This page is some initial thoughts and experiments with searching the Federated Wiki sites. What is id doing on Decentralised Academy site?
I have realised that my wiki-writing has been using it as a personal writing space: it helps me to think and to formulate my thoughts if I want to communicate them to others. I also use it to annotate links I found.
The answer is "What if you decentralised all the things, but you don't know where the things are? 
" source
The neighbourhood I usually search comes from people I know, so I find their wiki and reference it on my site: Ward, Mike, David etc.
At the same time I moan about the fact that many people I know use different tools for publishing that I can't federate properly - GitHub, MediaWiki, email (damn email, for publishing?), and Evernote. So I decided that I should crawl the Federation a bit more and get connected with people who already use Federated Wiki for their own purposes, so that I don't feel lonely.
Experiments
First I started to look at what people I reference fork from, and finally came across search.fed.wiki.org . This search is a bit weird - for example if I look for the word "trust" in this search and tick "sites" it will give me a bunch of welcome pages of the wikis that might have the word trust mentioned in them.
To find a page, not a wiki, I can either tick "page" when I search, which will give me a bunch of pages (and takes a while to load) or I click on each welcome page individually and do an inner search, which, afaik, searches the first paragraph of the pages in that single site or its neighbourhood, provided there is one.
Weird thing is, that many of the sites which the search.fed.wiki.org suggests when I request "trust" in "sites" don't find the word I was looking for in the "inner search"
Related pages and wikis
We're experimenting with full text search of the federation. For the moment this is both easier to code and easier to host than searching the web. search
Here we take a look at options for searching the federation and individual Fedwiki sites.
# Revisited
I repeated David's experiment. I searched for "trust" and stumbled across a page I was looking for just yesterday and had given up assuming I never wrote it. Now there are two versions with an interesting contrast.
Ward Cunningham and Kent Beck recall their work together in the Tektronix Computer Research Lab and how that has influenced their attitudes about programming throughout their careers.
Ed Roman reminds me ... It's been almost a year since I spoke at hack.summit.
The interesting thing is how I remember the outline for the largely unscripted conversation. In the original I recalled three important aspects of a talk we had given, the second involving trust, hence the hit.
Be Beautiful. Trust that you can know this quality and that you can produce it, even embody it, bountifully.
Our televised conversation recalled a talk we had prepared 25 years ago. I remember the title exactly, Be Valuable. Why did I not search for that? Because wiki's search today is lame compared to google.
Search for "trust" returns 411 titles alphabetized by domain name. A better approach would be to figure out how many ways the word is used and return the most useful site for each meaning.
Search for "be valuable", returns 433 titles, even more than the vague search "trust". Common words in uncommon order won't find the desired result when search disregards adjacency in the text.
A slow and dismal experience to be sure.
Distributed Search could do better. In David's case repeated searches moving from locus to locus could move him towards the conversations he would find interesting. In my case, I was looking for my own work so full text searching out from there is exactly what I needed.