Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!rutgers!iuvax!maytag!watstat!dmurdoch
From: dmurdoch@watstat.waterloo.edu (Duncan Murdoch)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Stumpers
Message-ID: <1258@maytag.waterloo.edu>
Date: 11 Jan 90 15:23:56 GMT
References: <21990002@hpvcfs1.HP.COM> <25A51804.9795@maccs.dcss.mcmaster.ca> <355@marvin.moncam.co.uk>
Sender: daemon@maytag.waterloo.edu
Reply-To: dmurdoch@watstat.waterloo.edu (Duncan Murdoch)
Organization: U. of Waterloo, Ontario
Lines: 30

In article <355@marvin.moncam.co.uk> emmo@moncam.co.uk (Dave Emmerson) writes:
>Fine, I guess that anybody who hasn't grasped that those are the 'lost
>chains made accessible' by now is never going to. But nobody has explained
>how they could get 'lost' in the first place. 

The easiest way to lose chains is to open a file, write to it, and then
crash the system before you close it.  When you open it you create a size 0
directory entry with no clusters.  When you write to it you use up clusters
on the disk and the FAT is kept up to date, so the chain gets created.  It's
not until you close it that the link is made from the directory to the
chain.  (This is my experience in MSDOS 3.2 and 3.3; I don't know about
other versions.)

A more exotic way to create lots and lots of lost chains is to trash a 
directory or subdirectory.  The most common way I've heard of to do this
is to switch floppy disks during the Abort, Retry, Ignore? message, but
there are probably others, such as hardware errors losing sectors on a
hard disk, etc.  

One last way is to believe people who tell you that all that you need to
do to remove a file is to edit the directory entry using Norton or similar
program.  

>Computers aren't supposed to 'lose' things, or so I was told, this was
>part of their raison d'etre!

Filing cabinets don't lose things either, but I've got stuff in mine
that I can't find :-).

Duncan Murdoch