Re: Nautilus/Medusa search index enhancements

	Did you consider the space implications of indexing a potentially
limitless 'net space' on a very finite local hard-disk - and the
scalability issues of having every single client re-indexing the whole
network and/or the security implications of the IPC necessary to do
intra-machine authentication / index querying ? and the scale of the
work to make that happen ?

Sorry.  You didn't understand the proposal *at all*.
1. The proposal is to have NO machine index any NFS servers. Au contraire, the search daemon would relay any search queries to Medusas running on remote NFS machines. Which NFS machines? The ones currently mounted, of course! 2. Security implications? AFAIK, if you're running a NFS network, you're trusting each client on their UIDs, so it's a pretty safe bet that you can use identd to identify the user who searched. 3. The scale of work isn't big. It's just a couple if's and a way to accept remote connections. Well, I'm an optimist, typical for software developers.

	I suspect that the (sensible) limitation of making this per-user, in
order to get _something_ that works nicely, soon, out there is likely to
result in a working deployed tool that can be used by real users for a
very useful sub-set of your big vision.
Evidently, it's a good purpose. But for networks it will cause problems when it starts indexing files in NFS servers.

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