Title Network Browser

The browser shows five columns each with titles of pages holding links leading to the right. site

From one starting point this web application displays reachable titles following links forward or backwards two hops. Find a link to this interactive tool in the caption to the right.

We consider how pages might link to one another without regard to whether they are from the same site or even the same neighborhood.

If we think of links producing children then we show five generations of pages: dads and granddads before and sons and grandsons after, each in their own column.

We're using the normalized version of page titles called slugs. These are all lower case and punctuated with hyphens. We show these with hyphens rendered as spaces to approximate the true page title. See Search Index Downloads

For each title of the visible federation we know how many copies have been made and collectively what outbound links are present among them. slug-web.js

"how-to-wiki": { "forks": 37, "links": [ "add-pages", "add-paragraphs", "copy-pages", "find-sites", "follow-links" ] }

We sum the forks for each title and then show this and the title count as a column heading. Dividing these two numbers gives you the average degree of duplication.

# Operation

Start the network browser with one or more slugs written as the query parameter separated with commas.

http://search.fed.wiki.org:3030/title.html?tektronix

The center column shows the starting titles.

The left two columns show titles with links to the right.

The right two columns show titles with links to further right.

Click a link to bring it to the center and adjust remaining columns accordingly.

Shift-click a link to open the search interface searching for links to the selected title.

# Download

This app will load and cache a new data file when one is available. On a slow connection this can take minutes. However, for reliably fast use once can save the two required text files, html and js, and launch the app from these local files. title.html slug-web.js

file:///Users/ward/Downloads/title.html?tektronix

# Adjustments

We will make adjustments to the algorithms as we find ways that this application is more or less useful.

We found that there were a few widely used titles that polluted results with way too many unrelated links. We now ignore any title that appears in more than 50 sites.

Object.keys(slugs).filter( function(s){return slugs[s].forks>50}) ["scratch", "welcome-visitors", "ward-cunningham", "currently-working-on", "recent-changes"]

The algorithm change does not exclude high fan-in/fan-out sites as a starting point. welcome-visitors