Eric Lease Morgan wrote:

> On Mar 20, 2007, at 10:02 AM, Laurence Finston wrote:

>
> Laurence, this is very interesting, and thank you for brining it to
> our attention.

Thank you.

>
> To what degree do you see the development of the Exchange Utilities
> moving away from a Microsoft-based environment?

My intention is for the package to run on free systems using only free
software.  In practice, this means first and foremost GNU/Linux, to be
followed, if possible, by FreeBSD.  All dependencies on Microsoft products
will be removed _unless_ the same functionality is available for free
systems.  This is in accordance with the GNU Coding Standards, as well as
my own wishes.  I would be perfectly happy for the package to run on
Microsoft systems, but this is a secondary consideration.

> Much of the code is
> written in C++ (++), but don't know how portable that is. Moreover,
> some of the descriptive text alludes to Microsoft SQL Server as the
> underlying database. Alas, many of us don't have access this RDMS.

The programs are written entirely in C++.  In the earlier parts, namely
`ATest', which accesses OAI servers, I used more features specific to
Visual Studio and Visual C++.  In `ZTest', which accesses data from Z39.50
servers using the YAZ package, tried to avoid using these features and
used standard C++ and the Standard Template Library as much as possible.
The most recent program, `scantest', which implements the beginnings of
what I call a "generic query language" uses GCC (the GNU Compiler
Collection) and runs under GNU/Linux.  Standard C++ is quite portable,
features from the Microsoft libraries are virtually non-portable and must
be replaced.  Threads may be a problem.  I plan to use Posix threads,
which are supported on all UNIX-like systems (to the best of my
knowledge).

I no longer have access to a Microsoft system and will not support the
Microsoft versions.  I plan to build the package up around `scantest',
which I will be renaming.  I will probably call it the "GNU Generic Query
Language Interpreter".  At some later date, I may port the projected
GNU/Linux version to Windows (or even MS-DOS).  However, I would prefer to
do this using GCC and cross-compilation rather than Visual Studio.

I don't plan to use Microsoft SQL Server anymore.  The SQL code I've
written seems to be reasonably portable.  I don't think it will be
difficult to adapt it to other database software using other versions of
SQL.

Of course, if I do find a library or other institution interested in
supporting this work, that institution would have a say in how I proceed,
as long as the requirements of the GNU Project are satisfied.

Thank you for your interest.

Laurence Finston