I rode a Greyhound bus on my crusade of saving a few bucks. I always had issues with buses, when I was a child I used to go from Celaya to Mexico City (and back) almost every weekend for a couple of years and I developed a phobia against the whole bus experience. Even though, getting to Pittsburgh from New York City was cheaper riding a bus than the other alternatives anyway, flying or train. Anyways, it was an awful experience that prevented me from being complete recovered after each travel leg.

Staying at the CMU dorms was also a nice and pleasant experience, feeling for three days like I was a kid on a nice American college. I might want to do that some time in the future pursuing a Master's degree (some day).

At YAPC, I had the nice chance to get to know a lot of people I'd only known by mailing lists and CPAN modules authors that I use on a daily basis. I also had attended YAPC 08 last year in Chicago, so I already knew a small group of mongers.

I delivered my talk about Perl and social networks on Tuesday. I gave it an small twist by moving it into a more motivational, business perspective for hackers, in the last minute. I was overall satisfied with it, and I also think some things were not well done and had to be changed. Fortunately, quite some people came to me and gave good, positive, constructive feedback about it, that I was very thankful to receive, and unfortunately some people also threw destructive criticism that only makes you think how douche-baggy people in the community you can be. In any way, I was happy and satisfied, some observations:

• I have to take out a lot of personal stuff I poured into it because people don't care (and it's not their fault, it's mine).
• Focusing more on actual code. I can try to focus on motivating people to adventure into social networks application development using Perl, but the YAPC audience is more into actual development and coding techniques than money-making ideas.
• I still have to practice my speech. I took the time to write my lines scripts for each one of the slides, but that makes the interaction with the people a bit more difficult to establish.

And some interesting points around YAPC in general:

• In the auction, event that is done every year helping TPF raise some funds and that a similar approach would be great to have in conferences like DebConf, the highest bid for a lunch with Larry Wall reached one thousand dollars. I was a bit surprised.
• They took us in a tour at Heinz Field, home of the Pittsburgh Steelers, it was awesome, although it was already a bit late and couldn't take good pictures of the field. Pictures are on this Flickr set.
• I had some time at the end of my stay in Pittsburgh so I took the opportunity to go to PNC Park and watch the Pittsburgh Pirates kick one of their greatest rivals (as I learned there, because of the proximity), the Cleveland Indians. Pictures are on this set.
• Some people outside the venue also joined me on the conference as I setup an improvised ustream transmission. Unfortunately, I never got to press the "Start recording" button, so no video of it. Bummer. However, God willing, I will record my next delivery of the talk on YAPC::EU in Lisbon, in early August.

I've placed the slides on a PDF here. You can also clone the entire repository containing the source Keynote file.

So, I spent a few minutes and came up with this simple Ruby script that uses John Nunemaker's awesome Twitter gem.

#!/opt/local/bin/ruby

require "rubygems"
require "twitter"

httpauth = Twitter::HTTPAuth.new(
	ARGV[0] || 'yehyeh',
	ARGV[1] || 'kissm3'
)

base = Twitter::Base.new(httpauth)

i = 0
(base.follower_ids - base.friend_ids).each do |id|
  i += 1
  begin
    base.friendship_create id
  rescue Twitter::General => e
    puts "#{e.class}: #{e.message}"
  end
end
puts "#{i} new friendships."

i = 0
(base.friend_ids - base.follower_ids).each do |id|
  i += 1
  base.friendship_destroy id
end
puts "#{i} destroyed friendships."

puts "#{base.friend_ids.size} friends now."
puts "#{base.follower_ids.size} followers now."

What this little code does is exactly that, it will start following the followers you don't follow yet, and it will stop following the people that don't follow you back, right? Got it? It's basically synchronizing your friends with your followers. As said, this is particularly helpful when you are maintaining a community account and want to keep up befriending your kind followers Handling the exception Twitter::General on line 18 is only done because the twitter gem raises it even when you are trying to befriend an account to which you have already requested friendship (like pending requests to protected updates accounts) or those of suspended accounts (spammers).

Well, let's just do it already.

#!/opt/local/bin/perl

use Modern::Perl;

use Net::Twitter;
use DateTime::Format::DateParse;
use DateTime;

binmode STDOUT, ":utf8";

my $t = Net::Twitter->new(
    user => shift @ARGV || 'lazy_fuck',
    password => shift @ARGV || 'bl0wm3');

my $whitelist = [qw/maggit rubx axiombox/];

my $then = DateTime->now;
$then->subtract(days => 5);

my $x = 1;
for my $i (1..80) {

    my $tweets = $t->user_timeline({ page => $i, count => 200 })
        or die "No fish for you, loser.\n";

    for my $h (@$tweets) {
        next unless $h->{"in_reply_to_screen_name"};
        next if grep { $_ eq $h->{"in_reply_to_screen_name"} } @$whitelist;

        my $date = DateTime::Format::DateParse->parse_datetime($h->{"created_at"});
        next unless $date < $then;

        say $x, ": (", $h->{"id"}, ") ", $h->{"text"};
        $t->destroy_status($h->{"id"});
        $x++;
    }
}

This little fucker will try to fetch your latest 16'000 tweets (if you have… twat, *grin*, more than that, you've got real issues and I cannot help there, get a shrink or something).

