We have done a lot of work over the last weeks. Not only did we make a lot of features work better, we also implemented the promised “OSI certified” indicator that documents whether the used license of a project is following the defined standards by OSI. There might still be some small errors in terms of the allocation of this indicators, but we keep working making it right. We have also upgraded the forum component and redesigned the front page of the directory. Small steps, but hopefully positive for the users. Traffic is increasing constantly and the popularity hitlist is changing – Alfresco made it to second, KnowledgeTree entered the top five. A recent Podcast done by Stephen Walli features the EOS Directory and gives insights from the maker, the community and the analyst perspective.
Back in February of this year, Nat Torkington asked “Is ‘Open Source’ Now Completely Meaningless?“Obviously we don’t think so, having just launched this directory, but from the beginning, one of the challenges has been just what definition of Open Source we should use in determining which projects to include.Dana Blankenhorn posted recently on ZDNet (“Optaros EOS will take the licensing question seriously“) talking about some of the issues and our plans for handling them. (For more on the subject see Michael Tiemann’s “Will the Real Open Source CRM Please Stand Up, ” and Matt Aslett’s “Centric CRM and SocialText respond to open source hard line,” as well as Blankenhorn’s earlier posts on the subject)The simplest criteria we could use to determine whether to include a certain project is whether the software is available under an OSI Approved License. If the license being used is not currently OSI approved (as is true of, for example, with the Affero GPL), we could try to determine whether the license would be approved if submitted, by testing the license against the open source definition. (This approach would run the risk of marking as “unapprovable” licenses that OSI might approve, or of marking licenses as “approvable” that OSI might reject.)However, there are projects out there which are widely used in enterprises, and which describe themselves as open source. For example, projects which require the prominent display of a logo in any derived work (often called “badgeware” or “logoware”), or projects which use an Affero GPL like clause to close what they see as an ASP or services “loophole. These licenses have not met with OSI approval, and yet many would consider such projects open source and wonder why they aren’t listed.Our approach has been, and will continue to be, to provide as much clarity as possible about the licenses under which various projects are available.As my colleague Dave Gynn put in his discussion with Blankenhorn, our plan is to positively identify projects whose licenses are OSI approved, and provide a basic overview of what that means and why we think it is important.This way, projects whose licenses are not OSI approved will still be included in the directory with what license they are released under clearly identified.We’re also trying to find a fair but clear way to identify and present to users what prevents any particular license from being OSI approved. This can be complicated, however, since we do not want to position ourselves as speaking for the OSI (it is not up to Optaros to determine what is or is not OSI approved licensing) or for the project (presenting the projects’ own rationale for its license does not mean endorsing it).I’m also interested in what the community of users here has to say. What guidelines would you like to see applied to projects being considered for inclusion in the EOS Directory?
When we started to develop our Open Source Directory we were taking an Enterprise perspective from start on. We wanted to develop a tool that helps CIOs and IT decision makers to easily find and pre-select open source technology. For this purpose we defined our rating system. We went one step further with the online version and opened the platform for EOS users to also enter their rating. In an ideal world this “user rating” would follow the same rules and guidelines as the “Optaros rating”. Both the “Optaros rating” and the “user rating” are trying to assess the “enterprise readiness”. We defined 5 levels of “enterprise readiness”:4 stars:Product/projects matches or is superior to best – proprietary/closed source – available and widely used products. It has proven to be a standard in a specific category that you can’t pass.3 stars:Product/project is mature, fulfills the important requirements and is supported well, it is enterprise ready and conforms to typically found needs/requirements in enterprise production environments2 stars:The right tool for the many situations, more investigations needed, not top league yet. Weaknesses may be driven by technology or lack of functionality or support. A proof of concept is recommended.1 star:Not recommended (yet) to be used immediately in a broad and enterprise wide context, other than early proofs of concept or deployment in a controlled way (e.g. as component that understood by the developer)0 star:Not recommended to be used in an enterprise, probably not worth further investigation. These technologies are not shown in the directory at all.When looking at the “user ratings” people often seem to have only one objective, bumping up the technology to 4 stars. But it takes a lot to reach the 4-star-level in reality. With this we would like to ask people doing ratings to stick to the criteria defined and make a reasonable judgement.
Only few days old, we have already received a lot of feedback from the EOS user community. We are happy that people are pleased with the site and find it quite usable. Aside of generic comments and feedbacks we also receive many proposals for projects and technologies to be added. We plan to implement a “propose project” functionality for this but for the time being email@example.com is the easiest way to communicate missing projects. We have a fairly rigid process to check candidates and we have right now quite a long list of candidates to be added. Let me reiterate what it takes in our eyes to be included in the catalogue:
- The technology is “open source”
- The technology can be considered enterprise ready (see criteria), we don’t list products that do not achieve at least one star
- The technology fits into one of the categories we cover in EOS
- The technology clearly is well represented within enterprises across the world
We are aware that there are probably quite a number of projects out there that fit these threshold criterias and are not listed (yet). But, well, that’s the reason we brought it online. We want the input and feedback of the enterprise users out there on the web. Thank you for contributing!
Not only did we launch our EOS at 10:02 CET today, we also experienced already significant peak loads and had to fix a couple of things to make the site faster. We apologize for any inconveniences! We already gathered quite a lot of feedback from our first day users and will make sure we can incorporate many of the ideas as soon as possible. With postings on Heise Online and Heise Open we attracted many interested people from Germany. Matt Asay blogged on EOS as well and we hope many more will cover it.
Today is the day! We launch our BETA version of EOS. With 285 projects of which more than 80% have been updated during the last weeks of reviews, discussion, benchmarking and more discussions. It’s not a surprise to see that 3/4 of the projects have actually progressed and their rating has improved. However there’s also another quarter of projects that we are less convinced about compared to our last big review cycle at the end of 2006. And then there are more than 30 projects that were added to our directory. These are projects that people made us aware off after the publication of our PDF Open Source Catalogue and others that we used ourselves in projects during the last months. Included in the 285 projects are a significant number that we think are a benchmark not only for other open source products but also for the commercial software world. This includes projects such as the Apache web server, FreeRadius, Firefox, Gnu gcc, Hibernate, ISC Bind, JBoss AS, PHP, Python, RedHat Linux, Spring, SuSE Linux, Tomcat or vim. We are sure not everybody will agree with our “Optaros Enterprise Readiness Rating”, but that’s the reason why we think taking EOS online is such an important step. People can both add their own “Enterprise Readiness Rating” as a “User Rating” as well as adding their comments into the forum. This will help us to improve the quality and objectivity of our ratings and at the end make the Directory even better for the user.Over the coming weeks and months we will use this blog to announce news, be it rating changes, entry candidates and new additions. But we also will talk about support models, commercial open source and open source licenses.We hope our users will find EOS a valuable source of information and that our hopes in making collective intelligence add value to this directory will substantiate.But not only the directory entries will change over time, we also want to continuously improve the EOS application and add more features to it to make it even more useful. “Perpetual Beta” we call it.Stay tuned, there’s more to come.