mumrah

Before 0.8, you are correct that if you lose a disk, you lose the data. However, just because a broker goes down, that doesn't mean you lose the data on it - it just becomes unavailable to consumers. The log files backing Kafka partitions are fully durable append only files. So the durability guarantees are pretty good, even before 0.8.

What you're talking about is failover and fault tolerance, which are greatly improved in 0.8 with the addition of replication.

It is a mixture of Java and Scala, though mostly Scala. GitHub mirror of the source: https://github.com/apache/kafka

Quora discussion: http://www.quora.com/RabbitMQ/RabbitMQ-vs-Kafka-which-one-fo...

RabbitMQ developer on the kafka-users list: http://mail-archives.apache.org/mod_mbox/kafka-users/201306....

SO discussion on several queuing systems: http://stackoverflow.com/questions/731233/activemq-or-rabbit...

This is the first release as an Apache top-level project and represents many months of hard work.

A few of the major improvements (from https://archive.apache.org/dist/kafka/0.8.0/RELEASE_NOTES.ht...):

  * Intra-cluster replication support
  * Support multiple data directories
  * Many new internal metrics
  * Time based log segment rollout

Plus many bug fixes and other improvements.

I also got an unsolicited email from them asking for feedback on the product. I didn't really mind since it is relevant to my area, but it was still a bit odd.

Since 4.0, Lucene's memory usage has dropped quite a bit actually.

This article makes no mention of spouse or kids, both of which can be major distractions. For adults, it's very hard to not see someone as "here" even though they are just down the hall; for kids - it's impossible. The compromise my wife and I have is that I take time during the day to help out around the house or run errands, but I make up the time at night.

Also, I pretty much have my headphones on all the time, which is not so different from when I worked in a real office.

Even the low end Sennheisers (like the HD202 on my head) offer a much better experience than earbuds. At 27 USD on Amazon, there's not much excuse for people who scoff at expensive audiophile headphones.

I see in no way how this is an improvement.

Where is the option "I would like to, but I have a toddler"

I am quite sure it will not come close to Lucene/Solr in terms of performance or capabilities.

+1 for use of ampersand ligature in "&c", i.e., "et cetera"

You are correct. It is not an implementation of Paxos, but a Paxos-inspired protocol. http://wiki.apache.org/hadoop/ZooKeeper/PaxosRun

This seems to be comparing ElasticSearch to Solr 3.x which is disingenuous. Solr 4.0 has been released for months now and includes many new features around distributed search and indexing.

Just to point out a few corrections: * Solr does have a transaction log. When using soft commits, the tx log is used to recover any volatile writes in the event of a crash * Solr replicates in a similar manner to ES. It no longer does full-index replication as indicated by the article. A write comes in and is routed to the shard leader and replicas. * Solr too has "High Availability built-in" with automatic failover of shards

One fundamental difference the article does not address is the robustness of Solr's cluster management. Solr uses ZooKeeper under the hood which implements the Paxos algorithm to avoid issues like "split-brain syndrome". ElasticSearch has implemented its own distributed coordination and is susceptible to such issues.

I never understood why this is easier than just manually doing the FOIL. 1312 = (10 + 3)(10 + 2) = 1010 + 102 + 310 + 32. Drawing all those lines and counting intersections seems tedious.

Pull request might not be a terrible attack vector. I'm sure there are plenty of people out there who don't look carefully and just merge them in.

In my experience, heating and cooling is the bulk of residential energy use. If you live in an older home, make sure to get an energy audit from the utility company (usually free - maybe you just need some additional insulation.

Also, thermal curtains (aka, black-out curtains) can help if you have a lot of window surface area.

Of course Mathematica is more succinct - it is a domain specific language, not general purpose like most of the others compared. The only meaningful comparisons to Mathematic would be R and MATLAB (other DSLs).

I must have missed something a while back, but Arduino being sold at Radio Shack is news to me! I think this is great for the maker movement and people trying to get started with Ardiuno. RS is selling a book and various shields - in store no less! I recall reading that RS was going back to its DIY roots [1] - I guess they were serious about it

1. http://blog.radioshack.com/2011/05/19/radioshack-the-diy-com...

Stickers are a neat idea, but why not just some special drawn symbol? Would be way cool if you could define your own.

"Mathematica is one of the more complex software systems ever constructed"

Wolfram, et al really thinks they are hot shit.

Average rating as a measure of "goodness" is wrought with statistical problems. Without the context of other statistical modes, looking at mean is pretty useless. However, people don't want to look at summary statistics for each item (mean, median, mode, std/var, skew, etc). So we try to come up with scalar metrics that capture "goodness" or "coolness" or whatever. Popularity (how ever you define it) is a common one to use. Here's a good comparison of popularity models: http://blog.linkibol.com/2010/05/07/how-to-build-a-popularit.... In the past I've been pretty happy with "Bayesian average" - it's simple to implement and gives good results.

But if you really want to dig into it, you have to consider all kinds of stuff like bimodal distribution of ratings (controversial items), rater quality/consistency, age or ratings, etc, etc.

It's really not as simple as you'd think!

Since we're all reminiscing about high-school geek antics: In my Physics class, we had semester long ongoing assignments that were markedly harder than the nightly homework. You could turn in the problems at your own leisure. They were mostly applied problems that involved some math we hadn't learned yet (like calculus). The problems were the same for each student, but we all had different parameters, so our answers would be different.

After figuring them all out a few weeks into the semester, I started writing up some BASIC programs on my TI-86 that would take in student's parameters and spit out solutions. Long story short, I ended up selling answers to some jocks and got caught (I guess the teacher was suspicious when C students were getting these hard problems correct). End result was: made my teacher simultaneously proud/disappointed, earned a few bucks, learned about corruption/greed.

All in all, it was a good learning experience and I don't regret it (though giving away answers for free would have been more altruistic I guess)

As with most agriculture in the US (assuming you're talking about the US), family farms are the exception - not the rule.

It's just an annoying convention around here. My best guess is because the street address can be one of 5 or 6 cities (Raleigh, Durham, Chapel Hill, Cary, Wake Forest, Morrisville, Apex, etc) and so people talk in terms of the metropolitan area which is Raleigh-Durham-Chapel Hill or the "triangle".

That's correct. I believe they are using it in some new "high availability" stuff coming down the road

I was hoping for eigen faces (http://en.wikipedia.org/wiki/Eigenface), but this was cool nonetheless!

I wonder what motivation they had for developing their own BigTable/Dynamo implementation rather than going with Cassandra or HBase.

Now there are three Apache projects with quite a bit of overlap (distributed columnar/key-value database). Hooray for fragmentation!

Aren't they missing a cartel in NM - specifically a large production facility in ABQ?

This reminds me of some early MongoDB benchmarks. Claims of extremely high write throughput when really it's writing to memory and relying on mmap to eventually sync the data to disk.

Making write speed claims when you're not touching the disk is like bragging about how fast you can cat a file to /dev/null.