Error logging approaches

Question

I would like to hear how people have implemented various error
logging approaches in Forte applications.
Forte produces log files automatically for service partitions in
$FORTE_ROOT/log and pops up dialog boxes with uncaught exceptions
for client partitions.
Some of the current problems with the existing Forte error logging
include the famously painfull task of determining which log file
in the $FORTE_ROOT/log pertains to your partition and how to
correctly log client errors.
We have taken the approach of opening our own error file for each
Forte project containing a start class method. All exceptions are
then caught and written to the application error file. This takes
the guess work out of finding the correct error file but raises
some interesting design questions:
1) Should service objects maintain their own error files or should
all problems just be handled as exceptions to be logged by the
calling method. ( In some cases the calling method may not care
that a service object is having a problem. In this case the service
object may try an alternative approach. If the request is finally
filled an error with the service object's functioning has occurred
but not with the calling method's functionality ).
2) Should all error messages be handled by a central message
logger operating as a service object. In this case if various clients
are running on different platforms all error message from an
application are centralized. This saves the effort of checking error
files for the same application on multiple platforms. It also means
that a separate error file does not need to be created for each
client that is instantiated. The message logger service in this
case could be accessed by reference among several applications.
Paul Buchkowski
Merrill Lynch Canada
[email protected]

Answer 1

Never mind...my password to access file server contained a character that was not acceptable. Once changed, access restored.