Very Computer

Is Faircom's C-Tree in some way object oriented?  I ask
this because our company is evaluating an "Object Oriented
Human Resource System" called "ICONtrol", which uses C-Tree
as its underlying database.  I'm trying to figure out what
aspect of the product is OO; its interface? its internals?
its underlying repository?

Call me a skeptic, but I've seen products claim to
be OO for every reason from "Well, we talk OLE, don't we!",
to "But isn't everthing really just an object...".
Curious minds want to know...

Jim Long

-- Michael
--
------------------------------------------------------------------------------


------------------------------------------------------------------------------

Quote:>C-Tree is an implementation of ISAM.  That stands for Indexed Sequential
>something or other.  It's a way of storing bytes sequentially in one file
>while maintaining an index of a record's position in the sequential file
>in another file.  These are referred to as the Data and Index files.  Each

--
-------------------------------------------------------------
G?ran Hasse                              tel:  + 46 8 7730148
Raditex AB                               fax:  + 46 8 7730902

C-tree, as a lower-level database engine than an RDBMS is a good basis
for developing an OO system. I wrote a multi-user classified
advertising system a couple of years ago, using this approach.

I'm now working on a "middleware" product which makes it easier to
make OO use of c-tree and other record-oriented databases, linking
specifically to diferent gui class libraries. The current project
is a c-tree backend and zApp frontend. I have an ambitious beta
tester who's going to try Symantec's Mac lib (TCL) with dtF as a backend.
However, SQL backends present an entirely different set of problems,
so I'm making no guarantess about that project.

OOFILE allows automatic object part-explosions, BLOB storage,
keyword and phonetic indexing, lookup tables and multi-value fields. The
main aim of the project is to get queries as readable as high-level
4GLs, but integrated into c++ frameworks for performance and total
extensibility. (The current project is a port of a complex 4th Dimension
database, from the Mac.)

I also know a few other people have used c-tree as a backend for OO
systems. Its fast variable length records and flexible indexing are ideal.

Andy Dent (A.D. Software - Mac, DOS & Windows GUI dev.)
94 Bermuda Dve, BALLAJURA  Western Australia  6066
Phone/Fax: 09 249 2719 (local)  +619 249 2719 (International)

TIA

--
Lee McAllister

: C-tree, as a lower-level database engine than an RDBMS is a good basis
: for developing an OO system. I wrote a multi-user classified
: advertising system a couple of years ago, using this approach.

: I'm now working on a "middleware" product which makes it easier to
: make OO use of c-tree and other record-oriented databases, linking
: specifically to diferent gui class libraries. The current project
: is a c-tree backend and zApp frontend. I have an ambitious beta
: tester who's going to try Symantec's Mac lib (TCL) with dtF as a backend.
: However, SQL backends present an entirely different set of problems,
: so I'm making no guarantess about that project.

: OOFILE allows automatic object part-explosions, BLOB storage,
: keyword and phonetic indexing, lookup tables and multi-value fields. The
: main aim of the project is to get queries as readable as high-level
: 4GLs, but integrated into c++ frameworks for performance and total
: extensibility. (The current project is a port of a complex 4th Dimension
: database, from the Mac.)

Hi Andy.
  May I ask you several questions?
  1) what is "BLOB"?
  2)In "Object-Oriented Databse Management"(ISBN 0-13-104092-8), Kemper wrote
    :we describe an indexing scheme,called "access support relation"(ASR) that
    goes beyond the common indexing as konwn from relational system.

    He used B-tree to retrieve data as you describe "C-tree" to load object
    from disk.Are they the same?
  3)Once I design a storage model ,called "FileCluster".like this:

     OID_array:contained every object status.
        |
     FileCluster:contained "Cluster Number"(CLN) which is derived from
        |        hash(OID).It is similar to C-tree way.
     Placement Tree:deapth first or beradth first.
        |
     Physical File organization

    This is my solution of OOFILE.

 Edward Shiang

In my first OO project with c-tree, I was storing advertising WP BLOBs
of around 2-5kb. I was very happy with the performance (on Classics).

:   2)In "Object-Oriented Databse Management"(ISBN 0-13-104092-8), Kemper wrote
:     :we describe an indexing scheme,called "access support relation"(ASR) that
:     goes beyond the common indexing as konwn from relational system.
Sorry, I don't know the reference so can't comment. The relations I'm
building into OOFILE include automatic loading of parts, multi-value fields
and one-one, one-many, many-many relations. I'm not implementing any of the
indexing strategies other than using c-tree (or other) indexes as OOFILE
is a layered product designed to sit on top of several other database
products (and under a lot of pressure to go SQL).

The main strengths of OOFILE are that it gives you integration with gui's as
well as high level database modelling eg:
likelyVoters = People.AlwaysVote=="Labor" || People.StayBought.Price<2000;
OR
likelyVoters = People.AlwaysVote=="Labor" || (People.StayBought &&
                                              People.Price<2000);

With the gui integration (I'm working on the zApp version) I'm dealing with
issues like multiple page entry forms. Most of the OO frameworks don't
support this so you need a class to manage the transition between pages.

Andy Dent (A.D. Software - Mac, DOS & Windows GUI dev.)
94 Bermuda Dve, BALLAJURA  Western Australia  6066
Phone/Fax: 09 249 2719 (local)  +619 249 2719 (International)