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

Re: Interfaces

Actually, my proposal is a mutant hybrid of 'define module' and
'define generic'.  Using it, you just stop using 'define generic'.

-----Original Message-----
From: P T Withington <ptw@callitrope.com>
To: Scott McKay <swm@mediaone.net>
Cc: info-dylan@ai.mit.edu <info-dylan@ai.mit.edu>
Date: Wednesday, July 05, 2000 2:32 PM
Subject: Re: Interfaces


>Scott McKay swm@mediaone.net on  2000-07-03 23:15 wrote:
>
>>
>> Kim Barrett wrote in message <200007031924.PAA16768@life.ai.mit.edu>...
>>
>> [Quote from Moon]
>>
>>> Dylan does not require programmers to make a textual copy of portions
>>> of the definitions of items exported by a module, like a C header file
>>> or an  Ada package declaration.  In my opinion Dylan modules still have
>>> a  well-defined interface; the difference is that we assume there will
>>> be the  obvious development tools for printing out that interface,
>>> instead of requiring programmers to pretend to be tools and copy text
>>> from one file to  another.  (Except of course Dylan still requires
>>> programmers to copy the -names- of the exports from their definitions
>>> into the define module statement.  We haven't thought of a good way to
>>> avoid that.)
>>
>>
>> If you want to declare an "interface" in Dylan right now, you copy the
>> names of functions into a module definition, and you copy the signature
>> from some set of methods into a 'define generic' form.  So Dylan does,
>> in fact, require you to make "a textual copy of portions of the
definitions
>> of items exported by a module" if you want to collate this into an
>> "interface".
>> And you get to copy a little bit into one place, and another bigger bit
>> into another place.  So I find Moon's argument less than persuasive...
>
>I would consider the generic's signature to be the original, not a copy.
>Then you implement the interface by writing methods that "subset" that
>signature.  A default method might copy the signature of the generic, but
>any more specific methods must have a more constrained signature.
>
>> I have in the past proposed deprecating 'define generic' and augmenting
>> 'define module' to take its place.  I haven't thought about this for a
year,
>> and I haven't written a real proposal, so don't bother to quiz me on it.
>
>I suppose if you could write the generic in your module declaration then
you
>wouldn't have to copy the name, as Moon complains.  But you would still
need
>'define generic' for intra-module use, would you not?
>