In quite a few cases it's possible to execute one-liners autonomously, for example from a crontab(5). But I'm assuming you mean you'd rather have a script file, which is easier to version/distribute/install/etc.

To see what the one-liner is doing, have a look at perlrun to see what the -p switch does, or you could even add -MO=Deparse to the perl arguments to see what B::Deparse makes of the one-liner:

$ perl -MO=Deparse -pe 's/^(?:L\d+|DETAIL|SPACE|SUMMARY),//'
LINE: while (defined($_ = <ARGV>)) {
    s/^(?:L\d+|DETAIL|SPACE|SUMMARY),//;
}
continue {
    die "-p destination: $!\n" unless print $_;
}
[download]

Simplifying that:

while (<>) {
    s/^(?:L\d+|DETAIL|SPACE|SUMMARY),//;
    print;
}
[download]

While it hopefully helps explain a little more what the one-liner is doing, and code is even something you could put in a file and have work, it's still not really "better" than just a one-liner, as it still lacks warnings, error checking, and it still writes to STDOUT. Here's how one might write the same functionality as a simple script with command-line arguments and a bit more error-checking:

#!/usr/bin/env perl
use warnings;
use strict;
die "Usage: $0 INFILE OUTFILE\n"
    unless @ARGV==2;
my ($infile,$outfile) = @ARGV;
open my $ifh, '<', $infile or die $!;
open my $ofh, '>', $outfile or die $!;
while (my $line=<$ifh>) {
    $line=~s/^(?:L\d+|DETAIL|SPACE|SUMMARY),//;
    print $ofh $line;
}
close $ofh;
close $ifh;
[download]

Note that the core functionality, the regex, remains the same. Also, this script does not make use of the magic ARGV handle (the <> operator) like the one-liner; it makes the input less magic and more explicit, although sometimes the more magic and also more flexible ARGV is preferable when writing UNIX-style scripts. TIMTOWTDI.