>>I was also wondering how this callback technique would work for "asynchronous" processing. If I have this scenario:
>>
>>- a Main (client) program instantiates 2 COM servers. Server A does some scanning of files on local drive C:, Server B does some scanning of files on network drive F:. Each may take several minutes to process but are likely to cause low CPU utilization.
>>
>>- in the meantime, the main program may be doing some CPU-intensive activity.
>>
>>What happens when Servers A and B "call back" that they've finished their tasks, but the main program is running something else (worst-case scenario, say a large SELECT - SQL)? Do the callbacks fail? Does COM somehow queue the callbacks to be processed the next time the client is available to do so? TIA.
>
>Check out File #9838. We have used this for "asynchronous" processing. It includes a callback for both errors and completion.

Thanks for the link. It looks promising that it's written by a Microsoftie :-)