That's depressing - I hadn't seen that. I had also come to the same conclusion about command numbering, but had decided I wouldn't change it. However, I suppose it has the same relationship as Orthogonal does to Befunge. --(this comment by Marz at 10:01, 25 August 2007 UTC; please sign your comments with ~~~~)
I wonder how long is the shortest possible implementation of a Turing-complete language in MGIFOS. I was just thinking maybe such a thing could be used to get around memory/disk space limitations by delegating to another language. Of course, any of the BF self-interpreters could be trivially translated into MGIFOS, but can anyone do better? :) — Smjg (talk) 15:49, 29 March 2012 (UTC)
What if they were numbered with base-9 1-8 instead? Would that make it any more efficient?
"><" ->19 = Less efficient by <0.03 orders of magnitude "+" ->378_9=314=More efficient by ~0.45 orders of magnitude ",[.,]" -> 67568_9=44936 =More efficient by ~0.97 orders of magnitude