mumrah

Apache Kafka 4.0 was just released. This is a major release which includes some significant improvements:

* Removal of Apache ZooKeeper as a dependency

* New consumer group protocol (KIP-848)

* Early access of Queues For Kafka (KIP-932)

Not for long. An early access version of KIP-932 Queues for Kafka will be released in 4.0 in a few weeks.

https://cwiki.apache.org/confluence/display/KAFKA/KIP-932%3A...

Part of the new design is that brokers and controllers can be run on separate JVMs.

Under 100 for radio and TNC, sure. See the NinoTNC http://tarpn.net/t/nino-tnc/n9600a/n9600a_info.html. For radios we use second-hand commercial radios that we reprogram for Ham use.

However, coax and antenna can easily exceed 100 alone.

I'm part of a TARPN and the main thing we do is chat. There is a simple chat program that runs on the network so we can do realtime multi-user chat. It's pretty fun.

BBS is also a pretty common application.

I think it's something like 30W over a single ethernet cable. I'm running a few devices with PoE that are not network enabled, just conveniently located near an ethernet cable that's not near a power outlet.

Packet radio has made some progress since the 80s :). Here in North Carolina we have a network called TARPN http://tarpn.net/t/packet_radio_networking.html. We use re-purposed commercial radios, RaspberryPis, and custom TNCs for 1200 baud packet links. Some folks are experimenting with higher speed links (9600).

These ISM modules will work on low power for close range links, but you can't just slap an amplifier and call it a day. There are regulatory issues with the tech they use (spread spectrum, wide bandwidth, encryption, etc).

Ordered a RPi Zero W from Adafruit on a Saturday, got it on a Wednesday. Took about two hours to get through a headless install of Raspbian [1], most of which was fighting with my Mac's Disk Utility.

Not knocking the BASIC Engine, I think it's cool as heck, just pointing out that getting started on RPi is not really as painful as it once was. You can even purchase an SD card preloaded with Raspbian!

1. https://learn.adafruit.com/raspberry-pi-zero-creation/overvi...

I would guess they have injectors for the filling at the centers of those cells. A Voronoi edge is equidistant from the two points it bisects, so it kind of makes sense that this would happen naturally if the filling is injected at a constant rate.

If you like this, I recommend checking out Learn You Some Erlang (for great good) http://learnyousomeerlang.com/content

It's for sale on Amazon: https://www.amazon.com/Calculus-Made-Easy-Silvanus-Thompson/...

Cheapest? A Chinese radio off Amazon, e.g., https://www.amazon.com/BaoFeng-UV-5R-Dual-Radio-Black/dp/B00.... This will allow you to tune in to local repeaters and listen. N.B., you cannot legally transmit without a license which is easy and cheap to obtain http://www.arrl.org/getting-licensed.

After getting licensed, you'll probably get frustrated with the UV-5R and look into a more capable hand held unit, or a mobile radio. These start at around $150 and go up from there (plus antennas, power supplies, etc).

Was this done using the DVDs? I'm curious about any potential licensing issues with the screen caps and subtitles. Did you have to get permission/sign something - or does this fall under fair use?

I'm skeptical how a small form factor oven like this could provide even heating that is required for many kinds of baking, even if it has convection fans. Also, having such a small volume means that most of the heat will escape when the door opens making it difficult to get that crucial "oven spring" for bread baking.

It seems like a perfectly acceptable (albeit pricy) alternative to a toaster over, however.

Lucene has had this capability for a while now. A regexp query (as defined by the RegExp class) is compiled into an automaton for quick comparison with the inverted index, also represented as an automaton. Automatons and FSTs are used extensively by Lucene these days. Out of curiosity, I did some playing around with Lucene's RegExp class when it first came out (or rather, when I first learned about it). It provides a really interesting way to build regular expressions: https://gist.github.com/mumrah/6104234

Javadoc: https://lucene.apache.org/core/4_0_0/core/org/apache/lucene/...

Source: https://github.com/apache/lucene-solr/blob/trunk/lucene/core...

My guess would be that they went with C to make it easier to hook into existing data systems like MySQL or Redis.

If you want sequential IDs with no chance of collisions, read up on "flake" IDs

* https://blog.twitter.com/2010/announcing-snowflake

* http://engineering.custommade.com/simpleflake-distributed-id...

* http://boundary.com/blog/2012/01/12/flake-a-decentralized-k-...

* https://github.com/mumrah/flake-java

Neat! I made one of these in Java a while back after reading the snowflake Twitter post (which has since been removed). Looks like Twitter has retired Snowflake: https://github.com/twitter/snowflake

https://github.com/mumrah/flake-java

Anyone know how this compares to the FST/DFA stuff in Lucene 4.x?

Part of the point of things like Chromecast is to avoid having to hook things up to a TV.

It's a better experience when you're dealing with a complete file and not an in-progress transcode anyways. With a file you can seek and pause/resume. That's much, much harder with a stream (if at all possible). The other trick is that the Chromecast does some poking around in the stream to figure out the total time. This requires that you support Range requests and return correct Content-Length headers.

I got something similar working in Java by simply exec'ing ffmpeg and getting a handle on the stream. Not being able to seek kind of sucks, and Chromecast does not like infinite streams. However, I've seen that he Chromecast will actually stream some MKVs just fine (so long as the underlying codecs are supported).

Looks nice! I've been using https://github.com/googlecast/CastHelloVideo-chrome plus some custom Java code to stream video content for a few months now. I'd love to have something like this as a UI instead of my cobbled mess. Drop me a PM or email @gmail if you want to compare notes. Cheers!

New releases of iOS usually go hand-in-hand with a new device. They design the OS to be optimal for the latest and greatest hardware, not so much the previous generations. As much as I like a good tinfoil hat theory, I'd say this is just a case of Apple adding new features (eye candy, mostly) and not worrying so much about the performance on older devices.

Guava has a nice explanation: https://code.google.com/p/guava-libraries/wiki/UsingAndAvoid.... Essentially, it forces you to actively think about the "absent" case when programming.

Very cool solution. Out of curiosity, how many iterations complete before finding the solution?

If you don't want to mess with ctypes, this sounds like a straightforward use case for namedtuple.

At LinkedIn, they started code naming projects after authors and literary works. E.g., Kafka, Camus, and a bunch of Harry Potter references. Kafka was donated to Apache by LinkedIn and the name stuck.

Use cases (from the project page): http://kafka.apache.org/documentation.html#uses

Java I/O is plenty fast. Just take a look at Hadoop, Lucene, Jetty, Netty, or any number of other pieces of high performance software written in Java.

Here's a good writeup talking about sequential I/O in Java: http://mechanical-sympathy.blogspot.com/2011/12/java-sequent...