Defining Chaos In Order

An interesting Webcast has just popped up, on MySQL’s usage in one of the biggest databases online: Netlog.

Netlog, one of the fastest-growing Web communities in Europe, need to ensure its database infrastructure will scale to handle its traffic increase of 300% last year. The company has 40 million registered members generating more than 6 billion page views every month. Sharding is a technique used for horizontal scaling of databases that Netlog is using.

Jurriaan Persyn, Lead Web Developer at Netlog, will be presenting frontend technologies to develop and improve the features of Netlog’s social network. If you’re interested in high performance, scalability, MySQL, PHP, caching, partitioning, Sphinx, federation or Netlog – join this webinar!

WHO:
Jurriaan Persyn, Lead Web Developer at Netlog

WHAT:
Database Sharding at Netlog with MySQL and PHP web presentation.

WHERE:
Simply access the web seminar from the comfort of your own office.

WHEN:
Wednesday, May 13th at 15:00 Central European Time (Brussels, Stockholm)
14:00 Western European Time (London)
16:00 Eastern European Time (Helsinki)
The presentation will be approximately 45 minutes long and will be followed by Q&A.

You can subscribe for this webcast here: http://www.mysql.com/news-and-events/web-seminars/display-334.html

Matti Tech database, mysql, netlog, PHP, webcast

A very impressive take on creating Benjamin Buttons face, using the most modern CGI known to man. Did you even realise the first hour of the movie was a completely computer animated face? Every action, reaction, motion and emotion of Brad Pitt’s face was interpreted by a computer, and transformed to a face which aged roughly 45 years. And it looked 100% real.

All in all, the movie A Curious Case of Benjamin Button is worth viewing, but especially so after watching that little making-off! Amazing what technology can accomplish these days …

Matti Tech cgi, motion capture

In case you missed it, the four leading characters behind popular bittorrent tracker The Pirate Bay have been convicted for “assisting in making copyright content available“.

Not to worry, they say, but what does the verdict really mean?

Well, for starters, the website & tracker of The Pirate Bay are a free service provided to the people. A means to share files over a commonly used protocol (BitTorrent). It is only that – a free service. It does not judge its content to be good, or bad. It doesn’t matter whether all files are legal, or illegal. The Pirate Bay provides the platform to exchange them, but does not encourage nor discourage sharing illegal files. A website (or software application) that uses a popular protocol.

The sentence in itself (1y prison for all defendants + a couple million in fines) isn’t important. It’s the fact they were convicted in the first place. They face jail-time because users (ab)used The Pirate Bay’s service. They used their platform to share illegal files.

This is a huge deal, as this verdict means any service provider can be found guilty for what their visitors do. Any ISP could be held (partly) responsible for content that travels over their network, whether legal or illegal. Any website owner can be held responsible for links that are placed in forums, images posted in galleries, … If this goes on – ISP’s could eventually be forced to filter all traffic, and take responsibility to prevent illegal usages. When will the war on privacy start then?

Since “assisting making copyright content available” is a insanely wide verdict, this can eventually be tracked back to software engineers who create applications that can be abused, ISPs that host website that violate copyright in any way, website administrators that happen to manage a hijacked forum, … When you go back far enough, even the backbone carriers are to blame for sending the network traffic out to other networks in the first place.

While the verdict isn’t final just yet (they appealed – d’uh), it is definately something to keep watch on. It’s a very public case with a lot of press attention, and its outcome could mean a great deal for future lawsuits concerning any kind of internet service provider, or software engineer.

Update 21/04/2009
It appears BT is among the first to block access to The Pirate Bay, and it seems several other bittorrent trackers are taking precautionary measures to protect themselves.

And what’s the deal with Google’s Torrent Search?

Matti Tech bittorrent, lawsuit, legal, tpb

Matti hosting case, lowercase, mysql

Had this strange issue a couple of days ago, where the following compile command didn’t give me the expected results (ability to use imagecreatefromjpeg() function). 

./configure --with-gd --with-jpeg-dir=/usr/lib
make
make test

The “make test” yielded the following results.

SKIP gif –> jpeg conversion test [ext/gd/tests/gif2jpg.phpt] reason: jpeg support unavailable
SKIP jpeg <–> png conversion test [ext/gd/tests/jpeg2png.phpt] reason: jpeg support unavailable
SKIP jpeg <–> gd1/gd2 conversion test [ext/gd/tests/jpg2gd.phpt] reason: jpeg support unavailable

Not exactly what I expected, because the libjpeg library existed in /usr/lib/libjpeg.so. Here’s what managed to fix it, running a “make clean” first, and changing the order of the “–with-gd” and “–with-jpeg-dir” parameters.

make clean
./configure --with-jpeg-dir=/usr/lib --with-gd
make
make test

And a make test now passes on the jpeg!

PASS gif –> jpeg conversion test [ext/gd/tests/gif2jpg.phpt]
PASS jpeg <–> png conversion test [ext/gd/tests/jpeg2png.phpt]
PASS jpeg <–> gd1/gd2 conversion test [ext/gd/tests/jpg2gd.phpt]

Tricky bugger this was …

Matti PHP --with-jpeg-dir, compile, gd, libjpeg, PHP

While everyone agrees that a clean URL is important in SEO (Search Engine Optimization), I think it’s valued a lot less than it ought to be (in other words; it’s more important than most think). Whenever you search for certain keywords, those keywords are highlighted in the title, description & url of the search result pages.

The Title & Description speak for themselves. The title is what used to get the most attention, and the description may not mean a lot in SEO-points itself (content & semantics are they way to go), but its value is mostly to get users to click from the search result page, onto your webpage. It’s your entry-path. Your pursuading argument to win a visitor.

And that’s what the URL has been graduately evolving into. Nowadays, internet users tend to ignore the webpages’ title, and look at the URL in result pages.

Take the following search result for “dedicated servers“.

While the title is displayed in a larger font than the URL, most people will look straight down at the URL to see if their keywords are present. Because after all, if the webpages’ filename contains our keywords, it’s bound to contain the info we’re looking for. Right? ;-)

It also has the advantage of getting a lot of information, in one tiny spot. The domain name usually contains the company title – which is often stuffed at the end of a webpage title, to contain more keywords. It usually indicates a language as well, so you know straight away what you’ll dive into.

Thus it’s important to change URLs such as

domain.com/index.php?module_id=4501&lang_id=501&product_id=2101

to

domain.com/en/productinfo/teddy-bear-slippers/

A clean URL, that shows what the page will contain, in a blink of an eye. 

While the Title-tag remains a great asset to get good rankings, it’s no longer the deciding factor to get users to click through. It’s time to spend some time, personalising & keyword-izing (if that’s even a word) your URLs!

Matti Seo keywords, mod_rewrite, seo, url