Tapes and close to a quarter-century.

[Jon maddog Hall]

bmcculley at rcn.com said:
> Usually but not always.

"Usually" usually means "not always".  But you brought up a good point.  Logical
"records" often bridged "physical" records....and in the case of the start-stop tapes
the usual real "physical" record was the block.

Along the same lines, some tapes contained "source code" and the systems wrote them
as 80-byte records (we will not go into the issues of six-bit bytes, eight-bit bytes,
nibbles, etc.) with each line padded to the full 80-character "card image".  So you
often had 800 or 8000 or 80000 byte blocks, with the record length being 80 bytes.

I remember how odd it was to me that Unix simply put a "new-line" character at the
end of every line, and did not have to pad the record with blanks.  In days of
five megabyte disk drives (no screaming about how big that is, o.k.?) this was
a big savings.

bmcculley at rcn.com said:
> If the next block was supplied in time the tape kept moving and the data was
> written following a gap of some indeterminate length.  If more data was not
> supplied in time, a horrendously expensive stop-reposition-start sequence was
> initiated.  Thus the time-out interval was usually not very tight, and gaps
> were loose.

I was under the impression that if actual data was not written in time that the
streaming tape would write "null data" (not the same as the traditional "inter-record
gap", until it either received more data to write or it gave up and stopped.  When
it got more data it started your "reposition-start" again.  When it did reposition,
it repositioned before the end of the "good data" and then started writing.

We overcame this in Ultrix by utilizing a ring buffer which managed to deliver
the data fast enough to the tape drive to keep it "streaming".

The inter-record (really inter-block) gap on a start-stop, 9-track IBM tape was .75
inch.

- Jon "maddog" Hall
(ex-TK50 junkie)
-- 
Jon "maddog" Hall
Executive Director           Linux International(R)
email: maddog at li.org         80 Amherst St. 
Voice: +1.603.672.4557       Amherst, N.H. 03031-3032 U.S.A.
WWW: http://www.li.org

Board Member: Uniforum Association
Board Member Emeritus: USENIX Association (2000-2006)

(R)Linux is a registered trademark of Linus Torvalds in several countries.
(R)Linux International is a registered trademark in the USA used pursuant
   to a license from Linux Mark Institute, authorized licensor of Linus
   Torvalds, owner of the Linux trademark on a worldwide basis
(R)UNIX is a registered trademark of The Open Group in the USA and other
   countries.