10:59 <Hodgestar> Question for language academics --in a scientific paper, would it be more accurate to say Julia uses JIT or JAOT compilation? Is JAOT distinguished from, say, C++ templating, in academic literature? And does anyone other than the Julia community use the term? Does JIT cover what Julia does, or not?

11:00 <Hodgestar> (I am helping some Julia coders write a paper and want to not sound silly -- although I don't know if any physicists will notice or not).

11:02 <nimaje> what should that J stand for in JAOT? AOT is "ahead of time" I guess

11:03 <nimaje> is it http://www.apogee.com/resources/jaot ?

11:21 <Hodgestar> nimaje: The "J" is for "Just" -- i.e. Just Ahead Of Time.

11:23 <Hodgestar> nimaje: My semi-informed opinion is that how Julia compiles is somewhat similar to the Java JAOT option you linked to.

11:23 <Hodgestar> nimaje: Also, thanks!

11:28 <nimaje> so, that JAOT stuff does slow, interpreted analyse runs and later does AOT based on the analyse runs?

12:27 <cfbolz> Hodgestar: that's really a matter of definition

12:28 <cfbolz> you can define "JIT" as "a compiler that executes at runtime"

12:28 <cfbolz> then Julia does JIT

12:28 <cfbolz> or you can define "JIT" as "a compiler that executes at runtime and incorporates some time of runtime feedback", which julia doesn't do

12:32 <fijal> yay, taxonomy!

12:33 <fijal> my pet theory is that no matter what field are you exploring, there is someone absolutely obsessed with taxonomy

16:39 <Hodgestar> cfbolz: Thanks!

16:52 <jean-paul[m]> the hypothesis test suite fails on pypy with a bunch of "zoneinfo._common.ZoneInfoNotFoundError: 'No time zone found with key UTC'" exceptions. anyone have any ideas about this?

17:33 <jean-paul[m]> It seems adding tzdata as a test dependency fixed it.