[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: new documents

On Mon, 22 Mar 1999, Raphael Hertzog wrote:

> > 2) Reduce cost of ownership. I can see the value to the debian distribution of
> > all the complicated machinery to make it easier for package maintenance, eg. a
> > link back to the package for each variable, the IsDefault flag, variable meta
> > data, etc. These things are excellent ideas and would prove invaluable to
> > debian developers. However, I would hope that this configuration system would
> > become widespread and adopted by all linux software and distributions.
> > To implement all the debian package related things a special debian hierarchy.
> > I don't believe the power of the system would be decreased by this approach
> > and this would keep the system easily accessible to all developers.
> They would be useful for everyone, not only for Debian. I don't see the
> need of taking it away...
> The metadata could be implemented in a neutral way so that nothing would be
> Debian specific.

By creating a DataBase Interface specification you not only get to play
around with where the database is, but also with where pieces of it are. 
By taking a {front,back}end or {front,middle,back}end <shrug> approach
to the appropriate modules you get the flexibility you need to support
whatever system you want, without rewritting a bunch of code.

There will be stuff that is constant across systems, and stuff that
depends on the system being used (Debian, RH, etc.); the trick is to
identify which is which, and devise a scheme to abstract the function in
a way that is universally applicable.  Using the scheme on a RH system
should be as simple as replacing the dpkg specific DB and DBI module
with RH specific ones.  As the specs get written it will become obvious
where different ways of doing things necessitate an abstraction of the

The price of flexibility is the need to write lots of specs,
and put up with an end result that is less efficient than you would

- Bruce

Reply to: