Hacker News new | ask | show | jobs
by sirsar 4446 days ago
LabVIEW is concurrent by default; control flow is done by linking the outputs of one function to the inputs of another. This makes writing concurrent loops ridiculously easy: just put two loops next to each other.

I rarely use it because organization is such a pain, but its "data-flow" paradigm does simplify a lot of logic.

3 comments

Happy to see another LV user here.

Organization gets much, much easier once you get the big architecture concepts - produce/consumer patterns, messenger-queues, event structures, actor framework. Yes, it can be painful, but the newer functionalities (clean-up, riddiculously easy sub-vi creation,..) subtly improve it.

Btw: If you are using LabVIEW and have never used Quickdrop, try it. It is the most underutilized and amazing feature.

Another example would be Bloom language. Here is a great talk explaining motivation for using "disorderly" languages in distributed systems: http://vimeo.com/53904989
Yeah there's a pretty large class of data processing tasks that are made pretty trivial by LabVIEW-likes visual dataflow languages. One I've seen for XML work that was pretty good was called "SDE"

http://www.futurepointsystems.com/?page=products