What's the point of journals these days, especially in computing with reproducible results (i.e. authors publish a paper on arxiv, code & dataset on github)?
Designing your code and data so it's truly reproducible is a ton of work, and most researchers are focused on the paper.