MeatballDistribution

MeatballWiki | RecentChanges | Random Page | Indices | Categories

This page is dedicated to the idea of having a wiki as a distributed application.

There are several possible uses of a distributed wiki;

redundant storage
private previewing
serverload distribution

there are also several implementation approaches to realizing some of these uses;

a single dedicated distributed application (see distriwiki)
an interwiki-protocol for index sharing
an interwiki-protocol for data exchange

==Killer Scenarios==

Below, write some scenarios that illustrate why and how a peer-to-peer distributed wiki might be useful.

Make sure the scenarios are specific and credible. Don't write in general terms of "the user". Make it a concrete, compelling story with concrete compelling characters. Unleash the writer in you ;-).

SiteThatIsAVictimOfItsSuccess
ShyPupilStartingToWrite
UnreliableServer? -- Individual wiki servers go offline for numerous reasons. Hard drive crash. Backhoe cuts through its net connection. System upgrades to latest version of Perl; later we discover that the wiki engine is incompatible with that version. Sysop loses interest, is not willing to put forth the effort to keep it online. Power goes out for a few days. Cat knocks the power plug out of the wall outlet (ask LionKimbro about this one). If we have a collection of servers, and any particular wiki page is hosted on more than one on-line server, then if any one server goes offline, that page can still be accessed through some other on-line server. Even though any one server may be unreliable, the servers are connected to each other in a system that (hopefully) compensates and is overall more reliable. See DavidCary, CommunityWiki:DistributedWiki.

MeatballDistribution

Discussion