Orchid XP v8 wrote:
> Obviously there's also a truckload of ways AmigaOS is *different* from 
> Unix. [There are no device files, pathnames have a syntax more like 
> MS-DOS, configuration is always stored in binary files not text files...]

Ah. OK.  I was looking at AmigaOS as the software and how you interface 
to the operating system.  I wasn't looking at AmigaOS as the 
command-line commands that come with AmigaOS.

I mean, really, CP/M had "more" and "type" and such.

> Now I would suspect that would tend to break horribly as soon as some 
> new application is added that expects everything to be in the normal 
> locations...

That would be where the symlinks come into it. :-)

> [Basically I have absolutely no clue how the traditional Unix file 
> layout is supposed to work. I don't know why, for example, we have /bin, 
> /root/bin, /usr/bin, /usr/sbin, /usr/share/bin...]

Legacy, and the fact that way too many programs hard-coded file paths 
instead of the OS doing something like providing APIs for finding out 
such stuff or having some efficient central storage like a registry. Of 
course, the #! syntax *requires* hard-coded paths in every file, so it's 
kind of difficult to improve easily in that sense.

You started out with two disk drives: the root, and the users' 
directories. But they were small, so you wound up putting things like 
the mail spool on the user disk (hence, /usr/spool) and things like 
executables you don't need before the system is booted enough to mount a 
second disk someplace like /usr/bin. And there were executables that 
only the administator could run anyway, so to cut down the number of 
directories everyone else had to search, you put those in /sbin and 
/usr/sbin, depending on whether they were the kind that needed the 
drives mounted or not. (Check a file system: better be on the root. 
Delete a user: Didn't need to be on the root.)

Then you had network-shared file systems, so you needed /usr/share as a 
place to mount that, and /var as a place that you could refer to with a 
common name but was actually different storage on each machine.

Then /usr got so overused for everything that it didn't make sense to 
actually put, you know, users there. So you got /home or any of a 
half-dozen variations on that.

Nowadays, it looks like this, mostly:
http://en.wikipedia.org/wiki/Filesystem_hierarchy_standard

> They had trouble. Somebody walked past and the T-adaptor on their PC 
> fell apart, breaking the ring. 

They set up the ring wrong, then. :-)  Each loop is supposed to go back 
to the hub that detects a broken connection and bypasses it.

-- 
   Darren New / San Diego, CA, USA (PST)
     "That's pretty. Where's that?"
          "It's the Age of Channelwood."
     "We should go there on vacation some time."

Post a reply to this message