A different approach
jhuntwork at linuxfromscratch.org
Fri Jan 7 13:32:29 PST 2005
So, it's been good to have a break. But I think I (and the devs) are
ready to start working on alfs again. After thinking over our current
position, I think it's best we approach things a little differently
To date, much of my approach has been to try to pick up the pieces where
Kevin left off, trying to use ideas that have been jotted down, and as
has been suggested, get a fully documented process in place before any
development takes place. But, thinking about it more, it seems to me
that while we definitely need a structure and goals to hit, much of that
can be defined as we work. A simple skeleton is really all that is
necessary for an SRS, at least at the moment.
Therefore, I'd like to set forth a milestone structure. We know the
basic things we'd like done, so a simple milestone can be set forth to
be reached. While we work we can add/define smaller increments that
will help fulfill the milestone. Advanced features that are not
necessary for a basic structure can be put off for future milestones.
As we develop the tools and milestones are reached, a more comprehensive
documentation can be written on what we've done and will do. For this,
James, you'll have to be in good communication with Jamie and Boris, and
For one of the very first aspects, I think we need to design a client
server model and communication protocol. That way any client can be
built to communicate with the server and log errors.
This is the direction I'd like to see us take in development. By
definitely stating where we're going, to, I hope to dissolve any
confusion that's been hanging around the project, and get firmly in
A list of the milestones is being drawn up currently.
More information about the alfs-discuss