non-native keypunchers

[robin at utafll.uta.edu]

I'm fairly certain that the TLG decision (alluded to) was based
not upon supposition, but upon tests.  It's particularly relevant
in that double-typing method would tend to fail precisely at the point
where knowing what "should be there" in the copy text will generate
independent identical errors, and the diff won't be the wiser.
Otherwise, independent identical mistakes are rare (if they are
"random").

You could check w/ Ted Brunner or others at the project:

tlg at uic.bitnet

PS  I have not followed this thread in detail, but am surprised
not to have heard any concern about validating the structures
of these texts with something like SGML.  What will be used for
"markup"?  The TLG beta code uses its own set of ID markers, but I
feel sure (now) that many wish it were something "standard"
that could be validated by (eg.) a public domain SGML parser.

Best wishes, and congratulations to Dominic and others for taking
the initiative.

Robin Cover

-------------------------------------------------------------------------
Robin Cover                Email: robin at utafll.uta.edu  ("uta-ef-el-el")
6634 Sarah Drive            **In case of link failure, try:
Dallas, TX  75236  USA     Email: Robin.Cover at sil.org
Tel: (1 214) 296-1783      Email: robin at ling.uta.edu
FAX: (1 214) 709-2433      Email: zrcc1001 at vm.cis.smu.edu