SourceForge exec on the open source breeding ground
- 23 November, 2007 08:04
We hear from SourceForge and SourceForge.net Marketplace Vice President Mike Rudolph about the world's largest Open Source software development web site, which, as of today, claims 162,855 registered projects and 1,732,552 registered users.
In Feb last year, SourceForge (then VA Software), which owns SourceForge.net, reported its first profitable quarter. Do you expect it is all up from here?
Our network of web sites serves an average of 32 million unique visitors per month. Our full 2007 fiscal year was strong. We achieved our second consecutive year of profitability, putting up 35 percent revenue growth year over year. We're quite pleased with these results.
How many new projects, on average, are added to SourceForge.net a day?
Approximately 75 - 100 per day.
SourceForge.net competes with RubyForge, Tigris.org, BountySource, BerliOS, JavaForge and GNU Savannah. How does it differentiate itself from these other services?
It's not that we all compete, the reality is: each of these sites are there to help developers. Some of these that you mention are focused on narrow corners of the market. We're different in that we tend to be the custodian of open source development. We're the largest. We've been around the longest. It's a proven and trusted development platform.
We're very broad plus quite international, with reach into almost every country on the planet. US-based projects only account for fewer than 20 percent of the projects on the site. We think this is a real strength as international participation is a fundamental strength of OSS. SourceForge.net reflects that.
SourceForge's Ross Turk said that the site's staff was 'beefed up from about 5 to 30 early last year. How many staff does it now employ?
About 100 employees.
After Feb 2006, SourceForge replaced a lot of its infrastructure. What changes were made?
We're always working on making the site better. Due to security concerns, I don't really want to comment on too many specifics about our infrastructure. We've been doing a lot of work modernizing our CVS - Subversion - and our main web site components, in addition to major updates to our central database infrastructure. This is to insure a fast, scalable environment for SourceForge.net projects.
What are the storage and bandwidth requirements for keeping SourceForge running?
Obviously we've got some very hefty storage and bandwidth requirements. During the month of September this year, we served almost 63 million files for a total of over 647 terabytes through our worldwide mirror network. When new mirrors petition to be included in our rotation, we require a minimum of 3TB of storage and 100Mbps of dedicated bandwidth.
What has been the most significant recent improvement to the site?
SourceForge.net has really seen regular improvement over the past several years, and this year was particularly active. There have been quite a few new features added including: code search, wikis, a community hub, improved navigation and, of course, the soon-to-be-announced SourceForge.net Marketplace.
What differences has it made?
I believe that it's a consensus within the community that SourceForge.net is bigger, faster, and just plain better than ever.
Earlier this year SourceForge embedded Krugle's search mechanism, which is designed to specifically search for code, into its site to make for easier developers searching for similar projects. Has this been successful?
Adding search code search functionalities was a response to user requests, and I think it's been quite successful. It's quite a job to have to look for specific snippets of code within a certain language or set of documentation. General text search doesn't always get the right answers, due to syntax or embedded loops or other things specific to code. The easier and more efficient the search mechanism on the site becomes, the more efficient the open source developers can be.
Is this SourceForge/Krugle partnership indicative of a broader phenomenon in search, where it is getting increasingly specialised in order to deal with the growing amount of information online?
Yes, absolutely. There is so much data out there and search is the natural entry point. Generalized search misses a lot of what's important or ranks it poorly. Krugle is one great example of focusing on a core type of data and providing extremely useful results. I think this trend will continue into the future.
Have you seen a change in the type of projects hosted on SourceForge?
We are seeing more commercial open source applications, absolutely. It's been quite noticeable. Part of the reason is definitely that open source is now being more widely adopted in enterprise environments. This is a major movement within open source software.
Zimbra, which was acquired by Yahoo for US$350 million and JBoss now owned by RedHat both started out on SourceForge. What are some other examples of SourceForge hosted projects that have made it into the big league, and how many of these "success" stories have you seen?
It is a point of pride for us that a lot of alumni have done very well for themselves as commercial companies built around open source products that originated on SourceForge. We think this is a fabulous model of development and are proud to play a part. Others include SugarCRM, Zenoss, JasperSoft, Hyperic, Alfresco, SourceFire, dimdim, OpenNMS and more.
What drives SourceForge's success?
We exist to enable the community to be successful, whether that is developing, distributing, or using OSS. Our position as a platform allows the OSS community to interact and be successful, and we stay out of the way as much as possible. This may be what makes us popular.
Will SourceForge.net utilise Web.2.0?
We've added wiki capabilities this year and we'll continue to improve functionality of the site over time based on user demand.