I use a naming convention that best suites my programming environment (IDE). When using Visual Studio, with Intelli-sense, I generally will not use any kind of pre or post typing notation as the IDE provides this context with a mouse-over (and you should be able to see its type from its usage in the code). As far as component delineation, as with UI constructs, I favor the noun-verb or noun-noun method of naming; i.e., MyStream_Read, MyStream_Write, MyStream_Flush, etc., or UserName_lbl, UserName_tb, etc. (with or without the underscore, but I do favor caps when not using underscore delimiters). I prefer this method as most IDEs present these entities in sort order. In less automated coding environments (NotePad++) I would use more of the constructs you describe, however, I always maintain sort order naming so its easy to see all labels, functions, or events associated with a given entity or programming construct.

The art of conversation is not only saying the right thing at the right time, but to leave unsaid the wrong thing at a most tempting moment!

SCENARIO: From a C++ ATL program I'm submitting database queries to an MSSQL 2000 system in the following formats: Select * from dbo.table (returns a rowset) Select * from dbo.table for xml auto (returns an XML stream) Exec dbo.sproc ... (could return a rowset or an XML stream) I can't tell before I execute the query whether it will return a rowset or an XML stream, however, after the query is executed I can determine which type of output was returned by checking the rowset's number of columns and first column name. PROBLEM: I now have the output in a rowset. This is fine if the output is a rowset, but, if its an XML stream I have a problem. How can I get the output from the rowset into a stream object AND have it properly translate the XML to a readable format? -- modified at 11:17 Friday 24th August, 2007

The art of conversation is not only saying the right thing at the right time, but to leave unsaid the wrong thing at a most tempting moment!