I've been digging into old p5p mails to research some of our bugs and
Policies, and I found this message from Larry from back in the day. I
think it is an interesting statement that we should keep in mind when
we consider the subject of backwards compatibility in various
contexts.
From message id: 20080118012232.GA24932@wall.org
--------------------------------------------------------------
Perl has broken backward compatibility in the past where I judged
very few people would be affected. This feels like that to me.
If we always put our thumb on the scales on the side of backward
compatibility, we never change anything. There's a fair balance to
be struck here, because most of the escape routes to the future
*might* have been used by someone already for some other purpose.
For instance, someone might already have defined a private "feature"
pragma that is inconsistent with the new one. We shouldn't let that
level of probability slow us down.
--
perl -Mre=debug -e "/just|another|perl|hacker/"
Policies, and I found this message from Larry from back in the day. I
think it is an interesting statement that we should keep in mind when
we consider the subject of backwards compatibility in various
contexts.
From message id: 20080118012232.GA24932@wall.org
--------------------------------------------------------------
Perl has broken backward compatibility in the past where I judged
very few people would be affected. This feels like that to me.
If we always put our thumb on the scales on the side of backward
compatibility, we never change anything. There's a fair balance to
be struck here, because most of the escape routes to the future
*might* have been used by someone already for some other purpose.
For instance, someone might already have defined a private "feature"
pragma that is inconsistent with the new one. We shouldn't let that
level of probability slow us down.
--
perl -Mre=debug -e "/just|another|perl|hacker/"