Log of the #nice channel on irc.freenode.net

Using timezone: Central European Time
* CIA-3 leaves02:40
* CIA-3 joins02:53
* arjanb leaves03:40
* CIA-3 leaves10:27
* CIA-3 joins
* CIA-3 leaves11:11
* CIA-3 joins
* arjanb joins12:17
bruxe eckel is seriously having a go at java generics12:57
<arjanb>he has written a follow up13:04
<Bluelive_>read it13:06
what do you think ?
it doesnt even mention the problem of not being able to generate an object file of a c++ style generic13:07
<arjanb>i think he has become annoyed with java while being a promoter before13:08
weird, the java generic implementation seem very 'java' ish
<arjanb>he'll do a keynote at pycon about "latent typing" that explains his mindset in these articles13:19
quite a different point of view he has: "But look at what the interface does: it weakens the type constraints by first decoupling the interface from implementation - so you only have an outline of what the type looks like, but no semantics attached."13:25
<Bluelive_>? i thought that was the point13:26
ive never seen the implementation of half the classes i use13:27
iam not sure what 'semantics' mean here13:28
* bonniot joins13:35
i saw a few comments from you on eckels blog ;)
<Bluelive_>i really think he is missing a few points14:07
he's a dynamic typing advocate, and that shows
did you read the followup article?14:10
strousoup did an interview not that long ago about making templates less implicitly typed14:11
<arjanb>i think have read that too but i can't find the link16:04
<Bluelive_>i think it was on artima16:06
btw im currently thinking about not allowing access to fields if a type is possibly null16:10
or would that cause it to be too ugly ?
<arjanb>what is possibly null the field or the object?16:18
<bonniot>that's what we do: x.f is invalid if x is possibly null16:20
<Bluelive_>that would make null something really unhandable16:24
<bonniot>no, because most types do not include null16:26
when they do, then it's because null can really happen, so you better handle it16:27
<Bluelive_>thats all fine and well when using powerfull inference16:28
Alpha isnt that powerfull yet16:29
<arjanb>why do you add the nullness thing to Alpha now?16:31
<Bluelive_>for better contracts on function parameters and a small optimalisation to not check for nullpointerexception16:33
* CIA-3 leaves16:47
* CIA-3 joins16:48
* CIA-3 leaves18:09
* CIA-3 joins
* CIA-3 leaves18:18
* CIA-3 joins
* CIA-3 leaves18:20
* CIA-3 joins
* CIA-3 leaves18:23
* CIA-3 joins18:24
* CIA-3 leaves18:28
* CIA-3 joins
* CIA-3 leaves18:31
* CIA-3 joins
* mars_ joins20:35
slt tlm
* mars_ leaves20:49
* [CAA]toutpuissan joins21:55
* [CAA]toutpuissan leaves
* CIA-3 leaves22:16
* CIA-3 joins

Generated by Sualtam