Front page | perl.perl6.language |
Postings from April 2003
--- Jonathan Scott Duff <duff@cbi.tamucc.edu> wrote:
> On Wed, Apr 09, 2003 at 06:05:47PM -0700, Larry Wall wrote:
> > Though the danger of treating patterns as too powerful for just
> > text is that we run the risk of making patterns too powerful for
> text.
> > However lofty we aspire to get, we gotta keep our feet on the
> ground.
>
> I agree completely. But I think this is an area where hard things can
> be possible.
>
> The default rule engine should be optimized for homogeneous strings
> composed of characters as that's the common case. But if we can tell
> perl that our "string" is heterogeneous or that it's composed of
> object
> "characters" then it's free to use a different rule engine that
> doesn't make those assumptions and we've just made some nifty things
> possible.
Grammar WallStreet;
rule day
{
/./ # Expects database stock-price data. See class Ticker.
}
rule trend(&xer_than)
{
<day> { $0.min = $1.min; $0.max = $1.max; }
(
<day> <( &xer_than($1.min, $0.min)
&& &xer_than($1.max, $0.max) )>
{ $0.max = $1.max; }
)+
}
my $up_trend = rx/<trend(infix:>)/;
my $down_trend = rx/<trend(infix:<)/;
my $buy_signal = rx/<$down_trend><$up_trend> <( $2.length == 1 )>/;
my $sell_signal = rx/<$up_trend><$down_trend> <( $2.length == 1)>/;
broker()
{
my $wire = new Stock::Ticker('NYSE');
given $wire
{
when <$buy_signal> { buy(); }
when <$sell_signal> { sell(); }
}
}
Nifty is one way to put it...
=Austin
Thread Previous
|
Thread Next