It'll make 80 requests for your timeline (remember Twitter gives you a 100-request hour limit), so only if the reply doesn't come from a certain people AND the reply is older than a given period of time (I'm setting it to one week for me), it'll get rid of it. If a friend doesn't see the reply in a week, she probably never will. After that, it just destroys the tweet (or it tries at least, from my experience, Twitter is still experimenting a hell lot of issues on their service).

That way you can cheat the system removing useless tweets that no one (not even you or the recipient) cares about anymore. Or just… don't give a shit, and you are all set too.

The first time I ran it, I went from like 6k tweets to 2500, which was a nice drop . If you feel like it, just grab it and customize to fit your needs. You will be needing the CPAN modules Modern::Perl, DateTime, DateTime::Format::DateParse and of course, Net::Twitter.

Having @debian on Twitter is actually kind of awesome!

You will have to register your application on Twitter previously. You can do so here. If this is a web application that you will be building, you can provide a callback URL which is the page where the user will get redirected once she has granted access to your application. If you just want to test, setting a desktop application is probably the way to go.

Once you have registered your application, you will get two strings, key and secret consumer. Refer to general OAuth documentation for deeper details.

Install Net::Twitter::OAuth as any other Perl module:

$ sudo cpan Net::Twitter::OAuth

Now, using it is very simple:

my $client = Net::Twitter::OAuth->new(
  consumer_key    => "YOUR-CONSUMER-KEY",
  consumer_secret => "YOUR-CONSUMER-SECRET",
);

No transactions or requests have been made yet. Here you need the user's access and secret tokens. If you already have them, which means that the user has already gone through the authorization process, you have to pass it now (you already stored them on database, a configuration file or whatever the data model you use):

if ($access_token && $access_token_secret) {
  $client->oauth->access_token($access_token);
  $client->oauth->access_token_secret($access_token_secret);
}

Now you can query Twitter so it can provide you access:

unless ($client->oauth->authorized) {
  # The client is not yet authorized: Do it now
  print "Authorize this app at ", $client->oauth->get_authorization_url, " and hit RET\n";

  <STDIN>; # wait for input

  my($access_token, $access_token_secret) = $client->oauth->request_access_token;
  save_tokens($access_token, $access_token_secret); # if necessary
}

All these snippets come from the example's of Net::Twitter::OAuth. So basically, if you are not authorized, which means that either the user hasn't even been prompted for authorization or denied access before, then you get the authorization URL which you can give to the user to visit.

Once the user has granted access, you can call request_access_token which will return the user's tokens. Here's where you can save those tokens for future use.

After that block, you are pretty much done and can use the regular Net::Twitter methods:

my $res = $client->update({ status => 'me ownz oauth!!1' });

Soon, a real life application post using HTTP::Engine, KiokuDB and Net::Twitter::OAuth.

During the last week, I received notification that the talk I've proposed for both YAPC::NA and YAPC::EU was accepted.

I will be talking about Perl in the Time of Social Networks, which is mainly two broad things: how Perl can actively power a modern social networking site (I'll be releasing a free -as in speech- social network website, similar to Lovd or RailSpace, built with pure Catalyst); and about high-load applications dealing with social networks' APIs and users (a Moose-based universal social network distributor), with Twitter, Facebook, MySpace, Blogger, LastFM, and others. My current contractor has been sponsoring the development and research for this since almost two years now, and we believe it's ready to go public: there's a lot of work backing us up.

I'll be presenting the talk in Pittsburgh, PA, on June 23rd at YAPC|10. If you are around the area, come and attend the talk!

On YAPC::EU::2009, I will be presenting in Lisbon, Portugal, on August 4th, which is conveniently placed just after DebConf 9, which will happen in Cáceres, Spain the week before (I will also be there).

Also, I must say, they flew away immediately with work buddies, so from the 8 stickers the guys kindly sent, I only kept two, but that's cool. I guess

Thanks GitHub!

I love regular expressions (to which some people say there's nothing regular about) and I just found TweetBeep recently, which is a simple alerts system. There's nothing really new or innovative about it, given that you can get the results yourself using advanced search on Twitter and consuming the results from a feed. However, I for sure, I don't want to spend time setting any of that up, and they do that quite easily, which is the whole difference, ease of use. You can get the results on time intervals via mail. Rocks.

So, one of my alerts is about regular expressions: I'm receiving everything Twitter tweets about them, whether regex, regexp, regexes, regexps, etc. are mentioned, I get them. Now, I feel a bit nerdish/dorky about it given that I like to reply on those tweets on totally random people and try help them whenever it's possible (or just learn a bit more from other regex gurus), but what the hell.

Now, the good thing about this is that being a "good person" and fellow regex buddy with people is starting to pay off. Whether I'm getting new very interesting followers, I also found a nice opportunities niche to take advantage of, and a couple of gigs have come up. I worship building strong relationships with people (as in long-term clients), and when starting to get to know some persons via a common interesting thing such as a regular expressions is just awesome, little piece of awesomeness.

For the geeky part of this post, if my regular expression Twitter alert had to be written as a regex, this would be it:

/regex(?:(ps?|es))?/

The questions now is, should I start following /regular\s*expres{1,2}ions?/ too? Maybe I will