[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ulysse::rdb_vms_competition

Title:DEC Rdb against the World
Moderator:HERON::GODFRIND
Created:Fri Jun 12 1987
Last Modified:Thu Feb 23 1995
Last Successful Update:Fri Jun 06 1997
Number of topics:1348
Total number of notes:5438

586.0. "ORACLE problem with VAX C?" by WR1FOR::KARIMDA () Thu Mar 01 1990 19:50

    I don't think this is the right conference to put this notes. However,
    since I don't know the where the ORACLE conference is or if there
    is one at all? I thought this would be my best bet. 
    
    Anyway, one of my customers who I am told is in the final decesion stages
    of buying into ORACLE (for some undisclosed reasons) asked that
    they had heard that the current version of ORACLE has a major problem
    with VAX C under VMS. They think it is because ORACLE uses some names
    that are common to the VAX C Runtime Library and therefore causing
    the problem. They want to know if there is some truth to this, since
    it would definitely impact their final decesion. 
    
    My knowledge of ORACLE is very limited. Any response would be of
    great help. Since I am trying to sway the decesion towards Rdb/VMS.
    Thanks in advance.
    
    
    ....dabir
    
    P.S. Also could some one point me to the ORACLE notes conference, if
         there is one at all.  
T.RTitleUserPersonal
Name
DateLines
586.1C problem? Maybe.SRFSUP::LANGSTONRdb is *the* Relational db for the VAXFri Mar 02 1990 03:1662
    I think I remember hearing something about Oracle having a lot of
    names (some of which probably clash with VMS's reserved words).
    
    But there are many other things about Oracle which can motivate
    you and your customer and the sales representative to learn more
    about the virtues of VAX Rdb/VMS and the many reasons beneficial
    to all for choosing the best relational database on the VAX for
    the VAX.
    
    Do the above-mentionied folks a favor and look through this conference 
    a little. Try dir/tit="oracle".  You'll find a lot of good stuff to 
    help you help your customer.
    
    Just a few things, some of which you've probably heard before: 
    
    For your customer:
    
    	Oracle is made to run on myriad hardware platforms and, due to
    the "least common denominator factor," can't fully take advantage
    of any one hardware/software architecture.
    
    	Rdb has a pricing structure which makes it easy for your customer,
    if he's an MIS Director, for example, to fit it into his budget.
    He can, for example, buy an Rdb Full Development Option license for a 
    MicroVAX, on which he can develop his applications, and get a Run
    Time Only license for his (probably larger) production VAXes.
    Oracle can be very expensive.  The five year cost-of-ownership
    differences between an Rdb (and tools) application and one developed
    with Oracle can be large indeed.  Encourage your customer to go
    through the exercise of determining what the costs might be.
    
    	Oracle's latest version doesn't take advantage of the high
    availability provided by VAXclusters, or the SMP architecture of
    your and my favorite hardware platform.
    
    	Oracle is not integrated with the VMS operating system and the
    many other Digtal products and tools with which your customer might
    be familiar.

    For your sales team:
    
    Much of Oracle's revenue comes from consulting, which your team
    will correspondingly loose.
    
    An Oracle-on-Sequent solution is less expensive than an Oracle-on-
    VAX solution.  The Oracle rep can't wait to tell your customer all
    about it.

    
    Also, look in nova::pm01:[booth.public] for some excellent up-to-date
    information that will help you.
    
�   Anyway, one of my customers who I am told is in the final decesion stages
�   of buying into ORACLE (for some undisclosed reasons)
    
    Put on your best consultative face and ask your customer to share
    with you what his reasons are.  The answers may be enlightening
    and help you enlighten him and get the sale.
    
    Keep us posted here, and let us know how you're doing.
    
    Bruce
586.2LSE support?DPDMAI::DAVISGBEscapee from New Hampshire...Sat Mar 03 1990 02:258
    and ask if they plan to use the Language Sensitive editor to assist in
    the development of the 'C' code.  If so, they can use it with VAX SQL
    also....if it's Oracle's SQL, forget it....
    
    Good luck,
    
    Gil
    
586.3Runtime library MAIL::DUNCANGGerry Duncan @KCO - DTN 452-3445Sat Mar 03 1990 03:4426
    During our V6 benchmark in Sept 89, we encountered quite a few problems
    with the install and mandatory link of Oracle.  On their distribution
    tape, Oracle supplies a copy of the VAX C runtime library
    (ORACTRL.EXE).  Per the Oracle VMS install guide:
    
    "... this file [ORACRTL.EXE] contains a copy of the VMS C runtime
    library distributed by Oracle Corporation.  This copy is slightly
    different from the DEC version of this library, VAXCRTL.EXE, that
    already resides on your system in SYS$LIBRARY.  If used, ORACRTL.EXE is
    linked as a shared image to ORACLE.EXE when you install the Oracle
    RDBMS.  You may use Oracle's copy of this library, or you may link
    ORACLE.EXE to VAXCRTL.EXE, but you must link to one of these runtime
    libraryies to produce the executable Oracle code."
    
    When we tried to link to the VAXC library we got error on routines
    DOPRINT and MALLOC.  It seems that VMS 5.2 changed the names of these
    routines to something else so Oracle wouldn't link.  I can't remember
    the exact sequence of events but it seems like we extracted these two
    routines from the ORACRTL.EXE and inserted them into the VAXCRTL.EXE
    ...or... we may have link to the ORACRTL.EXE and not attempted to link
    to the VAXCRTL version at all.  
    
    In either event, it seems to work OK but left me with an uneasy feeling
    about that versions of the modules in the ORACRTL library.
    
    -- gerry
586.4Precompiler Problems Too !!!MAIL::DUNCANGGerry Duncan @KCO - DTN 452-3445Sat Mar 03 1990 03:5657
    As a part of a system sale at my Oracle customer, we provided software
    services to convert the HP Unix C code to VAX C and decided to use
    Oracle's C precompiler.  (All other code at this customer site is Cobol
    and makes direct calls to Oracle or passes dynamic SQL.)  This
    particular SWS person is one of our very best people and an excellent
    C developer.  In short, he ran into some significant problems with the
    Oracle C precompiler and his report to me follows.  
    
    Since I don't know C, hopefully it will have more meaning to the
    masses.
    
    -- gerry
    
    
    	[... begin of email ...]
    
	Subject: Oracle Pro*C problem descriptions you asked for



	The problems that we encountered at xxxxxx  were not major yet I had
	never run into them when using a precompiler for Rdb.

	The Oracle Pro*C precompiler does not appear to precompile
	C's #define statements.  This makes it rather difficult to define
	a constant which specifies the current size of a relation element
	(a table column in Oracle terms).  For example --

	    EXEC BEGIN DECLARE SECTION
		VARCHAR Test_identifier[K_TEST_ID_SIZE];
		VARCHAR Result[K_WELL_MAXIMUM][K_RESULT_SIZE];
		.
		.
		.
	    EXEC END DECLARE SECTION

	    Where K_TEST_ID_SIZE might be #defined to say 10.

	    In the code for xxxxxx we were forced to use a literal values 
	    ("hard code") as in the following example.

	    EXEC BEGIN DECLARE SECTION
		VARCHAR Test_identifier[10];
		VARCHAR Result[96][10];
		.
		.
		.
	    EXEC END DECLARE SECTION

	If Oracle provided a "CDD" for source code inclusion of table
	structures into a routine, the above situation would not have
	been as much of a problem.

	[... end of email...]