[Simh] Way out idea for simh

[Bob Supnik]

On 4/21/2016 3:47 PM, Ken Cornetet wrote:
> In summary, "CU mode" offers a tremendous amount of power for simh users, and requires very little effort to implement.

Very little effort? Every paper tape implementation in SimH is unique. 
The "central library" for paper tape, insofar as there is one, is fgetc 
and fputc. Besides, the proposed CU mode does exactly what the reader 
and punch do today anyway - namely, read or write data from or to a 
single file. Paper tape devices that use non-ASCII character sets 
provide translation to and from ASCII for text.

Magtapes? On input, exactly how are records blocked? On output, how is 
record blocking disentangled? Most older operating systems had specific 
requirements for magtape formats, in terms of record sizing and 
blocking. While card images are a reasonable expectation for text files 
in some environments, it is far from universal as a format (because it 
was very space inefficient, among other things). And what guest OS, 
aside from Unix, will allow output of arbitrary character sequences to a 
raw device from a console command? Assuming there even is an interactive 
console in the simulated environment...

As I have said before, this issue needs to be addressed simulator by 
simulator. For example, look at Dave Pitts' solution for IBM 7094 IBSYS, 
which expected an auxiliary 1401 to deal with the messy details and slow 
processing time of character devices. He wrote utilities that turn text 
files into IBSYS job decks and turn IBSYS output magtapes into text 
files, as well as a generalized command procedure for running jobs.

If the file transfer issue is not addressed to your satisfaction in a 
particular simulated environment, then take it up with the developer or 
maintainer of that environment.

/Bob Supnik