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

The future of the Scheme standard

   Date: Tue, 23 Apr 1996 12:04:54 -0400
   From: "ozan s. yigit" <oz@nexus.yorku.ca>

   Alan [in reference to Implementors]

   > After all, they can implement the things they agree upon, and
   > they can prevent the Authors from ever agreeing to anything different!

   for the longest time, i was under the mistaken impression that "Author"
   meant "Implementor" as well as a contributor to R*RS, and i even thought
   that i was an author [with a little "a" perhaps]. 

Well if "Author" -meant- "Implementor" then several of the people whose names
appear on the cover on R4RS wouldn't qualify as Authors, so that can't be
quite right.  Certainly many of the Authors are Implementors.  Probably
most Implementors are Authors (in fact, I can't think of any who aren't).

The point I was trying to make in the text you quoted above was that if the
Implementors get together and create a de facto standard, they can also use
their positions as Authors to prevent the Authors from ever agreeing to
anything -contrary- to their de facto standard.  This is a simple logical
consequence of the facts that (1) the Implementors are a subset of the
Authors, and (2) the Authors never agree to anything without unanimity.

Note that this does not require any -malice- on the part of the
Implementors.  I'm not saying that they are going to conspire among
themselves: "let's implement this condition system and then prevent the
Authors from ever agreeing to anything else, ok?"  I'm saying that an
implementor, sitting in an Author's meeting, who has just spent two months
hacking one condition system into his Scheme implementation, will not be
inclined to vote for a -different- condition system.

   thanks for making me
   [a mere implementor] so excluded all of a sudden. no joke.

I honestly don't see how you get this out of what I wrote.  I certainly
apologize for any offense you may have taken.  I do agree with you that the
introduction of an explicit distinction between users and implementors is
devisive.  I'm not responsible for introducing that distinction and I've
been arguing against it as best I can.