The Perl 5 object system is among the most flexible in the programming world.
Maybe if you compare it to C++ and Java. Sure, it's
Turing-complete, so theoretically
anything could be implemented in it. But I prefer to
keep my sanity. The Perl6 object model will have a
number of substantial improvements, and I'll be
looking forward to that. Yes, I know it will be a
while yet, but as the Apocalypsen roll in, more and
more of the design (at least) will be clear. It may
even be possible to start before Perl6 is
released as such.
for(unpack("C*",'GGGG?GGGG?O__\?WccW?{GCw?Wcc{?Wcc~?Wcc{?~cc'
.'W?')){$j=$_-63;++$a;for$p(0..7){$h[$p][$a]=$j%2;$j/=2}}for$
p(0..7){for$a(1..45){$_=($h[$p-1][$a])?'#':' ';print}print$/}
| [reply]
[d/l] |
Maybe if you compare it to C++ and Java.
Or Ada. Or Python. Or Ruby. Should I list all the languages I know? Would that make Perl's OO support any more or less flexible?
You didn't answer my question. What precisely is missing from Perl 5's object system that is keeping you from implementing buffers now?
Don't get me wrong, I'm looking forward to Perl 6 too. But most of what I've seen looks like syntactic sugar. Are there really applications that can be written in Perl 6 that simply cannot be written in Perl 5? I won't believe it until you produce concrete evidence.
-sam
| [reply] |
Or Ada. Or Python. Or Ruby.
I tried Python and didn't like it, although I didn't
get far enough to know how strong the object model is,
because the B&D type system turned me right off.
I can't comment on Ada or Ruby except to say that
neither is known for being object oriented. Try
programming a medium-sized project in Inform for
example, and then come back and look at Perl5's
object system.
Are there really applications that can be
written in Perl 6 that simply cannot be written in
Perl 5?
Why, are you in some kind of special hurry?
No, of course not. For that matter there is nothing
that can be written in Perl 6 that simply cannot be
written in assembly language. Congratulations, you've
discovered Turing completeness. It's a matter of how
much time and effort I want to spend creating a
module to make programming easier in a language
that in three years will be obsolete. (By the time
I finished we'd be more than halfway there... it
won't take me one quarter as long to do it in Perl6.)
What precisely is missing from Perl 5's
object system that is keeping you from implementing
buffers now?
Just for starters how about sane syntax, a decent
inheritance mechanism, traits and properties,
and the ability for methods to determine exactly
what context they've been called in, and a lot
of stuff that makes anonymous routines easier
to work with? How about the rest of the language
being more geared toward expecting stuff to be
object-oriented? How about that Perl5's object
system is such a convoluted hack that I only
barely understand it, whereas just from
reading the Apocalypse articles I already
have a very firm grasp of Perl6's object model,
because it makes sense.
Yeah, I could fake a lot of that with existing
mechanisms (hashes, especially), but it'll be a
whole lot easier when it's built into the language.
Perl5 is pretty cool, but Perl6 is going to rock
significantly.
for(unpack("C*",'GGGG?GGGG?O__\?WccW?{GCw?Wcc{?Wcc~?Wcc{?~cc'
.'W?')){$j=$_-63;++$a;for$p(0..7){$h[$p][$a]=$j%2;$j/=2}}for$
p(0..7){for$a(1..45){$_=($h[$p-1][$a])?'#':' ';print}print$/}
| [reply]
[d/l] |
