The technique is explained in more detail here: http://lukeroberts.us/2009/12/freelensing/. The freelensing group on flickr has some great work too: http://www.flickr.com/groups/freelensing/.

Try it out!
-Casey

Toronto Island Sunset from Casey Banner on Vimeo.

-Casey

Calculus Timelapse from Casey Banner on Vimeo.

-Casey

The first problem I anticipated running into would be making remote queries to the Twitter API via AJAX. Since browsers enforce strict rules on what domains you can make remote requests to (you can only make requests to the same domain as the page running the javascript), I wouldn’t be able to get the JSON feed that way. This is a security limitation, so people can’t make malicious remote requests.

Normally what you do in a situation like this is to create a server-side proxy running on the same domain as your page, which makes the actual request to the remote domain, and returns the results. Since the proxy is on the same domain, the AJAX request is allowed and everything works fine. However, I had resolved not to use any server-side code to grab the Twitter feed, so that option was out.

After a little bit of research, I found about a slightly sneaky technique that people are using to load remote JSON without needing a server-side proxy. This is how it works:

If you add a <script> tag to the pages’s DOM, the browser will immediately run its contents. So, all you need to do is add a <script> tag to your page, with the URL of the remote JSON you want evaluated as the ’src’ attribute. Once the tag is added to the DOM, the browser fetches and evaluates the script.

Alone, this doesn’t do us much good. Sure, we can evaluate a bunch of JSON, but we can’t access that data from the other Javascript on the page since it wasn’t actually stored anywhere. Fortunately for us, most APIs that provide data in the JSON format also accept a ‘callback’ parameter. Provide that parameter, and the API will return your JSON, nicely wrapped in a function call to whatever you provided as your callback. Now, when that <script> tag gets evaluated, your callback function gets called with the data as its argument. Cross-domain JSON without anything done by us server-side!

Here is what it looks like in practice:

function remote_query() {
    url = 'http://api.twitter.com/1/statuses/user_timeline/kcbanner.json?callback=twitter';
    var script = document.createElement('script');
    script.src = url;
    document.body.appendChild(script);
}
 
function twitter(data) {
    // Do something with the results.
}

If you load that twitter url, you’ll see that it contains something like:

twitter([ A whole bunch of JSON ]);

So once the browser loads that script tag, it gets evaluated, and our twitter function gets called. No need for any XMLHttpRequests or server-side scripts.

Hopefully this was helpful!

Update: Apparently this is a well known technique called JSONP. I completely missed that in my googling. Thanks commenters!

-Casey

URxvt.perl-ext-common: default,tabbed

to your ~/.Xdefaults file, and restart your terminal.

You’ll get something like this:

The keyboard shortcuts are what make this awesome. Shift-down to create a new tab (or click ‘new’), and Shift-left/right to cycle between tabs.

-Casey

Anyway, what got me thinking about this was listening to some awesome deep house mixes from http://soulshine.org/. He also has a great Sunday Snooze, and Laidback Niceness series for down-tempo, chilled out stuff.

-Casey

Now to combat this horrible problem I whipped up this script to use the Twitter search API to grab the last bunch of @ replies between two users.

Use it like:

$ python banter.py nathanfillion simonpegg

import urllib2
import simplejson
 
 
def tweet_date_cmp(a, b):
    if a['id'] > b['id']:
        return 1
    elif a['id'] == b['id']:
        return 0
    else: #a['id'] < b['id']:
        return -1
 
def get_thread(user_a, user_b):
    url_t = "http://search.twitter.com/search.json?from=%s&to=%s"
 
    to_a = simplejson.loads(
        urllib2.urlopen(url_t % (user_b, user_a)).read())
    to_b = simplejson.loads(
        urllib2.urlopen(url_t % (user_a, user_b)).read())
 
    thread = []
    thread.extend(to_a['results'])
    thread.extend(to_b['results'])
 
    return sorted(thread, tweet_date_cmp)
 
 
if __name__ == '__main__':
    import sys
 
    thread = get_thread(sys.argv[1], sys.argv[2])
    for item in thread:
        print "%s: %s" % (item['from_user'], item['text'])

I might code this up into a simple web app to make it more accessible.

Have fun, and don’t stalk too many people.
-Casey ]]> http://www.caseybanner.ca/2010/03/31/banter-view-twitter-conversations/feed/ 0 http://www.caseybanner.ca/2010/03/29/fabrication-of-waves/ http://www.caseybanner.ca/2010/03/29/fabrication-of-waves/#comments Mon, 29 Mar 2010 19:59:52 +0000 Casey http://www.caseybanner.ca/?p=231 Recently I have succumbed to the never-ending struggle between studying things that I actually enjoy. Now, don’t jump on me and tell me that I should study things that I enjoy. I enjoy programming, but taking a course on it where you are forced into a backwards method of doing things starts to take its toll. Anyway, I digress.

I’ve been really getting into ambient and minimal stuff; for some reason I tend to tap my foot to a minimal track more than something with a crazy beat.

Here are a couple tracks that I’ve been working on over the past few days. Perhaps some slight tweaking is needed but I’m fairly happy with how these turned out.

Depth Perception by kcbanner

Cleanroom by kcbanner

Enjoy. If you wanted higher bitrate versions, I would be happy to provide them.

Calculated Scribblings  by  kcbanner

Enjoy!

-Casey

I don’t know why but I really really like timelapses. Yeoow.

-Casey