Subject: Re: Harlequin vs. Allegro
From: Erik Naggum <clerik@naggum.no>
Date: 1998/06/29
Newsgroups: comp.lang.lisp
Message-ID: <3108096841138643@naggum.no>


* Scott L. Burson
| And it is my understanding that their Windows products, also called
| "Allegro", were, at least originally, of yet another lineage; I'm not
| sure about the situation today.

  the latest Windows version, ACL 3.0.2, was a completely different product
  from the Unix version, ACL 4.3.1, much to my dismay.  ACL 5.0 is the same
  product on both, with Windows-specific stuff coming from the ACL 3.0.2
  heritage, like the IDE.  the very real and noticeable shortcomings of the
  Windows offering that I ran forcefully into when one of my first clients
  insisted on NT (which I didn't know how crappy was at the time) cost me a
  tremendous amount of work.  I have recommended strongly against ACL for
  Windows since then, and I have always been precise that I'm talking about
  the Unix version when praising Allegro Common Lisp, but the consolidated
  version should be as good as the Unix version is.  I'm very happy that
  this particular confusion is now history.

| It may be true, or more nearly true, now, but I don't trust Franz in this
| area, and if I wanted to develop a Windows application I would use their
| (or someone's) Windows product and not try to cross-develop from Unix.

  yup, been there, tried that.  cross-development wasn't even close to
  doable between ACL 4.3 (an ANSI CL) and ACL 3.0.2 (barely CLtL1 with some
  CLOS), so I have also yet to fully regain the necessary confidence, but
  watching the bug fixes and other stuff about ACL 5.0 strongly reinforces
  their claims to common sources.  maybe we can finally get to the point
  where I wanted to be almost two years ago when this client wanted NT:
  that one can run Allegro CL and Emacs under Neanderthal Technology and
  not have to deal with that ill-designed "operating system" at all.

  to be honest, part of the reason I like and use Common Lisp is that I
  don't have deal with all the demented Unix heritage, either -- it's a lot
  of work, but possible, to write clean and neat abstractions on top of the
  many Unices and not have to deal with their low-level crud and mistakes.

  speaking of mistakes, I _want_ file versioning.  :if-exists :clobber is
  _not_ my favorite option, and I want (close ... :abort t) to leave the
  old file intact.  that Unix does not offer a way to do this so it's quite
  hard to get it right has been a pet peeve of mine for years, long before
  I shifted focus to Common Lisp.  safe file I/O generally appears to be up
  to the application.  e.g., I don't want to make a file available to the
  world until it has been fully written and closed, but that means I have
  to use temporary filenames and do a rename on close manually under Unix.
  this is annoying.  what's more annoying is that I haven't found a way to
  hook into the operation of OPEN and CLOSE to automate this behavior.

#:Erik
-- 
  http://www.naggum.no/spam.html is about my spam protection scheme and how
  to guarantee that you reach me.  in brief: if you reply to a news article
  of mine, be sure to include an In-Reply-To or References header with the
  message-ID of that message in it.  otherwise, you need to read that page.