Wiki Zip

First mentioned on WikiWipeout, this is an interesting idea to extend the "saveability" of a Wiki. Don't forget, malicious users aren't the only source of lost data.

What would be the most efficient means to do this? Instinctively, a co-Wiki program to send the raw database would be the absolute fastest, but introduces annoying dependencies. A HTML-ized tarball of all nodes might be useful, but there would be a lack of metadata.


Try this: On your wiki, the page WikiZip has a clicker to a zip file called

When your wiki's server detects a download on this file, it zips up all your Wiki pages and metadata pages and returns this.

MiniRubyWiki now has this feature. --PhlIp

or this:

Disable the deletion of 'any' files on the server (yes, modify the OS ... ). Append or create new versions of everything in a constant stream. Allow the creator of a page to designate an "official" version, and let users choose if they want the most up to date version or the approved version. Hope disk space keeps getting cheaper and cheaper (it will). Stick to text-only at first.

Let's make a distributed mini wiki server that initially uses only RAM (the first users can easily boot from a CDROM when they want to try out participating without jacking their disk partitions), searches for and automatically stores data redundantly on nearby (lowest latency) nodes. Initially it would be like 4 full synchronized copies of everything, soon some genius will adapt RAID concepts to save space and add cascading cache (like bittorrent, Of course we could run out of RAM quickly with photos and video, so we will need to be able to resize and create partitions and/or use free space on existing hard drives partitions (better idea!) in the near future.

The number one feature of this system is that it is Public Domain (no copyright license what-so-ever) so that it is free to be adapted into commercial products and free to evolve as fast as possible. For more on this idea and others, see


EditText of this page (last edited September 2, 2004) or FindPage with title or text search