>Fabio-
>
>>Any test is useless, because 1ns after the success test result,
>>the connection can be died.
>
>While your point is a good one, I disagree that it's "useless". That said, trapping the error in a Try/Catch block would be a great idea.

Nancy, ?

i written:
Send the data and trap the errors.
How the developer trapped the error is another issue.

"Useless" it is relative to an eventual test estimate of valid static connection.
In other words:

IF ConnectionIsGood(h) && this is useless
 TRY
   SendCommand(h,@cmd)
 CATCH
   ...
 ENDTRY
ENDIF

Fabio