[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Need to avoid breaking abstractions? Then *DON'T*.




On Wed, 4 Jan 2006, Michael Sperber wrote:

>
> Without opacity and with reflection, a client who gets its hands on a
> record can get at its fields, thus breaking the abstractions provided
> by the maker of that record.

And this is as it should be.

The need for opacity reminds me of the old story of the fool
and the doctor.

Remember the old saw about a patient who comes in, and
while the doctor is watching, wrenches his arm out of its
socket, complaining, "Doctor, doctor, it hurts when I do this!"
And the doctor's calm, reasonable response:  "Don't do that."

We are faced with much the same situation in the desire for
opacity.  It has become the vogue for "doctors" (language
designers) to devise elaborate trusses to prevent people from
intentionally wrenching their arms out of the socket, but
these are, in some sense, silly since reasonable and intelligent
people do no such thing except in the most extreme circumstances
(say to escape a deadly trap, or facilitate foreign-function
interface) and in that case ought not be prevented.

If programmers do not want to break abstractions, then the mere
ability to break abstractions is not a problem.  And if they do
want to break abstractions, I presume that if they are intelligent
people then they have good reasons for doing so, and in that case
the language features that inhibit it are actively harmful.

				Bear