delorie.com/archives/browse.cgi   search  
Mail Archives: opendos/2003/09/26/22:11:03

X-Authentication-Warning: delorie.com: mail set sender to opendos-bounces using -f
Date: Sat, 27 Sep 2003 12:10:35 +1000 (EST)
From: DONALD PEDDER <jims_son AT jedi DOT apana DOT org DOT au>
To: opendos AT delorie DOT com
Subject: Re: confirm before over-write
In-Reply-To: <3F7473D7.30780.9F1A1BC@localhost>
Message-ID: <Pine.GSO.4.56.0309271106270.2035@jedi.apana.org.au>
References: <3F71E1FA DOT 24000 DOT A11D25E AT localhost> <3F7473D7 DOT 30780 DOT 9F1A1BC AT localhost>
MIME-Version: 1.0
Reply-To: opendos AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: opendos AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com

> >    I haven't used replace before. I see that you can specify to only write
> > files that don't exist, but then you'd have to use the command again if
> > you DO want to replace some existing files.
>
> Yes, but that's not an unreasonable thing to do.

   I disagree - having to use 2 commands instead of one makes things
unnecessarily complicated and time-consuming. The purpose of an OS it to
make it easy to use the computer (otherwise everyone would have to know
assembler in order to operate their computer), so why then deliberately
make the commands difficult to use? This defeats the purpose of having an
OS.


> >    Copy or xcopy would be good. I have no idea why confirm would be turned
> > off by default - that makes no sense (it's one of the things I don't like
> > about Unix too - you can overwrite/remove something you didn't intend to
> > with a typo).
>
> Because anybody using the commandline is expected to know what they are
> doing.

   I know what I'm doing, but like most humans, my memory is fallible, my
fingers sometimes unco-ordinated. I might use a filename which I've
already used, and thus un-intentionally blast away a file I wasn't done
with, or do the same thing by mis-typing a letter. I'm a human-being, with
human flaws, not a robot.
   The single most important aspect in programming is error-checking,
which is making sure the user did the right thing with their input!
Apparently the DR-DOS programmers don't mind us accidentally trashing our
files.


> Xcopy is a real pain as the command options vary so drastically between
> versions. To give anything but the most basic commands requires knowing
> which OS & version you've got.

   For one person with one computer writing files for their own use (the
scenario that DOS was designed for), this isn't a problem.


> With Copy, confirm is off by default because older versions of MS-DOS
> and PC-DOS didn't *have* a confirm option. So having it enabled by
> default would break a *lot* of batch files and other things.

   So provide an option to turn it on - it can't even be turned on! Most
users don't even HAVE any batch files, never mind lots. Only a handful
actually go as far as doing some programming of their computer.

   The above argument is like saying "we've put a turbo in your car, but
we won't let you use it because older cars might break down trying to keep
up with you". Why then did I get a turbo? More to the point, why did I get
a turbo that not only is off by default, but can't even be switched on?


> Trust me, I've had to deal with install files and the like that are
> *ancient* (because they had to work on things like DOS 2.1 or 3.0). And
> getting them to work with options set to behave differently can take
> *hours*.
>
> So consider what that'd mean to companies that may have hundreds or
> thousands of such files to "fix".

   What about the personal users, for whom the personal computer and DOS
were designed for? Shouldn't they be the number one priority? Most
companies use NT or Unix, not DOS (Novell servers run on DOS, and that's
about it).
   If you're working for a company, you're working for a company - you
get paid for the hours you spend doing stuff. Doing something for YOURSELF
goes unpaid though, not to mention no amount of hours can recover an
overwritten file (only deleted ones can be recovered).


> Thus, the defaults on any "old" commands will be
> such that they are compatible with the oldest
> version of DOS that is likely to still be used.
> Which for some rather common things (especially the
> sort of "embedded" systems that are the main target
> of DR-DOS these days) can be as old as DOS 3.1.

   The problem isn't just that it's not the default, but that it can't
even be switched on! No confirm, no option. At least in Unix you can
switch it on! This is a pretty fatal flaw in an OS which is aimed at
personal users.


> Personally, my advice is to use DR-DOS as the OS, and 4dos as the
> command processor.

   So, I have to use 2 OS's because one couldn't be bothered maintaining a
common option? One which was put there because people DO forget things,
and DO make typo's. Are we all expected to be Automatons now? I'm sorry -
I know more about computers than most people I know (and I work in IT),
but I still have human flaws. I still forget things and make typo's. The
OS should be set-up to deal with that (God knows Windows comes up with a
million "are you sure?" boxes).

dp.

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019