Ideas
An open source list of half-baked ideas. Anyone is free to take these ideas and run with them.
Twitter audio client:
An account owner has an option to speak a string of words. Then people can just listen to their tweets being spoken aloud in their voice. If an account owner hasn't uploaded an audio sample then there could just be a default voice that the end user chooses.
As a user, instead of constantly checking twitter, you could just listen to speak in realtime while you work. It would probably be pretty weird.
Media could be introduced and autoplayed.
Discrete CSS diff
This idea has a few possible uses
-
Everytime someone tries to check in some css, they get instant feedback to let them know if they aren't introducing any new visual styles (rules) to the codebase. A list of classes referencing the same visual styles could be returned to promote reuse within a codebase.
-
Site checker. You could compare two different sites and get back and object that contained all of the unique css for each codebase in regards to rules (property value pairs).
-
Could compare the top 1000 sites on alexa and show the list of unique css amongst all of them vs the css that each site contains.
Navigation Visualizer
You could scrape websites and grab all links in the header and footer of each website. Could check for elements and id or classnames that are equal to a set of common names to denote nav: navbar, nav-bar, nav, site-links, etc.
Build a data visualization that shows common patterns in website navigation. i.e 80% of sites use store, 20% use shop. Things to potentially visualize:
- Whether links are in header or footer
- Order of links
- Text used for page names
Could divide visualizations by types of site, i.e e-commerce, news, startup, personal portfolio.
Automated component documentation
First grab all links for a given site. Could use something like:
wget
--spider --force-html -r -l2 $url 2>&1 \ | grep '^--' | awk '{ print $3 }' \ |
grep -v '\.\(css\|js\|png\|gif\|jpg\)$' \
> urls.txt
For each view generate an ast from the dom Across all pages, automate finding common markup patterns within a set threshold, build a documented component library for most common patterns.
Vim Analyzer
Every keystroke, command, and edit is tracked in vim.
You could build something that analyzes a .viminfo file that reports on your daily vim usage and gives you back hints, tips, and optimizations for more efficient text editing.
Could return a list of macros or functions for later use. Shows stats from daily usage and compares you to other people (opt-in).
You recorded 32 macros today
You recorded 16 macros that you've recorded previously
You used j more than once in a row 245 times. Consider using / or text objects to move more efficiently. This could save you x keystrokes and y time.