home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.unix.wizards:4636 comp.unix.shell:4661 comp.unix.misc:4127
- Newsgroups: comp.unix.wizards,comp.unix.shell,comp.unix.misc
- Path: sparky!uunet!caen!hellgate.utah.edu!eeide
- From: eeide%asylum.cs.utah.edu@cs.utah.edu (Eric Eide)
- Subject: Re: The Problem with UNIX
- Message-ID: <EEIDE.92Nov12154722@asylum.cs.utah.edu>
- In-reply-to: michael@chpc.utexas.edu's message of Thu, 12 Nov 92 19:37:07 GMT
- Organization: University of Utah Department of Computer Science
- References: <aldavi01.721333614@starbase.spd.louisville.edu>
- <1992Nov11.194557.16258@yarc.uucp>
- <EEIDE.92Nov12120339@asylum.cs.utah.edu>
- <1992Nov12.193707.27532@chpc.utexas.edu>
- Date: 12 Nov 92 15:47:22
- Lines: 47
-
- Michael Lemke (michael@chpc.utexas.edu) writes:
-
- Michael> Well, fixing typos is neat but it is not the essential problem. My
- Michael> main complaint about Unix on the user interface level is that there is
- Michael> no command line interpreter. What I mean is that after the shell
- Michael> munged your command line it is *completely* up to the program to
- Michael> interpret the command line and there is no system function available
- Michael> to parse even these `standard' options.
-
- It sounds as if you are proposing the invention of an application-independent
- command line parser that does more than just syntactic processing (as provided
- by getopt()). This would definitely be interesting -- but is it really
- possible? What would it look like? What would the universal meaning of `-l'?
-
- If there were such a command line parser, it would certainly reduce the amount
- of knowledge that my smart shell requires. As it is now, I basically have to
- describe the syntax of every program that I want to be "smart" about. The
- shell is an unusual program in the sense that most of its "command set" is
- implemented by entities that have no real connection to the shell itself. And
- obviously, it's hard to be smart about things you know nothing about :-). For
- this reason I've sometimes thought that I made a poor choice in deciding to
- make a "smart shell" for my thesis. But on the other hand, the shell is the
- most commonly used program with a command line interface, so perhaps a "smart
- csh" is much more valuable that a "smart ftp."
-
- Michael> Some programs use one letter chinese (you know, one character per
- Michael> word) and others (eg, find) use words (-print -name). And the real
- Michael> problem then starts when -l changes its meaning from command to
- Michael> command, some commands need spaces between the option and the
- Michael> argument, others don't, some take both, yech. This would all be
- Michael> solved if there were *one* system function that is used by all
- Michael> programs instead of having every program duplicate more or less the
- Michael> same functionality with different success.
-
- If I understand you correctly, you're basically making a call for standard
- syntax. I'm all for it. Now as for the standard semantics (e.g., the
- universal interpretation of `-l'), that's definitely up in the air.
-
- Michael> Does AI work now?
-
- You're begging the question of whether or not AI ever worked. For a suitable
- definition of AI, AI has always worked :-).
-
- --
- -------------------------------------------------------------------------------
- Eric Eide | University of Utah Department of Computer Science
- eeide@cs.utah.edu | Buddhist to hot dog vendor: "Make me one with everything."
-