Thought for the day: Always follow the path that maintains one's level of motivation and momentum.

For all of those who have asked about how things are at TurboLinux, the simple answer is as busy as ever, with an added bit of focus. The industry and the markets always help on focusing a company as well as the individual.

Onto more important things...

I'm still floundering on one bit. LDAP makes for a great top-down heirarchically controlled directory service with centralized control. However, there is a lot of fore-thought required, and seeding an ldap server for an organization is daunting. Its even more daunting to figure out how to seed a low leaf among multiple ldap servers (say a workgroup server). The only effective bottom-to-mid upward solution to date has been Netinfo, which lets the local cluster define domains and then one can tie them together after the fact. For a NOS -- management tools are key, but you have to be able to aptly manage your local data. Otherwise you end up with complex tools to manage impossibly complex data. One can only hope for an extremely high degree of centralization in an organization to limit this problem.

My feelings (which to a degree have been related to Luke Howard -- lukeh here), are that maybe a combination of the two is necessary. The equivalent un*x workstation/server mayhap needs an improved, secure Netinfo style authentication and management system, with adapters and tie-ins into higher level LDAP directory services. Agents must exist to help propogate loosely managed info into the larger centrally controlled LDAP DSes, and likewise global sentings can by profile be approved into the local realms of the clusters.

This could get ugly. But unless the end user or group manager can effectively manage his applications and data, directory services are a large undertaking that can quickly become a hindrance and would be summarily dropped.

What do you think? You can email myself as always with your ideas.

I'm also hoping to apply these a bit more back at Stanford. Should get more involved in my co-authored Linux distribution, SULINUX . I've wondered off from that for quite a bit...