What shall we talk about?

Recently, our founder and CTO Brian Behlendorf decided to step back a bit. He is still around but turned over the CTO cubicle to me. Quite coincidentally, but pretty much at the same time, CollabNet purchased SourceForge Enterprise Edition from VA Software, bringing the world’s two premier commercial web-hosted software collaboration environments under a single roof. If all that doesn’t call for a little blogging, I don’t quite know what does!

So, what shall we talk about? The past? Next steps? New beginnings? How about "all of the above"?

By this point in time, I think no one can doubt that the open source software development community has accomplished some amazing things, things in some cases that the enterprise community has attempted repeatedly and failed at. Such singular achievements come to mind as sendmail and BIND, the rolling collective stewardship of the UNIX family of operating systems, the famous "LAMP stack," or any number of other dramatic achievements. But at the same time, there’s no question but that the open source process has some liabilities that are really worrisome to the enterprise — you know the list: imprecise schedules, false-start and abandoned projects, and a tendency to ignore less "sexy" details like quality, localization, and customer support. The best open-source projects can equal or better the best enterprises in all these matters, but then there are all those distressing not-so-best projects.

There’s some kind of magic in the open source community process, but it needs some taming. CollabNet was founded (thanks, Brian!) to capture the magic of the open source process, civilize it just a bit, and bring it to the enterprise. They (before my time) provided a set of tools proven in the open-source arena, but didn’t hesitate to help the community produce better replacements where necessary, such as by sponsorship of Subversion development to replace CVS: the magic’s not in the tools, but in the processes that co-evolved with them. CollabNet added features like rich security and authorization systems, that are crucial to enterprise use but uninteresting (if not down-right antithetical) to the open-source community. Brian also brought with him, from experiences in the Apache community and elsewhere, both wisdom and contacts at the heart of the communitarian open-source magic.

Since I joined the company in 2002, we have added world-class hosting and support services, scalability, enterprise-focused features like project management and reporting, and have opened the product to partner and customer integrations through APIs. By this time, as well, surely no one has missed the significance of SourceForge. Through its public face sourceforge.net (as well as its several other properties), VA Software has served and fostered the open-source community immeasurably. Through their commercial offering, SourceForge Enterprise Edition, they’ve distilled that open-source magic into their own commercially successful brew. But VA Software found their business growing in other directions as well, into online media, and have discovered a need to focus their energies there.

Transferring SFEE to a more suitable home became very attractive. CollabNet is a great place for SFEE, and we’re delighted to bring over this complete team, well experienced in our core field, and their solid and innovative product. So where do we go from here? Hey, give me a while to work out the details of a combined product roadmap; I’ll get back to you on that. But I’m "new" around here, and we’re just getting started!

Tagged with: , , , , , , , ,
Posted in Subversion
One comment on “What shall we talk about?
  1. Brian Behlendorf says:

    Hi Jack! Great post, and you described the origins of the company very well.
    As you work out the details of a combined roadmap, there are no doubt going to be decisions to make that affect end-users of both CEE and SFEE. Realistically, those decisions will likely mean disruptive changes for users of one of those systems, perhaps for both. Disruptions should never be done for a frivolous reason, but sometimes there’s a good reason. Such changes are a great time to get rid of misfeatures that few people use, to clean up an architectural mistake, or to bring greater consistency between features.
    The only way to measure the magnitude of disruption from any proposed change is to measure how customers use the existing features, combining quantitative data (looking at traffic logs, or usability testing) with qualitative data (asking users what they use, anecdotally). Historically CollabNet has done very well with our upgrades, handling disruptive changes by doing both of the above, keeping in close conversation with key customers since there was a manageable number of them. Now, with SFEE, there are many more customers, and many more on-premises installations, so it’s harder to consult server logs or have that direct relationship with how customers are using the site.
    So, the biggest answer I can think of to “what should we talk about” here is: let’s tease out some of the decisions that need making, to the degree that they are about features and product strengths, and talk about them. I’d then hope to see customers replying – both kinds of customers, the ones that write the checks and the 1,100,000 users. Let’s be clear that it’s not a plebiscite – even if the majority of feedback went in one direction, a different direction might need to be taken, and there’s other channels of feedback that CollabNet will have to listen to. The cool thing here is that it’s not limited to ‘feedback’, but conversations that can lead to unexpected answers.
    Interesting times. 🙂
    Brian

Leave a Reply

Your email address will not be published. Required fields are marked *

*