[libreoffice-users] Re: Base record access unacceptably slow

Le 28/07/11 18:22, Tom Cloyd a écrit :

Hi Tom,


> Where the &^%$ is the management - The Document Foundation - in all
> this, right now, today? Do they even watch this list? In short, do they
> give a damn that the only theoretically viable alternative to Access
> (for ordinary users) is in real trouble? Why aren't they showing up here
> with some clarifying position statement?

Been there, seen it, done that. The clarifying statements of opinion I
have had from the developers - they are after the all the ones that
write the code, either as volunteers and/or paid employees, are those
that Michael Meeks has expressed (albeit in other words). If you, as a
user, want Base to be actively supported, corrected, developed...then
sign up to a commercial support contract where your concerns will be
taken as a business case user by that support company, be it Novell,
Suse, Redhat, Lanedo, or whoever else might be in that loop. I would
actually probably pay for support if I knew that it would go towards
fixing the things I want fixed in Base. However, the support contracts
being offered are, to my knowledge, general support for the whole of the
suite, not specifically oriented to Base as such, in fact some of them
are even more general than that, i.e OS-support based, and some of them
are rather pricy. Again, price would probably not be an issue if I was
guaranteed that the money would go to paying someone to develop on Base.


History calls :
If one looks back to how the current HSQLDB came to be in Base, this was
only due to the fact that Sun helped pay Fred Toussi (the guy who
developed HSQLDB) to help them integrate it into OOo - otherwise it was
a non-starter. Fred also had donations from the community to help get
the work done.

So there you go, even back then, Base would not have become what it is
today without independent funding of development. Perhaps, a new set of
funding is indeed what is required to bring about changes in Base as it
is now, and develop for the future.


> I'm desperate for time, a fix, and vision of a long-term solution to
> this mess. I have work to do today, a lot of it, and I can't do it. I
> can't solve the problem, and other than by implementing the
> regress-your-java solution idea (which I have yet to be successful
> with). No one else is solving it, either. For some, migrating to another
> backend is not a challenge. For the rest of us, it's unknown territory.
> I researched this a bit, and while there certainly IS stuff out there
> about how to do it, there's not a lot, and there are multiple levels of
> challenge with this solution anyway.

Well, I jumped onto the mysql bandwagon very early on, before Base 2
even came into being. As Heinz has said ODBC worked fairly well even
back then and it was lightning fast. Unfortunately, things are not going
so smoothly with that solution for me now on OSX, where I can't get it
to work, even with a commercial paid-up ODBC driver.

Sun also came along and developed the mysql connector extension. This
actually still works fairly well for me, but none of the developers are
building it and making it available on the extensions site, which is
what Sun used to do each time a new version of OOo was released. There
has not been one single release of the connector extension by the
LibreOffice project since its inception. After enquiring over at the
Apache OOo project, the mysql is not part of the software grant from
Oracle, and so because the libmysql library is GPL code, it can't be
included in the Apache repositories and therefore the connector
extension will not be built and hosted by the Apache project. In other
words, unless someone else independently and repeatedly builds andhosts
the connector for each new version of OOo/LibO that comes out, and for
each platform, even the native mysql connector is doomed.

The other alternative to MyODBC or the native connector : JDBC
Connector, but again, this is not without several major problems,
including performance from within LibO, and date string management, and
blob and object management and, and...all of which are problems that
mysql (now Oracle) were aware of and never bothered to fix.

Anyway, other avenues are out there, I am exploring them as I speak,
because I'm not going to keep flogging a dead horse for much longer. If
it can be shown that its not actually dead, then I'm all for helping out
testing, documenting, etc, but if TDF itself is not prepared to clearly
show that this module is one of its centres and remain true to its
declaration of "protecting the investments of the past 10 years", I'm at
a loss to see what difference my willingness will make.


Alex





-- 
Unsubscribe instructions: E-mail to users+help@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/users/
All messages sent to this list will be publicly archived and cannot be deleted