If I had to sum up software development in one word, it would be “don’t”. When you know a bit about what’s possible, it is easy to be seduced by the notion that custom software could solve all your problems. Your business is unique of course, and so your requirements are unique. It seemed so easy for the geeks to code up that prototype. It is easy to forget “Soon, cheap, good — choose two”. The cost of making software robust, scalable and flexible seems to surprise us every time. It is easy to forget the nightmare projects where, with budgets and deadlines long blown, still more resources are poured in before the plug is eventually, painfully pulled. A pretty good fit with an existing system is a pretty attractive alternative. So most of the time, when custom software development is suggested, mechanical repetition of the word “don’t” is the most valuable consultancy you can offer.
Why, then did we build custom software?
In 2002, my company GroupSense provided an online learning platform for post-graduate business school Advanced Business Education Ltd (ABEL), for the third year in a row. ABEL’s programme involves self-study and small group collaboration in between a series of block-courses. In 1998, when we first met to discuss elearning, the managers at ABEL showed me their course material. It was two large ringbinders that they sent out to their students in courier bags. They asked me what I thought they should do with it. I said “send it out in courier bags” and “let’s do something online that you can’t do with courier bags”. We agreed that displacing the course material printing cost to the students would benefit no-one. What they wanted was to make it easier for their students to collaborate in between face to face meetings.
At that stage, I provided online collaboration consulting, using platforms that were already available. Most corporates had a mail system with shared folders. Where they didn’t, or where participants spanned organizational boundaries, many options were already available on the Internet. There were several mature web forum tools, although none of them supported email participation very well. I favoured email lists as they enabled participants to use the client they were already using for collaboration. The best list servers also had online archives, that supported posting and file-sharing: the best of both worlds. I’d been recommending a service called MakeList.com, until it became eGroups.
In 1999, ABEL launched 30 online groups for students using eGroups. It was a small step, that extended the services ABEL provided, and enabled ABEL to begin to learn about how it could use the new medium. By 2000, eGroups had been acquired by Yahoo!, and ABEL provided the same 30 groups plus another 150 groups, in each of two semesters, for small group collaboration. In another small step, the web interface was de-emphasised in favour of simply getting the students to use the email interface. The usage and feedback were good, so we offered Yahoo! Groups in 2001 and 2002, by now teaching the students how to get a Yahoo! ID and use the web interface of their online groups. We also provided a website for ABEL, with static content, feedback surveys, some assessment results and links to the students’ Yahoo! Groups.
There were various problems with using Yahoo! Groups, however. Because Yahoo! had integrated Yahoo! Groups with various other properties, their registration system was complicated. It was difficult to point students to their Yahoo! Groups, because they all had different urls. The groups were located amongst arbitrary other Yahoo! Groups, as well as advertising that was often not appropriate for a business school. ABEL wanted their elearning initiative to provide more of a sense of place, and to reflect their brand. Finally, administration of 330 Yahoo! Groups was manual and tedious, and there was a risk that losing access to the single administration account would shut down the whole show.
Despite the problems with Yahoo! Groups, ABEL was sold on the benefits to their students of being able to interact with their online groups using email, the web or both. They asked us to find an elearning platform that provided the functionality of Yahoo! Groups within a website that they could control. Though we looked, we could not find software that provided that. We evaluated various elearning platforms, but they were all oriented towards content and had basic web forums, at best. There were still no list servers with a good web interface. What we did find was some open source components that went most of the way to what ABEL wanted. We also found some developers who could integrate the components quickly. ABEL agreed to the open source development approach, and GroupServer was developed and in production within a couple of months.
In the years that followed, the ABEL site was expanded to support online content, as well as surveys, workshop admission forms and photo sheets, notifications of assessment results, and sophisticated reporting of results and survey responses.
In 2004, GroupServer was spotted by Steven Clift of E-Democracy.Org. E-Democracy.Org had been running online public issues forums in Minnesota and other places, since the mid-nineties often using Yahoo! Groups themselves. They were looking for an alternative that they could have more control over, and recognised GroupServer as exactly what they were looking for. They provided some resources to develop features that would make GroupServer more useful for public sites. By May 2005, they had begun to migrate all of the E-Democracy.Org forums to a GroupServer site.
The approach of incrementally developing GroupServer has worked well for OnlineGroups.Net and our customers. We have managed to keep budgets and timeframes in check. It has taken five years, but GroupServer is now scalable to hundreds of thousands of users, and it has been in production all this time.
By 2006, organizations including Landcare Research, Local Government Online, and Steven Clift’s Democracies Online were using GroupServer. Other organizations began to ask for online groups sites, so the OnlineGroups.Net beta service was launched. Now that OnlineGroups.Net is now out of beta, and GroupServer 1.0 is available to download, there is a viable alternative to Yahoo! Groups, and Google Groups, for organizations that want email groups with a web forum interface, on a dedicated customizable website, with no advertising.