This document is basically my to-do and idea list.

Todo

  • auto periodic uploading, downloading
  • allow links in description
  • Add eval should make real nodes, not just concat a string
  • blinking eys
  • add "auto-scan downloaded files under ___KB" option.
  • http upload to special server.
  • Security hole: User could add <a xmlns="http://www.w3.org/1999/xhtml"> to description, use tag to run javascript in extention security context..
  • beautify report window, add quick report buttons.
  • instead of nuking the whole trustbase on each informerreload, do an incremental comparison. So if a site is not accesible (high traffic) you just keep the old one.
  • alert people when there are new informers, reports in their DB.
  • parse non-xml responseText for judgements and informers. (OUTFOXED REPORTS)
  • Add "add me" button to profile pages (use the help:about trick again.)
  • Add custom attributions - [implement as exported favorites! (folder name = category)]
  • Add equivalence (and subset) assertion
  • Optional delay between report downloads (for backgrounding)
  • add "Search RegEx" option for other search engines
  • Add "collect bugs" option, and collect bugs and give option to send.
  • Add RFC option. Implement as... attribute in trust? degree of "rfc"?!!!
  • Pop up window when an RFC is answered.
  • disregard initial www. and such.
  • add date checker to downloader. (save wanderingstan.com!)
  • add download reminder (ever X days). Do it on first boot-up.
  • add pictorial and textual display of your connection to informant in the review reports page (You->I1->I2->Reporter)
  • add everything from bookmarks? (use XSL to transform)
  • search within trustbase..but how? !!! GOt it! But you need to be google...
  • add options for custom css changes for markups
  • convert local file browsing pages to xml, translate with customizable xsl
  • add ".." to file browsing
  • add option to limit search depth by numer of reports (no more than 3000, eg)
  • ability to delete reports
  • skip fingerprint check of image formats
  • some quick way to jump to your own page...
  • catch window close, and trigger an upload at that time?
  • how to handle language? a language tag indicating the lang of the uri, (and one for the description?)
  • add "search in comments" feature. (e.g. "funny", "scary"...) -- more important, search del.icio.us[style] tags
  • animated gif for "outfoxed is loading" button at beginning.
  • in report box, indicate if review is for exact URI or for a generalized version.
  • for livejournal, automatically assume their friends are informers
  • restructure trustDB.js so it can be used to create temporary trustbases.
  • add page upload style option (?)
  • Report viewer as normal XHTML page (linked from help:about?suzanneLookUpUri=)
  • Add count of NEW reports, informers to sidebad
  • Re-work database/window coordination. Sample Code
  • Store TrustDB in it's own file, not in prefs.
  • Fix date format to conform to RDF standard
  • Add count of reports/users to report sidebar
  • Generalize basic URL's so that entire site get's rated
  • Add MPL text to documents.
  • Welcom wizard
  • link to users from add/remove
  • Report viewer
  • Reformat XML into groups
  • Add name, homepage info
  • Add timestamp to informer in trustseed.
  • Filter informer list while growing / by untrusted URI's
  • Add hop info to quick-lookup, sort returned summaries (by depth, then age of informer record)
  • Edit reports (sort of)
  • Limit random page to Good pages
  • BUG every window is montitoring downloads...that won't do! need to keep a cache of files we've already checked (or something)
  • Detect changes in browser tab.
  • Catch new URL as soon as navigation begins, not waiting for page to load.
  • add a "hire me!" line to about box.
  • Fix profile download bug. Insure upload/download is silky smooth.
  • sidebar viewing of reports
  • show report listing in a sidebar
  • check for errors returned by FTP
  • Beautify trust page xsl
  • Page status (good/bad) shown before users are asked to accept a certificate.
  • switch file format to SHA-1
  • modify SHA algo to pause.
  • modify MD5 algo to take breaths, so browser doesn't kill it.

    • Bugs

  • First informer gets added twice
  • Duplicate reports when same file accessed via different URI's: trustbase.org/stan.xml vs www.trustbase.org/stan.xml (Do we need an internal ID? I think so.)
  • Can't handle country codes after a standard domain. I.e. he considers everything from *.com.au to be from the same domain. Yikes!
  • Can't use special characters (<,>,') in eval/informer info ... remove the string hack
  • After report is added, summary db gets rebuilt 2 times
  • Download (in options) doesn't work.
  • Description attribute for user doesn't stick
  • XUL Stack doesn't work on Mac (options page)
  • Similar problems on Linux ... need to observe
  • "Download from URI" in Options doesn't download Profile info

    • Usability

  • HOw to have users distinguist between rating an enitre domain (google.com), subdomain (scholoar.google.com), or a page (wanderingstan.com/links.html)?

    • Moritz

    - Hitting OK on the  dangerous page alert should repress  further warnings for this page (for this session? for this visit? hmm...)
- Marking of links destroys the page layout and might not work in any context - ideal would be tooltip on rollover, dont know of that's possible/viable
   (Stan: title attribute? statusbar message? mouseover style effect?)
- Sidebar needs layout and maybe quicklinks for menu functions.

    Caryn

  • quick key for reports
  • observe other people's reports without trusting

    • Ideas

    security:
as small rsa passwords to server?

use security protocol for safe email sending.

slidewinder intro
TrustWeb
issue: bad sites will have their URL's showing up a lot on peoples trust pages, raising their google rank!!
Terminology
		The xml file is the ... "trust feed", "trust file", "report",
		One opinions/eval is a  "evaluation", "report", "opinion"
		The person/business giving you trust info is a "informer", "expert", "trust source",

    future:

    Options to compress trust info (propose standard)
Security
Incremental download, continous spidering in background.
users can already rate each others xml pages, so this "reviewing of the reviewers" can be used to determine who's report rises to the top.
detect large changes in an informers page, flag as "suspicious" until X days have passed  -- may have been hacked
filter "random page" results by browser history, to ensure you see something new.
seach only pages/sites that are trusted
speed optimization -- remove comment from quicklookup table (it can be looked up when needed)
Trust is best with No single point of failure
set custom hoplimits for search commenting, page commenting, etc... ("only hilite pages from me and my immediate informers..")
integrate with thunderbird for fingerprint check, email check?
touchgraph

    Concerns

  • Is this giving out private info? Only what you want. And if you just want to leave a bad reviewof that one plumbing company that ripped you off, that's fine too.
  • Is it libel? Hundreds and thousands of people all expressing their opinion, and then only to those who want to download it, cannot be libel.
    • Outfoxed
A trust aggregator
Get Outfoxed, not outfoxed.
The Fox is with you.
May the fox be with you.
lonefox, lonewolf ... don't browse alone

    Features

  • warnings on download from untrusted server
  • Will give user feedback before they accept a certificate or install a plugin
  • option to do fingerprint check on downloaded files.