<  The TU58 was never intended as a backup or production medium.  It
<was designed to replace the TU55/TU56 as a small file-structured
<random access device.  That's right, I said random access.  "But wait,

And also the tu60 cassette.

It was the main storage for the PDT11/130 (LSI-11 with 32kw ram and up to
six serial ports).

<tapes are sequentially accessed..."  Yes.  But TU58 tapes (and TU55/56
<tapes) have blocks with block numbers which can be accessed randomly
<like a disk.
<
<  Anyway, that's the strength of those tapes that they wanted to
<protect...not the actual encoding format.

The format was not all that secret.  Also at least two compnaies offered 
tu58 equivelent that was floppy based.

There is that and also at that time (1979) the TU58 was recording at 
approximately 800bits/inch which for that tape and time was decent.
The tapes were small, rugged and designed for repetitive access of
the same block (shoe shine). It was a the cheapest, simplest block 
addressable device for it's time.  It was also unique in that it put 
the least interface requirements on the host system, a serial port at 
any baud rate from 150b->38.4kb.  It also has a mode where it will drive 
the lsi-11 console ODT direct to self boot a romless system.

Also on the density thing.  The TU58 used a two track fixed head, all 
the streaming drives use a multi track format of far more than two tracks.  
If you went for the track density(at 800bpi) used on the travan drives the 
TU58 would easily be a couple megabytes though not near as reliable or a 
simple.

I may add that Bob Armstrong has designed a 8751/ram based semiconductor 
tu58.  The advanatage is that the load time is limited by baud rate rather 
than the rewind/fast forward speed of the tape.  If enough people are 
interested the PCB for it can be produced for resonable price.

I've committed to assembling one using wire wrap as I'm cash limited 
and the design is straightforward.


Allison