Quotable quote from Matz

[Hal E. Fulton]

This caught my eye and I wanted to make sure everyone
saw it. (I'm posting this without Matz's permission --
hope that's not rude.)

According to Matz:

Ah, please stop mentioning POLS whenever your personal
expectation is not satisfied. I'm sick of them.

Others may also be sick of this. But that is of no
consequence. However, if *Matz* is sick of it, we should
all pay attention.

Did you know POLS spelled backwards is SLOP?

Cheers,
Hal

 

[Shashank Date]

One can also refer to this slide from Matz's OSCON2003 presentation:

http://www.rubyist.net/~matz/slides/oscon2003/mgp00043.jpg

Or, to paraphrase another of Matz's lines from the past: "'Principle of
Least Surprise' means least surprise to *me*"

Yes, the *me* part needs to be underscored heavily. Otherwise, it is
ambiguous and people are naturally going to interpret it in their own
context.

I can understand why Matz feels sick when he is reminded about that
principle in the wrong context, but I take it as a tribute to his genius. He
is doing his job so well (of sticking to that principle) that people start
expecting the impossible (and I don't mean this in a bad way: I think it is
very obvious to fall in that trap ...we are all heavily biased by the
language that we know best. Ok, may be not all ;-)).

Certainly, if the *me* part is stated in the correct fashion (like, * for
the unbiased*) _in_the_same_breath_, then hopefully such "false positives"
will die down. Until then, Matz will have to chalk it up to "one more battle
wound to victory".

$0.02

--shanko

 

[Steven Lumos]

Yes, the *me* part needs to be underscored heavily. Otherwise, it is
ambiguous and people are naturally going to interpret it in their own
context.

The real problem is that THE Principal of Least Surprise is used in
the first place. That implies a universal. Some quick google
searches for POLS and POLA imply that most people consider it to apply
the the users and not the developer/language designer.

If anyone is really bothered that much by it, use a more clear name,
like DSM (Doesn't Surprise Matz), or ITM (Intuitive to Matz).

Besides, wouldn't it be much more fun for people to post "DSM
violations" which then go on to argue why Matz should be surprised?

Steve

 

[Hal Fulton]

The real problem is that THE Principal of Least Surprise is used in
the first place. That implies a universal. Some quick google
searches for POLS and POLA imply that most people consider it to apply
the the users and not the developer/language designer.

If anyone is really bothered that much by it, use a more clear name,
like DSM (Doesn't Surprise Matz), or ITM (Intuitive to Matz).

Besides, wouldn't it be much more fun for people to post "DSM
violations" which then go on to argue why Matz should be surprised?

Well, obviously the original meaning implied a universal. But it's
never really universal, because we don't all agree.

It's only a guideline or a cute way of saying that software ought to
be intuitive. There has never been an objective way of measuring
surprise. Someone will disagree with that, too.

I think the point is: Express your preference, but don't pretend it's
universal. Little if anything would be unanimous if put to a vote on
this group (or any other).

Or another point: You can't try to "push" Matz into changing something
(solely) on the basis that you think it's surprising. ("You" meaning
the generic "you" or "we.") Firstly, it won't work; and secondly, it's
gotten tiresome.

Or to put it more succinctly:

Matz.respond_to?pols) # false
# Or perhaps he responds but raises an exception

By the way: I'm always curious to know whether anyone can find a
verifiable reference to this principle earlier than the Geoffrey James
quote (in _The Tao of Programming_) about the Law of Least Astonishment.
That was '86 or so. If someone finds an earlier one, please email me.

Cheers,
Hal