Lately, all my primary technology news sources have been polluted with articles about Twitter and URLshorteningservices. A part of my dies inside every time I hear about one of these. I can understand the hype about Twitter, but the only reason there is a hype about URL shortening is because of Twitter.
I have an open suggestion to Twitter that would benefit them, their user base, and hopefully kill the URL Shortening market space in one go. (so it would be one less thing to pollute my daily tech news with).
Solution
Twitter, implement your own shortening service and integrate it with the service. There is #hashtags, and there is @mentions. Using the same convention and use either !% or ^ to implement URL redirection.
Example
"Hey @person, go check out ^s89n3"
If a user wants to shorten the url http://www.janitha.com, the service would map that url to a code such as "1a2b3c". Now if a user enters ^1a2b3c, it links directly to the original URL (or links to something like http://www.twitter.com/redir/1a2b3c that redirects to the correct place). It's really trivial, everyone and their grandmother is implementing URL Shortening services, so I am sure you can spare a fraction of your millions in funding to implement this simple feature.
Last night I was wondering how interesting it would be to have your friends Facebook status updates, twitter tweets, and other types of live feeds be read back as speech in real time as their updated.
A coffee cup later, jFeedSpeak was written. It's a Feed-to-Speech script I wrote in python and it accepts any number of RSS feeds (facebook status updates, twitter, Google news, pretty much any RSS you throw at it). The newest updates are piped in to Festival (Text to Speech engine) which you can listen to (and optionally pynotify that pops up a notification, but I commented that out for compatibility).
The default script uses the Twitter public feed, but it can become really annoying, so I highly suggest you modify and add more personalized feeds to it. Instructions and configuration inside the script. Enjoy!
That was my Sunday afternoon project. Inspired by @joshuabaer's tweet looking for such service, decided to implement one just for fun today. Sorting through the data was fun, loads of wget, bash scripts to cut, awk, grep and process, python to mix and match, some php to whip out a interface, and sweet little mysql db to store it all.
It has a resolution per city using area code & prefix (123-123-....), and it turns out the US has a total of 155,972 unique geographically based telephone sectors.
You can feed it a list of phone numbers in any order, or even a chunk of text with phone numbers inside it and it will try it's best to extract all the valid phone numbers. Valid formats include 1231231234, 123-123-1234, (123)123-1234, 123.123.1234, etc... yay for regular expressions.
I need to start using the google maps api more frequently, maybe next weekend project.
I was tired of seeing the same variety of CAPTCHA systems online that consist mainly of letters and numbers. These have gotten so complex, they are close to impossible for even humans to guess. Since my new site needed a captcha system to stop the comment spam, I decided to write my own captcha just for fun, except instead of letters, using shapes. It's very simple, yet quite effective (no spam comments since I launched, compared to several hundred spam comments with the last wordpress with a conventional captcha system).
The current version uses circles and boxes and prompts the user to enter the count. The shape overlaps and deformations make it pretty hard for a bot to guess (but not impossible). Compared to a using a regular captcha, the biggest strength in rolling your own captcha is that no one would try to create a bot or find a already existing one to break that since the captcha system is rare (such as this one). Security through obscurity in a way. So if you have a big comment spam problem, and have a hour or two to spare, I would highly recommend you create your own captcha variant.
When you are developing your own, make sure to bind the corresponding captcha image to the session, ip, unique key, and what ever else you want and make sure to have a expiration method as well as have a failure threshold to thwart brute force attacks.
So lately I've been hanging out a lot at Coffee Station and on campus using my Laptop. Doing this, one of the biggest annoyances I had was whenever I had to leave the area to get a coffee or something, I risk leaving the Laptop unattended. Using a cable lock every time was annoying, so instead I wrote a anti theft alarm script that utilizes the Laptop's built in accelerometer. Originally it's meant to be used for the HDAPS, or Hard Disk Active Protection System, to detect shock and prevent data corruption, but I will be using it for a different purpose. Here is a demo video...
This is how it works. Whenever I am about to leave the laptop, I start the script which locks the laptop and start the gnome screensaver. Any small bumps or tilting up to 20 degrees of the laptop will produce a small chirp sound where the volume is directly proportional to how tilted it is (so the more you tilt it, the more loud it becomes) as a initial warning to anyone who bumps or laptop or for small movements. Any movement beyond 20 degrees of tilt will start the alarm at the highest volume (and the ThinkPad speakers are Loud). The script monitors the gnome screensaver status, so the only way to stop the alarm or to unarm the system is to log in using the password or a finger swipe which the gnome screensaver.
If you want to get this working on your linux laptop, make sure you have HDAPS on your laptop and Tp_smapi installed. There are other anti theft implementations that use the HDAPS data you can find here, but none of them integrated well with the gnome screensaver or have the initial warning feature. That's all for now, hope you found this useful.
Driven by a strange fascination about how things and society work and a mysterious craving for new information and experiences, you will find me doing a little bit of everything.
Website Under Construction
These days I am slowly rewriting and redesigning the website from scratch, excuse me if you come across any bugs or non working parts of the site. Let me know your comments or suggestions.
