POV-Ray : Newsgroups : povray.off-topic : Prehistoric dust : Re: Dusty Server Time
4 Sep 2024 21:18:31 EDT (-0400)
  Re: Dusty  
From: Clarence1898
Date: 18 May 2010 20:25:00
Message: <web.4bf32fc3ecb621eff0b197720@news.povray.org>
Orchid XP v8 <voi### [at] devnull> wrote:
> >> Files only on disk? Or on tape too?
> >
> > Obviously tape files were contiguous.  I must be misunderstanding what
> > you're asking.
>
>  From what little I've seen, with punch cards it's a case of "please
> read this bunch of cards". You don't have filenames or anything.
> Presumably on magnetic disk you have a file *system* which describes
> logical files with names and things. I'm asking whether the same holds
> for tape, or whether it's just treated as an endless stream of bytes (or
> records or whatever).

If the mag tape has standard labels, the file attributes are defined in the tape
label header record. The tape label also contains the last 18 characters of the
dataset name. If its an unlabeled tape, it has no tape label headers, so the
program has to have the attributes hard coded in the source code.  98% of all
the tape we process are created internally.  We have a tape management system
that keeps track of what dataset is on what tape volume.  If the operator mounts
the wrong tape, it will be rejected and the operator will be re-prompted to
mount the right tape.  If its an external tape we just tell the TMS to bypass
validity checking.

>
> >> OK. But does the system know where the *fields* in a record are? Or
> >> just what size the records are?
> >
> > You compiled it into the program. Often when they weren't actually fixed
> > size, they were fixed size anyway and padded (like cards).  Or the size
> > was stored in the header of the file.
>
> Right. So it's a property of the program, not the system.

Yes, typically the record structure is defined in an include file, much like C.
Thus a common record format is available to all programs that access that file.

>
> >> Yeah, I think the term "mainframe" is probably obsolete now. There are
> >> probably more exact ways to describe what type of computer you mean.
> >
> > Could be.  There's something clearly between "small" and "large" now.
>
> Minicomputers! :-D

There can be so much overlap in performance of PCs, Minis, and mainframes, it
can still be difficult to classify where a particular machine goes.

>
> --
> http://blog.orphi.me.uk/
> http://www.zazzle.com/MathematicalOrchid*


Isaac


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.