Plurk

57 responses to this plurk (Jump to bottom)

  • shaf
    oh please don't make start having to code c# for living :-(
  • shaf
    'make me' I mean ;-)
  • Per says
    so now they are done taking features from Haskell, I wonder what language is next...
  • Schwarz says
    Intercal maybe?
  • Per says
    I am hoping for Perl or PHP personally!
  • Schwarz says
    yes, they shold include the PHP way of OOP ;-)
  • Ilyazz
    they need to do smth, as c# doesn't have any growth in popularity
  • amix
    i root more for PHP's equality relation: "false" == true -> true while "false" == false -> false :-P
  • Per says
    "anystring" == 0 > true is more fun
  • Per says
    anyways, just use === if you want strict equality :-)
  • amix
    actually it was "false" == 0 i thought about :-)
  • Per says
    that "any string equals 0" is pretty hard to debug when you encounter it the first time, trust me.. ;-)
  • amix
    yea, i could imagine (s_LOL)
  • Spand says
    Personally, I cant wait till I can do <?php ?> in C#, or maybe <?jscript ?>. oh the opportunities ! :-P
  • mxstone says
    I code in c# and I digg it as a language, I think a lot of the things in 3.5 make for lazy programming though.
  • mxstone says
    but the use of generics in c# is far more limited than their java brethren IMHO
  • amix
    i thought c#'s generics were more powerful than java's (because they are available at runtime) (s_unsure)
  • Telperion is
    not very familiar with C# :-P
  • amix
    knienmczak: then they are more powerful in C# as in java they are only available on compile time
  • Schwarz says
    amix: There is no F-bounds on generics in C# so you cannot express existensial types like you can in Java
  • Schwarz says
    The static part is far less expressive than the Java counterpart
  • Schwarz says
    but of course the runtime availabilty is great to have
  • Schwarz says
    Never understood why it isn't in java
  • amix
    maybe because that would require big changes in the JVM and Java bytecode (s_unsure)
  • Schwarz says
    well you could just put the type in a field
  • Schwarz says
    and let the compiler tie it together
  • Schwarz says
    that is how inner classes work...
  • amix
    jea, that could work. weird that they don't add it as it would open up for a lot of hackery and make reflection easier
  • Schwarz says
    basically yes
  • Schwarz says
    But in java the model is not that a type is passed as a parameter
  • Schwarz says
    Java generics are more expressive than that
  • Schwarz says
    that is what we just talked about. I think they should add the reflection to java
  • Schwarz says
    but it would be harder to use in java since the type parameters do not have to be proper types like in C#
  • sweetestsmiles says
    will consider migration if is useful for mathematical / computation work and others
  • Schwarz says
    for example you can say List<String> in both java and C# is you say
  • Schwarz says
    But in java a type can also be List<? extends String> which is not a proper type
  • Schwarz says
    that is slightly harder to make sense of through reflection, even if you had the type available
  • Schwarz says
    that may be why you can't in Java...
  • Spand says
    Pretty sure the official reason is the bytecode change.. and not because something wouldnt be possible via reflection :-)
  • Schwarz says
    yes the official reason was the bytecode change
  • Schwarz says
    It isn't that it wouldn't possible to do reflection. It is just that it would be hard to create a good API since the type system is
  • Schwarz says
    actually very complicated
  • Schwarz says
    most people only scratch the surface of the java generics
  • Spand says
    Well its not like it shields implementation details from the user now anyways.. It would almost be a shame to make it easy to use now :-P

Ads