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

Re: SRFI naming



>>>>> "al" == Alben Barkley Petrofsky <alben-barkley@petrofsky.org> writes:

>> From: sperber@informatik.uni-tuebingen.de (Michael Sperber [Mr.  Preprocessor])
>> >>>>> "al" == Alpine Petrofsky <alpine@petrofsky.org> writes:

al> They are the shortname serial numbers.  Did you ever read the message
al> that started this thread?
>> 
>> Yes, I did.  I'm sorry if I expressed myself poorly: I know what it's
>> supposed to mean, but I fail to see how it's easier to rememember than
>> the SRFI serial number.

al> I'll try answering that again.  I have a much easier time remembering
al> whether something was numbered 1 or 2 than I do trying to remember
al> whether it was numbered 1 or 2 or 3 or 4 ..... or 97.  

That's reasonable, but I, personally don't.

Here's a conjecture: in fact, you are most likely to remember the "1"
in preference to other serial numbers.  (Because you always knows it
exists, for one.) If in fact, there are, say, 3, the "3" (and
especially the "2") is just as difficult to remember as any other
serial number, especially now that you have to remember the name with
it.  The potential for confusion is pretty big, I'd say, because it's
going to mnemonically difficult to remember what exactly the
difference between LIST-LIB-1 and LIST-LIB-2 is.

al> I can use the name "Snookums", or any name I like, to refer to the
al> 43rd U.S. president, but I can't expect anyone to understand me
al> because "Snookums" is not a publicly-recognized name for him.
al> Fortunately, our presidents do have recognized names, and the
al> identifier bush-2 suffices to distinguish him from all past and future
al> presidents.  (Note that the names clinton-1 and bush-2, even though
al> they are not at all descriptive, are much easier to remember than 42
al> and 43.  Also, even though presidents johnson-1 and johnson-2 have no
al> particular connection to each other, these names are still superior to
al> 17 and 36.)

This example is exactly where I disagree with you: To identify
"johnson-2" you need to remember that there's a "johnson-1".

But let me again state: if you want some sort of mnemonic naming
scheme, by all means, write it up, and, if you want, submit it as a
SRFI.  It seems people are mostly expecting the SRFI editors to take
care of this chore.  There are several reasons why we presently don't:

- A unique naming scheme can serve its function if it's formulated in
  the context of a fixed set of proposals, effectively forming a
  library collection.  Ideally, this collection would be comprehensive
  in some sense of the word.  Presently, in almost any sense of the
  word, the collection that would result from the current SRFI set is
  not comprehensive.

- In absence of the context of a fixed set of proposals, all
  suggestions for naming schemes I've seen give a clear mnemonic
  preference to some SRFIs over others, usually based on chronology.
  This is likely to be the poorest choice possible in the long run,
  assuming that we accumulate more wisdom with time, and, therefore,
  better proposals.  It's also something I'd like to avoid for
  political reasons alone.

-- 
Cheers =8-} Mike
Friede, Völkerverständigung und überhaupt blabla