I think more people should follow and submit things to algorithms so that it goes higher.

/v/algorithms for those who can't type :P

etl/Java/Linux

I gave up and started with C++ and got that rolling. I have never used gradle and when it fails it really fails. It said it didn't show me 43 failures in a ~10 test file lol. I think it is actually the test failing but it says that it is failing with toString which I don't even have.

I can't even get the tests to work. It says error with toString(). I don't even use it and the tests don't either! Lol pretty goood...

Thanks! I saw the second link already, but I didn't see anything related to probes.

Yeah I think maybe it is the learning part that might be the issue. A very large majority of schools don't require students to really learn any parallel/concurrent programming. The most I ever saw was in the OS class but the examples are most theoretical or trivial. Oh yeah, I encounter the sleeping barbers problem all the time (in the general sense) /s.

To me it seems really strange that people are having these serious issues.

Most of the issues I always hear about stem from

• Nested mutual_exclusion_object_of_choice (MEOC) (dealock)
• Shared data with no MEOC. (races)
• No single simple point of entry (complicated)

That is just my opinion though. Maybe others have seen different issues?

DON'T TOUCH MY BLANKET!!! oh wait... wrong Linus...

Without a code of conduct does this mean I can belittle people all I want?

In theory (I tried and can't get it to totally work) you should be able to generate a huffman tree in which The encodings of "Hello World!" results in the output of "Hello World!".

There are a few subverses related to programming in some way. machine learning, compsci, computerscience, learnprogramming, etc... You can probably just think of a topic and find a subverse that exists.

Yeah you need to have created the subverse to add a new moderator it seems. You will need to post on /v/subverserequest to get control of this one. Good luck.

I can help moderate. I am the moderator of /v/algorithms

Also, maybe add some links to related subverses?

I think version 1.x you could. The new one would probably require you to extend some class.

And there is /v/computerscience.

The problem with /v/computerscience and /v/compsci is that they are very broad fields. I like algorithms and data-structures more than networking, operating systems, and databases.

I think places like /v/programming put more emphasis on the day to day programming and software development aspects than the more theoretical aspects of CS.

Too bad you don't have two versions, one for the white background and another for the Nightmode. If they have not been active in over a year you should request to take over [link]

I guess one thing is to make sure we are talking about the same version of Hadoop.

I have really only used version 2.x and don't know of a way to control where the data is placed (other than changing the source code for HDFS). I tried to find out and the only thing I found (in 5 minutes) was a post on StackOverflow and this blog post.

The HDFS is actually on each node. You have to place the data in the HDFS and it is distributed across the nodes.

For the mapper portion you can do a log based reduction I do believe. So you would have half your nodes stop doing anything and the other half merging it and one of the finished nodes. Then half those finish and the cycle continues until you have a single node.

I helped some exchange students over the summer setup and modify Hadoop to use a real-time scheduler.

Here is my high level description (I live leave the in-depth for others)

High Level Description

Hadoop is software which is installed on multiple machines for the purpose of running jobs which can be divided into multiple parts.

What is it for

You have a problem and it takes a long time to solve. Luck for you that your problem can be broken up into parts and worked on separately.

Hadoop wants

You write a program that describes how to break your problem up (map) and also how to put the different solutions back together (reduce).

How

Hadoop takes your program and uses the map portion to tell a machine what they should do. When done the reduce portion is used to take these partial solutions and create the actual solution you want.

Guess we will see. Programming already has 13K subscribers which is pretty cool.

Maybe? Most of what I have seen on /v/programming seems to have little to do with data-structures/algorithms though.

I hope people come and give data-structures and algorithms some much needed love.