The problem is, really, that Perl 5 and Perl 6 are two completely separate languages. You shouldn't even bother thinking about backwards compatibility; the developers aren't. Larry stepped on that before much of anything else.
From what I've read, Regexen are staying _basically_ similar (basically), but Damian Conway being the sick evil bastard he is, they're going to incorporate some higher capabilities into the engine that look an awful lot like his Parse::RecDescent grammar parsing module.
I mean, really, at their base, he's moving around flags a little, making whitespace default insignificant, and tinkering a few flags. It's when he starts making rules and backtracking capabilities (which exist in the current language, in clunky form, as he demonstrated) that it starts getting wacky.
What it comes down to, though, is that noone should be thinking 'Oh, man, now I'll have to port all my perl to perl 6.', because Perl 5 isn't going to die out or anything, just get phased out like perl 4 - left at a stable rev for support purposes.
Dealing with this kind of ethical quandary is all too often the system administrator's job - or at least our problem - because we are point of contact. If not us, who?
So far as a code of ethics goes, I feel the ACM Code of Ethics is sufficient to the needs of a sysadmin. It's not legally binding so far as I know, even if you are an ACM member, but in the absence of upper management restraining your ethical action, it is a good method for judging when to do what.
Other than that, I can only offer one caveat to the concept of instantly informing the users. Often after a break-in, some tactical delay is required to make sure we've identified all the breakin locations. Only after we're pretty damn sure do we step in and shut things down. On private systems informing the user isn't harmful, but when the information goes out via, the hazard of an annoyed cracker deciding to break something since he's on the way out anyway is too dangerous.
Of course I'm not claiming that not informing is an option, but I think a delay of up to a full working day is not unwarranted. Long enough to tie things down.
Slashdot Mirror
The problem is, really, that Perl 5 and Perl 6 are two completely separate languages. You shouldn't even bother thinking about backwards compatibility; the developers aren't. Larry stepped on that before much of anything else.
From what I've read, Regexen are staying _basically_ similar (basically), but Damian Conway being the sick evil bastard he is, they're going to incorporate some higher capabilities into the engine that look an awful lot like his Parse::RecDescent grammar parsing module.
I mean, really, at their base, he's moving around flags a little, making whitespace default insignificant, and tinkering a few flags. It's when he starts making rules and backtracking capabilities (which exist in the current language, in clunky form, as he demonstrated) that it starts getting wacky.
What it comes down to, though, is that noone should be thinking 'Oh, man, now I'll have to port all my perl to perl 6.', because Perl 5 isn't going to die out or anything, just get phased out like perl 4 - left at a stable rev for support purposes.
Dealing with this kind of ethical quandary is all too often the system administrator's job - or at least our problem - because we are point of contact. If not us, who?
So far as a code of ethics goes, I feel the ACM Code of Ethics is sufficient to the needs of a sysadmin. It's not legally binding so far as I know, even if you are an ACM member, but in the absence of upper management restraining your ethical action, it is a good method for judging when to do what.
Other than that, I can only offer one caveat to the concept of instantly informing the users. Often after a break-in, some tactical delay is required to make sure we've identified all the breakin locations. Only after we're pretty damn sure do we step in and shut things down. On private systems informing the user isn't harmful, but when the information goes out via, the hazard of an annoyed cracker deciding to break something since he's on the way out anyway is too dangerous.
Of course I'm not claiming that not informing is an option, but I think a delay of up to a full working day is not unwarranted. Long enough to tie things down.