Wed, 8 Aug 90 19:23:06 O
|
>I was hampered by the fact that the copy of COUNTRY FILE on LISTSERV's
>A-disk is dated 11/13/88 and contains possible errors (ET for Egypt
>while BITEARN NODES has EG; I haven't checked to see which is correct)
>and omissions (KW was not listed at all; that's Kuwait).
The file used to be manually maintained from the data at the bottom of
NETCNTRY EXEC, which is the "official" source of that information.
Unfortunately, NETCNTRY EXEC is an exec of around 600 lines containing a
lot of comments, a lot of EXEC2 code to provide a NAMEFIND-like function
to the user, and a small amount of country data in the middle (around the
end of the file). The only "safe" way to extract the data is to call the
EXEC, which takes around 0.5 to 1.0 sec of CPU time on a small 370
machine. I never managed to convince Berthold that this should be split
into an EXEC and a data file, so I made my own, and updated it manually
as new countries were added.
The problem is that this is something easily forgotten and particularly
error-prone. I am pretty sure that Egypt was initially entered as ET
because I remember it struck me as something funny at the time, but when
you update the file manually, looking for new countries, it is easy not
to notice that an error in an "old" country code has been corrected.
Anyway in order to avoid this problem in the future I wrote an EXEC that
builds the file from NETCNTRY EXEC (since I have a local NETSERV, I don't
need to worry about which version of NETCNTRY I am calling, by definition
I have the latest one available in my area). I will rebuild the file
whenever NETCNTRY EXEC is updated and include it in the release update
shipments. For now, I have made a 16E-006O fix file containing the
"current" version of COUNTRY FILE.
Eric
|
|
|