TRANSFER Programming Manual

Communication between Different Versions of TRANSFER
National Language and Character Set Support
40970 Tandem Computers Incorporated 7–9
shifted according to the rule for the specific character set type. The national use
characters are a special group of characters in the ASCII character set, as follows:
@ [ \ ] ^ ` { | } ~ # $
The TRANSFER name character set and upshift rule are defined independently of any
SCREEN COBOL character set type. With the TRANSFER delivery system, a character
map can be defined to use ASCII national use character codes to stand for characters in
the name character set. The upshift rule for these codes is fixed by the rule defined in
the name character set for the corresponding characters.
Communication Between
Different Versions of
TRANSFER
A TRANSFER system distributed over a network may include nodes configured for
various versions of the TRANSFER delivery system. Nodes running versions earlier
than B40 have a common predefined base character set. Those running B40 or later
versions, however, require a base character set explicitly configured by the system
administrator. The TRANSFER delivery system requires all of these B40 and later
nodes to use the same base character set.
In a distributed TRANSFER system where nodes are configured with different
versions of this product, communications between version B40 (or later) nodes and
nodes with earlier versions generally present no problems.
The following rules apply:
For applications running on a B40 or later node, references to correspondents and
distribution lists on a B20 or earlier node always succeed. Even if the name
character set on a B40 node does not contain Arabic digits or national use
characters, and one of those characters occurs in a name on a B20 node, the B40
node can still refer to the B20 name successfully. This is true even if the reference
is made by a correspondent who does not defer remote name resolution (as long
as the correspondent's character map includes all characters in the remote name).
For example, a correspondent on a B40 node can issue the READ-NEXT-NAME
UOW to retrieve the names of all correspondents on a B20 node. Furthermore,
that correspondent can use any of those remote names to do such tasks as
addressing packages with the SUBMIT-PACKAGE UOW or placing members on a
local distribution list with the ADD-MEMBER UOW.
For applications running on a B20 or earlier node, most (but not all) references to
objects on a B40 or later node succeed. Specifically, when system administrators
for B40 (or later) nodes configure name character sets that include characters that
are illegal on earlier-version nodes, names containing such characters cannot be
accessed from the earlier-version nodes.
For applications running on a B20 or earlier node, names that contain the
wild-card character (*) do not match names with characters that are illegal on such
a node. For instance, when you issue the GET-NEXT-NAME UOW from a B20
node for names on a B40 node, the TRANSFER delivery system skips over all
those names containing characters that are illegal on the B20 node. Furthermore, if
you supply illegal characters with such names, the TRANSFER delivery system
returns the E-INVALID-NAME error indication.