next message in archive
next message in thread
previous message in archive
previous message in thread
Index of Subjects
account holders to date and there are several more gr I am looking at this ongoing debate as seperate but intermingled items: 1. a website redesign requiring some semi automated yahoo style index that indexes halifax/ns/canada related sites including - those hosted on CCN machines by both 'groups' (in either a old style IP account, or a old style user account (and we all know there is at least one group being run from a user account) or a new stlye group acount (ie special case single-user)) 'people', individules running meaningful (and not so meaningful) accounts from there ~user directories as well as people who run IP sites by themselves. - those hosted somewhere else (which in reality will be significantly greater in numbers than CCN hosted sites) 2. A new system of handling groups should be created (existing system improved): There is confusion over what a IP is (note all the cases that I mentioned above of users doing IPish stuff in ~'s, IPs being one person.. etc etc etc) and there is a definite lack of automation toola for current IP accounts. We have a system in place for handling quotas/expiries/office tools: the database and tools associated with normal individual accounts. The general case is that we need data of 'important web sites' for two reasons: admin: money accounting, computer accounting, expiration, etc. (web sites important because they run on our machines and consume our resources) ccn-hoo index: necessary information to build (daily, on demand, on the fly) the ccn-yahoo index (web sites that are important because they have halifax/ns/canada content) (not that not all CCN hosted sites would qualify) There would be a number of approaches to this: 1) Leave things the way they are. No money/systems accounting of group accounts, automated indexing impossible for sites not hosted on CCN. (or a almost completely useless automated index of the sites we do host) 2) Keep everything in the (next generation) IPDB: Add on money/systems accounting abilities to the IPDB, and tell it about all the non ccn/info area sites that should be in the indexed. Now all of the indexed sites (potentially thousands) have entries for billing contacts, expiration dates, quotas, etc.. Clearly unnecessary. 3) Maintain the namedb, ipdb, and a stuff-to-indexdb: Make the necessary changes to the IPDB (giving it the functionality now existing in the namedb) and build the stuff-to-indexdb, and associated tools. 4) Move IP/group accounts into the namedb (or just put new ones there), discard the IPDB, and focus on building the stuff-to-index-db. 1) is clearly not going to happen. 2) would require the most amount of work, and have us keeping unnecessary data around. 3) sounds not to bad, but if there is a separate stuff-to-indexdb then the namedb can be striped of all of its indexing knowledge (about 75% of the IP-DIRS cache), actually it would have to be striped of that data (or just ignore it) lest it get out of sync with the stuff-to-indexdb. And what do you have if you strip the IPDB of the indexing fields? A non-working namedb. Which leads to 4) Which is exactly what were doing. Treat the special case multiple authors as special cases.[1] [1](This will be solved with the release of the next WebDAV spec.. it'll do multiple authors, versioning, etc, over HTTP. And in those Web Folders you see on newish Windows machines.)
next message in archive
next message in thread
previous message in archive
previous message in thread
Index of Subjects