Program committees and the conferences they serve may be part of the problem, and hence part of the solution as well. Instead of picking the best N papers so a

Program committees and the conferences they serve may be part of the problem, and hence part of the solution as well. Instead of picking the best N papers so as to fill out a conference schedule, pick the good papers and shorten the conference schedule if the number of good papers is <N. And then raise the standards to meet your expectations of reproducibility, code reviews, unit tests, etc. If a big conference like ISCA were to be shortened by one day by omitting the least worthy papers, you'd see much better work arriving the next year.