This is related to another post, but this post is more of a design and development thread, and the other post seems to be asking about off the shelf solutions.
My first experience deploying alfresco's groupware solution, albeit in a kind of backhanded way, failed miserably and is documented in this lessons learned document. That's OK, because I think I learned a bit about my audience and how to increase adoption.
My audience is comfortable with email. Not forums, not blogs, not wikis. The document library is about the only thing they used, but they have some "trust" issues which need to be addressed before they will willingly use that.
So in short, I now view email as the bridge to adoption. Now I am pondering ways to make alfresco's email interface more useful to my audience, and I'm open to suggestions. Initially, I was thinking to extend the "email extension" with RFC2822 fields ("Message-Id", "References", and "In-Reply-To"), the "List-Id" field, and provide some means for Alfresco to store an email list archive, providing a nice interface like MarkMail. All that would be necessary would be to subscribe Alfresco to the email list, and then you could use whatever email software you want. Alfresco's role would be to archive the email discussion list in the share site alongside the rest of the project information.
However, googling around this morning for better ideas, I found one. In a somewhat dated report titled Internet Groupware for Scientific Collaboration, I came across two concepts I really like. The first is that the model for "Collaborative Discussions" should be distinct from the view or implementation technology (e.g., email discussion list, usenet NNTP, forum software, etc.) The second concept I really like is the notion of a "nosy list" implemented by the "Roundup" issue tracker (see the Roundup documentation). This effectively allows focused email discussion lists to be dynamically created by end users without external help from a system administrator. The lists will be short lived, because they die out when people stop discussing the central topic.
As a matter of fact, the only thing I don't like about the nosy list concept is that the system creates accounts for email addresses it doesn't recognize, and I'd want it just to manage those addresses as addresses.
My purpose here is to spark some discussion about implementing a "nosy list" in Alfresco which serves as a gateway to the "forum", "topic", and "post" data types. The forums/topics/posts would have to be exposed in the share sites to be useful. In my opinion, there is nothing else which has the same potential to increase adoption.
Wow. So the report above was written in 2000, and 9.5 years later, the author offers this advice in a blog post (note that the topic is how to share equations on the web, but is easily generalizable to other applications):
Why didn’t I see, then, that the crux of the issue wasn’t XML and MathML and SVG, but rather the ability to “integrate directly into the shared spaces of the Web”? And that what ought to be integrated directly was the typesetting language already familiar to mathematicians, namely LaTeX?
The answer is that I needed (and still need) to be reminded that good-enough solutions here now, and familiar to people, often trump great solutions that aren’t here and wouldn’t be familiar if they were.
The moral: email, being familiar and good-enough, trumps less familiar interfaces.
