3 ms·

I've been watching Go from the sidelines for 2 years now. I do data engineering work and most of that is on the JVM (Spark, Scalding, Hadoop, Kafka). I wonder i

by josep2 10y ago

I've been watching Go from the sidelines for 2 years now. I do data engineering work and most of that is on the JVM (Spark, Scalding, Hadoop, Kafka). I wonder if there will ever be a push for Go in that space. I know NSQ has been implemented with some good success but I see a lot more systems and web server articles about go thank any data engineer or scientific computing.

goatlover 10y ago
Why would you pick Go over Julia for that domain? Julia was made specifically for scientific computing, it's performant, has better array and number support, and allows unicode for mathematical variable and function names. It also has excellent concurrency like Go, but most of the work on the language and libraries is centered around data engineering and scientific computing. People can use whatever language works for them, and Go has it's plusses, but I would hate to see popularity get in the way of a good new language like Julia in a domain it was intended for. It would be like if R were relatively new, and everyone wanted to use Javascript instead for statistical computing. Use the right tool for the job.
- josep2 10y ago
  I've looked at Julia once or twice but at the time it wasn't mature enough to do what I needed. Sounds like I should give it another look though.
- IndianAstronaut 10y ago
  People want one tool to do all their work. Which is why Python took off in the ML and scientific computing space. The same language you use for data analytics is the same one you use for ETL, building a web backend, interfacing with the system, etc.
elcct 10y ago
Have you seen this talk? https://www.youtube.com/watch?v=D5tDubyXLrQ https://www.youtube.com/watch?v=D5tDubyXLrQ
- josep2 10y ago
  I did see this. While interesting, it's pretty rudimentary in the Data world. Google is supposedly developing a TensorFlow API for Go, would be interesting to see deployments of that.
tapirl 10y ago
Although I haven't tried voth, I feel NATS is better than NSQ from the design view. Golang big data world still lacks of the core components, such as Hadoop and Spark in Java.
IndianAstronaut 10y ago
I am starting to use Go in the data engineering space. I use it for ETL processes I am running. I see Go right now at the stage Python was in 2009 in regards to scientific computing. The libraries are coming into place and the community enthusiasm is there.

by josep2 10y ago

goatlover 10y ago

Why would you pick Go over Julia for that domain? Julia was made specifically for scientific computing, it's performant, has better array and number support, and allows unicode for mathematical variable and function names. It also has excellent concurrency like Go, but most of the work on the language and libraries is centered around data engineering and scientific computing. People can use whatever language works for them, and Go has it's plusses, but I would hate to see popularity get in the way of a good new language like Julia in a domain it was intended for. It would be like if R were relatively new, and everyone wanted to use Javascript instead for statistical computing. Use the right tool for the job.

josep2 10y ago
I've looked at Julia once or twice but at the time it wasn't mature enough to do what I needed. Sounds like I should give it another look though.
IndianAstronaut 10y ago
People want one tool to do all their work. Which is why Python took off in the ML and scientific computing space. The same language you use for data analytics is the same one you use for ETL, building a web backend, interfacing with the system, etc.

elcct 10y ago

Have you seen this talk? https://www.youtube.com/watch?v=D5tDubyXLrQ https://www.youtube.com/watch?v=D5tDubyXLrQ

josep2 10y ago
I did see this. While interesting, it's pretty rudimentary in the Data world. Google is supposedly developing a TensorFlow API for Go, would be interesting to see deployments of that.

tapirl 10y ago

Although I haven't tried voth, I feel NATS is better than NSQ from the design view. Golang big data world still lacks of the core components, such as Hadoop and Spark in Java.

IndianAstronaut 10y ago

I am starting to use Go in the data engineering space. I use it for ETL processes I am running. I see Go right now at the stage Python was in 2009 in regards to scientific computing. The libraries are coming into place and the community enthusiasm is there.