Conference ulysse::rdb_vms_competition

148.0. "Integration feedback requested" by BANZAI::WASSERMAN (Deb Wasserman, DTN 264-1863) Wed Jun 15 1988 19:57

     _ _ _ _ _ _ _
    | | | | | | | |
    |d|i|g|i|t|a|l|          I n t e r o f f i c e   M e m o r a n d u m
    |_|_|_|_|_|_|_|
 
 
    To: Dist                       Date:   15 June 1988
                                   From:   Deb Wasserman
                                   Dept:   Database Systems Product Management
                                   DTN:    264-1863
                                   Loc:    TTB1-6/D05


    SUBJECT:  VAXset/DATABASE Integration


	As many of you may know, there has been a long-standing problem
	when some VAXset tools are used together with database precompilers.

	In general, the problem occurs when a VAXset product (LSE, SCA, PCA)
	as well as VAX DEBUG, attempts to point back to the user's original
	source file.  There is no correlation between the original source
	and the intermediate source generated by the precompiler.  

	A technical taskforce has been investigating the problem and will
	be producing a paper outlining proposed solutions.

	In addition to the technical research, Product Management is very
	interested in getting marketing and field feedback regarding the
	customers' view of this problem.  Any input which we receive from you
	is welcomed because it helps us understand the problem and its impact
	on our customers.

	Please consider the following questions:

	o  What specific problems cause the most complaints about using the
	   VAXset tools in a pre-processor environment?

	o  What do you think are the most important priorities on the list
	   of requirements (see attached list)?  What is needed to satisfy our
	   user base short term, and what work should be strategically planned
	   in the long term?

	o  What do you think is the impact on the customers' applications and
	   productivity?

	o  What is the size of problem?  What is the revenue impact?  Do you
	   think we're losing possible VAXset (or database) sales due to this
	   problem?


	Please send any and all thoughts you may have on this to me at	
	NOVA::WASSERMAN.  This is being also being posted to the PREPROCESSORS
	LSE, PCA, SCA and SQL notesfiles.



Possible list of requirements
(not prioritized, not complete):

DEBUG
	a)  The source code displayed when the user is stepping through
	    the program must come from either the original source file
	    or the intermediate source.  The user should be able to choose 
	    which he wants to see.
	b)  Add support for preprocessor language variables.
	
LSE

	a)  Error messages from preprocessors must be included in LSE
	    diagnostic files.
	b)  Error messages from the compiler must be correlated with the
	    original source file.  The user must be able to visit errors
	    in either the original source or in the intermediate source.
	c)  Support for templates (&RDB& flag)
	d)  Editing lines containing the &RDB& flag.
	e)  What to do about COMPILE/REVIEW - can't compile the intermediate
	    source.

SCA
	a)  Follow cross-reference information from original source.
	b)  Need support for analysis information on symbols that appear
	    only in the original source.
	c)  Add support for preprocessor language variables.

PCA
	a)  Line numbers in source display must refer to the original
	    source.

Also,
	a)  Preprocessors must pass error messages to the compilers so that
	    they can be included in compiler listings.