We recently spoke to Nicolas Vandenberghe, founder of Iterating.com, a Wiki-based software directory covering open source, commercial, and hosted software. The site provides ratings, articles, and detailed product feature comparisons.
DDJ: What was the genesis of Iterating.com?
NV: My concern was the big wave of new open-source software. I was thinking that there must be a big directory of all software available. I found several directories, but there was nothing comprehensive and nothing that included user ratings and reviews like you see in the consumer space. One of my startups was in the B2C eCommerce space, a universal shopping cart, so I had some expertise in this, but I realized if we used a standard approach it would not be comprehensive, because the software world is so big and fast moving. It hit me to make it wiki-based and engage people who have a product or an interest in an area. That was the "A-ha" moment, and I've been working on it for about 18 months.
All of the open-source solutions to do wiki-based sites were not appropriate. You have articles and categories, but the page is unstructured. But software is described with many attributes, such as operating systems, license, vendor or organization. What we had to build a "striki," a structured wiki. Structured data is usually done through databases, but after talking to several people, the most elegant, forward-looking approach was to structure the information with the Semantic Web and RDF. We looked at merging existing different vocabularies to describe authors and to cover reviews, and then we added on our own vocabularies. We wrote the site in Java using a tool from HP Labs called Jena, a tool to enter RDF statements. Jena stores them in a database and outputs them when you need them.
DDJ: So what are you announcing today on the site?
NV: We announced recently a Comparison Matrix feature, which allows you to compare several tools feature by feature. And it's editable, so if someone knows something has changed in the feature set of a product, users can edit the data. That's why we call it "Iterating," because the issue is usability; everything is editable and can be changed, but it's still displayed and understandable.
We currently list over 17,000 products. The site has moderators who add products, but we have imported many from other sites — 2200 from Debian, 4000 from Sourceforge, and some from Freshmeat. We have set up a set of web services to extract information from different sources. In an ideal world, we'd all have the same description of our data using RDF, not that many people have adopted the Semantic Web standards. So we have vocabulary mapping tables to map their description into our RDF description.
I think we are the first site covering not only open-source and commercial software, but also hosted software, things can you use online, So Google spreadsheet will be compared to Microsoft's Excel spreadsheet. You have one place where you can compare them in the same place and read what users say about what are real alternatives and what are not yet real alternatives. You can say, "I'm not going to bother installing Wordpress, if just going to Wordpress.com is good enough."
We're still growing, right now about five to ten products are added a day and new reviews each day. Every month we're about doubling. That's good, but the software is such a huge space and we aim to get to hundreds of 100,000s but if it keeps doubling the math should work.
Just like with other wikis, it's meant to be self policing, but we also have moderators. If a review is fair we won't removed, if it is well commented and well explained. Our goal is to be useful and comprehensive. Someone with a new product can list it on our site. We'll give a chance to anybody. Anyone can rate the products. Overall we don't have users putting in junk, but if a rating for an unknown product pops up to the top, and it's not objective enough, we not remove the product, but we'll remove the review.
DDJ: Have you received feedback from software vendors?
NV: We plan to talk to IBM, and we've talked to Microsoft and Oracle, and both have expressed interest in participating. Oracle's products are already listed. Everybody so far has said "yes. I'd like my products listed." You might think that commercial vendors don't want to be compared with open source. But the reality is it's better to be listed than not listed.
DDJ: Do you rely on any outside domain experts?
NV: We engage experts, mostly focused on current the upper categories: operating systems, software-engineering tools, web-design tools. For these classes, we've contacted people online and at conferences to enlist their help. This is important for the comparison matrix, we could not come up with every feature and category. We have experts work with us to start the process, or ask them the best places to get a list of features for a product.
We also have a couple of advisors, with Larry Augustin, the founder of Sourceforge, and former Microsoft Vice President Brian Roberts. They are both investors in Iteration, as well.
DDJ:What are the future plans for the site?
NV: Our first goal is to do a really good job for the top of the tree, the classes that we've selected. We want to engage the community and keep growing the product base. The big next step would be to develop a similar approach for listing service providers—listing people who are expert in a particular product. We'll use the Semantic Web approach on the attributes, then you can sort them by location, size, expertise, client references, and so on. Coverage will go from your one-man shop doing web design to IBM Global Services. We'll describe their expertise, the products they're expert in, and they'll get rated like everybody else. Clients can come back and rate the service. It will be a comprehensive listing. With the wiki approach, people could add their page and edit it. That will probably be coming by the end of the year.
