Subject: Re: conservative gc sucks
From: Erik Naggum <erik@naggum.no>
Date: 14 Jan 2003 15:38:50 +0000
Newsgroups: comp.lang.lisp
Message-ID: <3251547530539377@naggum.no>

  In the unlikely event anyone else reads Tom Lord's increasingly
  spurious excuses for articles...

* Andy Freeman
| Can this reliably work on Unix?

* Erik Naggum
| If you delete an open file, what exactly do you want to rely on?

* Tom Lord
| That the data in the deleted file is still accessible to programs
| holding a descriptor for that file.  That the file name may be
| re-mapped to a new inode, even while the previous inode is still in
| use.

  Are you speaking for Andy Freeman?  Did you understand the /context/
  of the question he asked?  Or are you just jumping into a discussion
  to which you have not been paying much attention so you can say
  something stupid and pretend it is somebody else's fault?

* Erik Naggum
| Why would you want to enroll such an open file in this
| mechanism?

  Recall that the mechanism discussed here is one of keeping track of
  a number of /files/ by name, which can be opened and closed at will.
  If the files cannot be opened and closed at will in this fashion,
  then use your goddamn brain and do not use this technique!  How hard
  can it be to grasp that just because a technique has limitations
  does not mean it is not useful where those limitations do not apply?
  What kind of a programmer are you if you are unable to grasp this?

* Tom Lord
| Because these semantics are part of how good unix programs achieve
| ACID semantics.

  Let me get this straight.  You want automatic flushing of buffers
  and closing of files upon garbage collection of streams so you can
  implement certain algorithms which you claim are not easy in Common
  Lisp as is (except that it is, you just use finalization).  When a
  simple scheme is proposed that would free you of all concerns for
  when files are opened and closed (which seemed to be the point of
  your desires until you came up with more problems after the old ones
  were solved) for some particular uses (I proposed only two such
  schemes depending on usage factors, obviously not the exclusive list
  of methods, but the /obvious/ choices for how to implement what you
  claimed was /obvious/ yet did not understand), your counter-argument
  is that it should not be used because you want /ACID semantics/ of
  these garbage collected and automatically opened and closed streams?

  Are you /insane/?  Or are you just making up problems as you go so
  that nobody can ever satisfy your increasingly irrational demands?

| It may be the case that sometime in the future, we'll be able to
| wrap unix filesystem calls in begin-/end-transaction wrappers -- but
| for a few decades now, people have made good use the very easily
| implemented semantics under discussion to achieve ACID properties
| without full-fledged transactions.

  Are you at all aware of what thread this is in?  Does the word
  "context" not have any meaning to you at all?  Is thinking so hard
  for you that you have to concentrate real hard and work long hours
  not to make bumbling mistakes?  Is that why your miserable excuse
  for a brain overloaded when you were given standard sociological
  arguments about models and you called them "jaw-droppingly dumb"?
  I take some solace in the fact that the arguments were given by a
  professor of sociology at the U of Oslo who has published dozens of
  papers and books.  When you, an obviously harebrained fool, pays no
  attention and provides evidence that you are so mind-numbingly
  retarded that you should be protected from yourself, that is one of
  the things he has discussed over the years -- how people who are
  poor on mental models seek out and need information that conforms to
  their models lest they become very nervous and agitated and feel an
  urge to escape rather than to think.

| Argument for argument's sake?

  Yes, that does indeed seem to be your modus operandi.

| If the earlier questions were honest,
| I've provided the answers to them.

  Except for the fact that they were asked of Andy Freeman, not of
  you.  Except that you do not appear to be particularly honest to
  begin with.  Except that you answered questions completely out of
  context.  Andy Freeman had already swayed from the context, but you
  showed me what kind of idiot you are when you decided to respond.
  You probably think that just because your feeble brainpower cannot
  deal with something, it must be dumb.  Astonishingly stupid people
  tend to behave that way.

  It must be fun to be you.  I sometimes wish I had accepted the drugs
  I have been offered at various parties of the years so I could have
  been able to /relate/ to people like you and how your brains work.

-- 
Erik Naggum, Oslo, Norway

Act from reason, and failure makes you rethink and study harder.
Act from faith, and failure makes you blame someone and push harder.