>Since I got a response, let me elaborate briefly.
>The only place I really experienced handcuffs was
>with the DATABASE concept. I had difficulty with
>things I normally do with custom applications.
>
>For example, I like to have my systems
>automatically create the necessary tables and
>indexes when it is first executed and subsequently
>update structures automatically with each new
>version. I have the code to accomplish this
>already written and all I have to do is include a
>file called SYSSTRUC.DBF and SYSTAGS.DBF in my
>project. This became difficult with the database
>concept.
>
>Other things like validation of data gave me
>difficulties because of problems with key values
>being considered invalid because the Primary Key
>already existed as a DELETED RECORD. Even
>appending a blank record became a challenge.
>
>Don't get me wrong ... There are many nice
>features in the database concept but I just don't
>think they have been fine-tuned enough yet to
>provide enough benefit to warrant the difficulties
>they bring to the table (no pun intended).
>
>Just to balance things a bit, I think the OOP
>concept and design tools are just great. I hate
>to hear any whispers at all about Micro-soft not
>continuing with this product. It

Don

Here is a tidbit I hope you can use. The Primary key is a 'virtual' key that exists only in the DBC -- the key is still a Candidate in actuality. The Primary portion resides in the DBC and is linked via the Table header. This is where the problem comes in. It would seem that is you do not use Primary Indexes but all condidates (or regular) then there is no direct link between the Table header & DBC so (hopefully) most of your problems will go away.

Yes the new data structure is a pain & I spent 2 weeks reindexing, etc before I finally figured it all out.


Good luck,
Tom