Re: filenaming proposal for IAFA description files

[Matti Aarnio <[e-mail]>]

> > Now, this filename will obviously have to be different for ...
> 
> Well, after reading over the new iafa-doc2, I have yet another suggestion
> to make.  Rather than limiting the names of the IAFA- files to a particular
> name/extension/case, I suggest you have a single file that meets all the
> restrictions of known operating system names.  This single file will give
> the mapping from IAFA- document name to system dependent name.  Software
> that is doing mirroring could then grab the one known name file, parse
> its contents and use that to obtain the other IAFA- documents.  This may
> [I hope] help in the automation of IAFA- data.
> 
> Implementable?

  From practicality viewpoint:  Yes, very much so!
That first file should be named something which is implementable at
all systems (6+3 ? 8+3 ?) and then all the rest is just a question of
simple mapping.  However those maps should contain BOTH directory info
and real filenames - for beasts like IBM CMS with a bit odd "directory"
structure...

Model:
	IAFA-foobar: "string for CWD argument" "target.file"
(And as it is all text anyway - ASCII or EBCDIC - that need not be defined ?)

> Aydin Edguer

	/Matti Aarnio <[e-mail]>