RE: Persistence Library Needed in Boost? (Was: [boost] A tip from the Dr.)

From: Jack M. Thompson (JMThompson_at_[hidden])
Date: 2002-01-24 23:14:37

Good input.

It should be accepted as a need.  I need something to manage persistent
storage all the time.  The trouble is that there are many mechanisms to
use, but I don't know of any solutions that meet all of the criteria to
be an international standard.

I have to admit that I don't really know exactly what the criteria is,
but I see it as being:
        1) completely platform independent
        2) free and highly available
        3) open source
        4) completely generic
        5) useful
         

My idea fits these needs quite well.  I haven't actually compiled it on
any more than a couple different windows platforms, but I know that it
is platform independent because I only used simple C++ and STL
containers (using STLPort on Windows CE).  Even if there are some
problems compiling on different systems because of things like naming
collisions or lack of support for namespaces or other facilities, the
concept is platform independent.

The only part that I know of that needs to be specialized for each
platform are the IO routines.  There are a couple of abstract objects in
the system that need to be derived into the Input and Output objects.
There is only a little work to be done, and it is necessary to make
these specialize-able.

I have to admit, I would be thrilled if I could be the one to start this
and see it through, but I would be very interested at this point in any
other libraries that meet the same criteria.  Do you know of any?

> That's rather presumptuous of you, isn't it? ;-}
I'm not sure what you are saying is presumptuous.  But I know that the
concept is rock solid.  I could possibly have a bug or two lurking
somewhere, but I don't think so.  I know there could be improvements
made to it, but the concept is still very useful.  

-----Original Message-----
From: Stewart, Robert [mailto:stewart_at_[hidden]] 
Sent: Thursday, January 24, 2002 5:19 AM
To: 'boost_at_[hidden]'
Subject: RE: Persistence Library Needed in Boost? (Was: [boost] A tip
from the Dr.)

> From:	jackthompson99 [SMTP:JMThompson_at_[hidden]]

[snip long description of persistence mechanism]

> So, I don't know if this is the best way to do it, but we do need 
> something simple that uses ONLY C++.  The parser could probably be 

Here is the first statement of your premise: A C++-only persistence
mechanism.  Provided that is accepted as a need, then your library may
be
the solution.

> faster and smaller, but the concept is rock solid.

That's rather presumptuous of you, isn't it? ;-}

> What do you think?????

There are numerous C++-only persistence mechanisms.  For example,
RogueWave
created one years ago, though I don't know how current it is, and it
certainly isn't open for standardization.  The question remains whether
this
is a worthwhile goal.  There are object databases that are less
intrusive
than your singly-rooted object hierarchy, and they are probably faster.
So
I have to ask, why reinvent the wheel?

How does your framework handle class templates?  How does your
singly-rooted
design interact with other singly-rooted hierarchies?  Is your binary
representation portable?

Have you tested your framework on different platforms?  Is your
technique
compiler specific?  Is it OS specific?  I'd guess you've implemented
your
framework with MSVC on Windows of some flavor. There may be problems
lurking
that will be showstoppers on other platforms.

I'm not trying to be negative, just realistic.  There may be great value
in
your proposition.

Rob
Susquehanna International Group, LLP
http://www.sig.com

Info: http://www.boost.org  Send unsubscribe requests to:
<mailto:boost-unsubscribe_at_[hidden]> 

Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/