Conference ulysse::rdb_vms_competition

	Last week I attended a COGNOS presentation where they spoke of
        STARBASE - their new relational database. They claim that
        it was written by the original authors of Rdb, is DSRI
        compliant, and will only be available on a VAX platform
        until 1990. After 1990, they will have versions for other
        vendors (eg ORACLE). Their other claim was 2 phase commit.
        I questioned them as to whether this was similar to SYBASE's
        effort, and they assured me that the 2PC is all there, you
        don't have to write your own co-ordinator.
        
        They spoke of STARBASE being an opportunity for Digital
        - because it runs under both VMS and ULTRIX. We don't have
        firm release dates for the VMS nor ULTRIX version (for 
        Australia) at the moment, but it would appear to be early 
        calendar 1989.
        
        At the end of the session, they handed out brochures, and
        the following is an extract from it:-
        
        "Cognos Incorporated will market its own relational database
        management system and distributed STAR technology in
        conjunction with its PowerHouse application development
        language. The new RDBMS is an open system, offering support
        of multiple platforms, databases and operating systems,
        and is fully integrated with the PowerHouse fourth-generation
        language. It will be available on the VAX/VMS platform late
        this year.
        
        "This technology will put Cognos in the forefront of meeting
        the distributed requirements of company-wide computing"
        says Michael Potter, Cognos Chairman and Chief Executive
        Officer. "Our objective is to provide our end users and
        application developers with access to data throughout the
        organization, regardless of the differences in hardware
        platform, operating system or database."
        
        The Cognos RDBMS is the company's own version of Interbase
        Software Corporation's Interbase software, a fully relational
        database which supports ANSI SQL. It will offer users of
        PowerHouse on VAX the flexibility to use RMS, Rdb/VMS, the
        Cognos RDBMS or any combination of the three systems.
        Interfaces will also be developed to other widely accepted
        RDBMS's."
        
        
        Julie

Yes, the original author of Rdb/ELN, our incredibly successful
offering....  What BS.

I'm a little confused....
    
    	> They claim that it was written by the original 
    	>authors of Rdb,....
    

        >    The Cognos RDBMS is the company's own version of Interbase
        >Software Corporation's Interbase software, a fully relational...
    
Their own version?  Does this mean that they copied Interbase
    functionality?
    
    Also, where did the article mention Rdb/Eln (referring to -.1)?
    
    
    
    

    StarBase is InterBase. I talked to Cognos. They made some small
    modifications. They didn't have to do much since InterBase uses
    OSRI (Open Standard Relational Interface), which isapparently a
    slightly modified DSRI.
                                                                 
    SMARTSTAR too has had little difficulty running on InterBase.
    
    ---- Michael Booth
    
    InterBase can't be too awful. Ashton-Tate bought an equity share
    in the company. Cognos isn't the only one interested in InterBase.

Interbase=Starbase.   Lots of companies are interested in Interbase.

Interbase runs OSRI, a superset of DSRI (claim)
Interbase runs DSRI programs without modifications (claim)


I do not know if interbase has been able to keep up with the unpublished
DSRI enhancements that Rdb/VMS uses.  My GUESS is that they have been able to.

Knowing Interbase, the company, it would not suprize me at all if the
2 above clains are true.

Starkey (Interbase) likes to claim that he invented Rdb, leaving off
which "Rdb" so that Rdb/VMS is implied.  In reality, he invented
Rdb/ELN which has not been very sucessful.

I believe that everyone is so interested in Interbase because its the
only serious database product out there that isn't a competitor of the
people who are buying it.  ORACLE, RTI, etc. offer competition to A-T's
DBASE and COGNOS' POWERHOUSE, so of course they don't want to get to
much in bed with them.  Also, I assume that Interbase is much more
willing to make a reasonable financial deal than the big guys.

As for DSRI compatibility, I can assure you that Interbase can not have
kept up, since a number of changes haven't been documented even
internally.  However, I expect that Interbase is quite compatible at the
DML/BLR level.  Its in the DDL/MBLR area that Interbase (and Rdb/ELN)
are quite incompatible with Rdb/VMS.  Also, the next version of Rdb/VMS
will have BLR extensions that Interbase knows nothing about.

>As for DSRI compatibility, I can assure you that Interbase can not have
>kept up, since a number of changes haven't been documented even
>internally.  However, I expect that Interbase is quite compatible at the
>DML/BLR level.  Its in the DDL/MBLR area that Interbase (and Rdb/ELN)
>are quite incompatible with Rdb/VMS.  Also, the next version of Rdb/VMS
>will have BLR extensions that Interbase knows nothing about.

       
       I must be missing something here.  I thought DSRI is published
       and available for use by third parties. The purpose, as I
       understand it, is to provide the basis for a richer Rdb
       application environment. 
       
       If so, how can there be undocumented changes to DSRI? Isn't
       there a formal procedure to make changes? Why not publish
       them?                                               
       
       __Julian
       

>       I must be missing something here.  I thought DSRI is published
>       and available for use by third parties. The purpose, as I
>       understand it, is to provide the basis for a richer Rdb
>       application environment. 

>       If so, how can there be undocumented changes to DSRI? Isn't
>       there a formal procedure to make changes? Why not publish
>       them?                                               
       
I'm not sure that there are many (if any) changes to what is published.
In other words, what is in the book should still work.

Just because an architecture is published doesn't mean that it cannot
grow.  Just because an architecture is published doesn't mean that all 
of it is published.

The amount of DSRI that is published was chosen specifically to permit
applications to layer ON TOP OF Rdb/VMS (and other DSRI
implementations).  We have not published sufficient detail to allow someone
else to build a DSRI-compliant run-time system.

There is certainly enough published for someone to build a psudeo-DSRI
compliant system, but there is no assurance that all applications will
be able to run on it.