Volunteers make Vancouver PHP Conference work

45

Author: Bruce Byfield

More than 225 developers attended the second Vancouver PHP Conference at the University of British Columbia’s Downtown Campus in Vancouver Canada this week. Organized by the Vancouver PHP Users Group, the conference attracted many of the best-known names in the PHP world, including Rasmus Lerdorf, Andrei Zmievski, Damien Seguy, and Zak Greant. The result was a well-rounded conference that shows what an experienced group of volunteer organizers can accomplish.

Many of the out-of-town speakers arrived over the weekend and began the conference with a ski trip to Whistler and a dinner in the trendy Yaletown district on Sunday. However, the conference did not officially kick off until Monday morning. Despite the discovery that name tags had not been printed alphabetically — making for “real random access,” as one volunteer commented — registration was largely complete by 9 a.m., when the first presentation started.

A selection of presentations

Fittingly, Rasmus Lerdorf, the creator of PHP, delivered the first presentation of the conference. Lerdorf gave an unstructured but interestingly varied talk. He began by suggesting that many people contribute to open source because, like any online activity, doing so stimulates production of the same hormone as orgasm or everyday social interaction. The rest of of Lerdorf’s presentation was largely technical. It included a demonstration of how Callgraph, a tool for graphically representing the calls made by a running program, could be used to help track down bottlenecks and speed up a program, and another of Yahoo Pipes, a graphical tool for building Web applications online. Lerdorf also told the audience, “Don’t worry too much about security,” arguing that most problems in PHP programs would be buffer overruns, and require local access to exploit. Instead, he suggested, security problems were more appropriately approached on the operating system level.

Lerdorf’s talk was followed by Andrei Zmievski’s report on the efforts to introduce Unicode support into the upcoming PHP 6.0. Zmievski began by explaining that, on the simplest level, Unicode is an effort to map every character of every written language in a single system. However, he added that PHP 6.0 will also support the Common Locale Data Repository, which includes such information as the currencies associated with a language and the preferred methods of collation, or sorting. In version 6.0, the currently limited locale awareness will be replaced by a detailed declaration of Unicode locales. As Zmievski explained, this change will allow such features as multiple character sets within the same string and the ability to customize sorting options. The new system will also support such transformations as converting between full- and half-width characters in Japanese, or Korean characters to Latin ones. Currently, he said, Uniform support in version 6.0 is 60% complete, and is expected to be finished by the end of 2007.

Another well-attended presentation was Jeff Barr’s “Web Scale Computing,” an overview of Amazon.com’s online services. Essentially, Web scale services are pay-as-you-go computing, available at what Barr describes as “massively competitive rates of pennies per megabytes for storing or transferring data.” These rates are calculated to be an appealing alternative to companies that might otherwise set up their own server farms for their intermittent needs. After outlining the services offered by Amazon, Barr demonstrated some of the tools that already exist to manage his company’s Web services from the desktop.

On Tuesday morning, programming began with Brian Aker’s untitled talk. Aker, who is director of architecture for MySQL, spoke about the strategic problem that currently preoccupies him: how to scale databases for major Web 2.0 such as Amazon.com, MySpace, and Flickr. According to Aker, these problems include data caching, splitting applications over multiple partitions, running events within databases, and routing — all problems that are perennial with databases, but that become especially acute on the scales faced by such sites. Problems unique to such sites include finding power for their server farms and managing the waste heat that they give off, and geographical replication. Aker explained the range of solutions for the perennial problems, but admitted that the problems unique to the giant sites remain mostly unsolved.

Most of the tracks over the two days were heavily technical and focused on PHP. Topics varied from Damien Seguy’s “PHP Tips and Tricks” and Marcus Böerger’s “Standard PHP Library Updated” to Rick James’ “Microsoft Tools and Platforms for PHP” and Kevin Schroeder’s “Caffeinated PHP: Using Java to Extend PHP.”

However, the conference also branched into other topics. Zak Greant gave a heavily attended talk on “Copyright, Contracts, and Licenses” in which he explained the basics of these concepts and what programmers needed to know about them. Perrick Penet delivered a talk on “Open Source in Europe” and Andre Zmievski gave a presentation on “VIM for Programmers” based on the premise that developers do not really need feature-rich IDEs for their work.

In addition, each day ended with a series of five-minute Lightning Talks, for which attendees could volunteer on-site. Some chose to give a Lightning Talk to show a PHP trick they had learned, while others used their time to promote their company, project, or conference. The format proved an effective way to wind down a day of presentations.

On Monday night, the conference threw its doors open to the public to present “Open Source in Vancouver,” a panel discussion with Brian Aker, Zak Greant, David Ascher, Perrick Penet, Greg Dean, and me. Realizing the broadness of the topic, we quickly threw the discussion open to questions from the audience. Ignoring the title, the discussion began with a debate about whether open source was stagnating or becoming more popular. However, the subject of patents quickly arose, and dominated most of the two hours allotted for the panel. As a presenter, I found the interactiveness of the discussion a perfect end to a long day, with as much insight given by the audience as by any of us on the panel.

The advantages of volunteer organizers

Compared to other conferences, the second Vancouver PHP Conference unfolded with unobtrusive efficiency. The credit rests entirely with the volunteer organizers. Several of them had participated in the first Vancouver PHP Conference in 2004, so they knew what to anticipate. The 15-minute break between presentations was a welcome acknowledgement of the fact that humans and schedules rarely mix well, and the organizers’ quickness to troubleshoot the small problems that inevitably arose, and to keep everyone informed, ensured a generally smooth event.

The conference was a small event, with the majority of attendees being local, but, to its credit, it managed to attract many of the leaders in the field. Although these two facts might seem contradictory, Peter Gordon, one of the main organizers, suggests that they were actually closely related. A small group organized by volunteers, he explains, was precisely the type of audience that may of the keynote speakers prefer.

“They see user groups and conferences by volunteers as promoting the work they do,” Gordon explains. “So they’ll come up [to Vancouver] for airfare and a hotel room. They don’t ask a fee or anything like that. Putting it together on the volunteer level means we get a whole lot more support from just about everyone else who gets involved in the event.”

The formula seems to work. The conference was filled to capacity for its venue, and Gordon tells me that the organizers are already looking forward to another one next year.

Bruce Byfield is a computer journalist who writes regularly for NewsForge, Linux.com, and IT Manager’s Journal.

Category:

  • PHP