Conference ulysse::rdb_vms_competition

Codd & Date Consultiong charge an ENORMOUS amount of money to do the
evaluation.  Since each vendor pays for it, one might question the
validity of the evaluation.

One vendor we talked to, who explicitly put in the effort to meet all of
Codd's requirements (over other product requirements), paid for the
evaluation, was rated as meeting everything, and then marketed the hell
out of it, told us it didn't help them AT ALL.

So, the key question is "Should we pay an enormous amount of money to
Codd so he can make some worthless statement about our conformance to
his model or should we spend the money on some more effective marketing
program?" 

    I thought the fee was around $30,000. That's not exactly enormous.
    
    I have only seen one vendor (oracle) market their claim to adherence
    to the Codd & Date model. I don't believe they ever specified that
    C & D evaluated their product. The only other vendor that comes
    to mind in Cincom, who did, very briefly, publicize that C & D said
    they were more relational than DB2. However, at the time Supra did
    not run on a VAX, and they were competing with DB2, which is vastly
    less expensive than IBM Supra.
    
    If either of those two companies are the ones being used, their
    experience may not be relevant to our market.
    
    Nonetheless, it is interesting that the C & D evaluation did them
    no particular good. The question remains, was that because no one
    cares about "true relationalism", or that the database vendor didn't
    know how to leverage the evaluation (or that competing against DB2
    is useless).
    
    ---- Michael Booth

    Re: .0
    
    >First has anybody got an electronic copy of his rules ?
    >
    >And a comparison of how Rdb/VMS matches ?
    
    Look in COOKIE::DSRI.  There's an essay describing Codd's Rules
    and our evaluation of DSRI's compliance to those rules.

    Bear in mind that the DSRI 'essay' is dated now.

>    Bear in mind that the DSRI 'essay' is dated now.

It must be really dated by now.  Has anyone updated this to reflect our current
Rdb/VMS product?

>>>>    Bear in mind that the DSRI 'essay' is dated now.

>>It must be really dated by now.  Has anyone updated this to reflect our
>>current Rdb/VMS product?

Well, you can always go look at it and see how dated it is...

I am the author of the analysis of DSRI and Rdb/VMS vs Codd's 12 rules.  I
have updated it a couple of times, including once about a year or so ago
(9 March, 1989), so it's not too vastly out of date.

I do *not* plan to update it any more.  It's reasonably time-consuming and
I doubt that the overall effort in terms of sales or even education of
sales force for incremental updates is worth it.  If someone on the
Rdb/VMS development team wants to update it, that's fine with me, though.

Enjoy,
   Jim

    I think that an account of how Rdb or DSRI compares to these 12 rules
    would be a lot less useful now then when it was first done.
    
    Firstly, Codd has just published "The Relational Model Version 2"
    (Addison-Wesley, don't have the ISBN) which is a whole bookful of rules
    about the relational model, so those who are interested in how we
    compare to his rules will expect us to measure our products against a
    lot more than 12 rules.
    
    But I think that these rules are beginning to outlive their usefulness.
    Maybe they helped when people weren't really aware of what relational
    meant and why it was good to be relational. But I think that people
    now know what it means and can form their own judgements about how
    useful the relational model is, and how important particular features
    of it are to their applications.
    
    	Andrew.