On Sun, 25 Oct 1998 chip(_at_)perlsupport(_dot_)com wrote:
According to Joshua Pritikin:
On Sat, 24 Oct 1998 chip(_at_)perlsupport(_dot_)com wrote:
"No unreal content exists."
Laws by fiat are stupid. Drawing no conclusiions yet tempered by
experience is smart.
Um, I've been manipulating content and metadata with Perl for over a
decade. So that conclusion *was* tempered by experience.
But that's backwards! When I look at perl, I see that much is generic and
also quite a few short-cuts. The interesting part is deciding what is
important enough to earn a short-cut or CORE-level hook. The ?? operator
is a good example: _it_evolved_. There was a good deal code written
without ??. But people noticed they were typing more than they thought
they should. There was a great deal of pondering and we settled on ??.
To me, this is the right way to design things (code with existing tools
then ponder short-cuts).
So what I am trying to say is that when you declare "meta-data must be
separate from text", it seems to me like you are skipping steps whether
you have 10 years of experience or not. Where is the prototype
implementation that doesn't need core language support? Granted, it is
worth a little brainstorming to make sure that perl6 can swallow
absolutely every useful idea. But as I mentioned before, it is better to
_design_for_change_ than to try for a perfect, all encompassing design.
Now I'm not sure how much you have got implemented at this point, but it
would seem more relavent to study sneaky ways to automate C++ binary
backward compatibility. Or maybe we have enough CPU to ask people to
recompile everything for each new release? If so, we'll probably need a
module manager to automatically keep all the binaries in sync?
Yet never allow doubt to delay action.
Yes, you too can have an aphorism for any occasion!
I wasn't just being cute. :-(
