Subject: Re: binary search?
From: Erik Naggum <erik@naggum.no>
Date: 1999/03/10
Newsgroups: comp.lang.lisp
Message-ID: <3130078479648733@naggum.no>

* Johan Kullstam <kullstam@ne.mediaone.net>
| how does a fill pointer let you *insert* new items to the *middle* of
| a vector?

  you provided us with a sample argument, not the definitive argument that
  I asked Sam Steingold to provide, and I answered you with another sample
  argument to show that your sample argument was not the definitive
  argument I was looking for.  why do you need to defend yourself?

| i was assuming that you would want to maintain the list in a sorted state
| in order to apply a binary sort to it.

  as you can see, your assumptions spawned your answer, and my assumptions
  spawned mine.  since I didn't ask you, but Sam Steingold, I could frankly
  not care _less_ whether you think your assumptions are better than mine.

| i suppose you could fill and then shift everything over with a copy.  or
| have i failed to completely grasp the full power of the fill-pointer?

  I think there are several things you have failed to grasp, but I can't
  say whether the fill-pointer is among them or not.  however, with this:

| this was never meant to be serious.  ...  i don't think sam was
| completeyl serious either.

  you show that you failed _utterly_ to understand my question to Sam.

| it's just a fit of whimsy to see if it were any way possible to justify
| binary searching a list.

  I didn't know you controlled Sam Steingold and know what he thinks, what
  motivates him, and how he will react.  I still doubt that you do.

  the useless opinions and generally silly responses from Johan Kullstam
  having been noted, I would like to ask Sam to answer the question, as I
  happen to believe that one might sometimes want to use a list instead of
  a vector, depending on a host of circumstances, just as I tried to show
  that one might want to use a vector where the standard idiom is to use a
  list.  I can think of several algorithms that usefully might delay
  representational issues, but no applications of such algorithms.  Sam?

#:Erik