're doing, the odds of a given executive hire working out will be about 50/50. That is, about 50% of the time you'll screw up and ultimately have to replace the
805.
by jbellis 16y ago
Allowing your users to determine how short is short enough would be a good thing.
806.
by jbellis 16y ago
Well, it's a mixed bag. Top causes of death in the US, 2009: Heart disease: 631,636. From what I have read, heart attacks are painful but usually over with quickly. Living with a pacemaker or heart surgery can suck but I'd definitely take
807.
by jbellis 16y ago
I remember looking at their engine 3? years ago. They'd started w/ an OSS engine, then pulled development in-house and left the OSS tree mostly unattended, then announced it would be fully OSS again "real soon." Glad I didn't wait.
808.
by jbellis 17y ago
I dunno, as someone who does a lot of free support, the quote in context sounds blunt but not asshole-ish to me. (Especially if Canonical has a track record of trying to get free work out of him or other kernel developers.) "I've already
809.
by jbellis 17y ago
It's a little snarky, but Ted isn't the first to have observed that Canonical has very high developer turnover rates.
810.
by jbellis 17y ago
Less hand-waving, more economics: http://www.themoneyillusion.com/?p=4954

More ›