This spring I have been doing some work on a web application, part of a deliverable in one of my courses (INF5270).
We were a team of four writers/programmers/designers/whatever needed. The deadline was about an hour ago, and we delivered our system 3 hours before that for once. So, for personal reflection I thought I’d sum up my experiences. Especially in the light of having read through a better part of Getting Real by 37signals.
The theme of the course is, as stated in its title, Design of Interactive Web Sites. The curriculum is somewhat broad, but the main focus is on Information Architecture and Sociability. Our task was to build a social web site (and also write up some reports about it and reflect on the theory of the course).
We decided on a theme early on; books. We were going to build a social web site about books, or literature if you want it dandy. We started investigating other sites within the theme, and found some. Looked into them, and what they did and reflected about what we would want to put in.
At this step, we only had some thoughts, and also a report that we had to deliver about our findings and how they related to the theory taught. The next step was, as Getting Real suggests, making the screens in HTML. Only, we skipped the part with sketches and drawings and drew directly on a document with ideas. I think this is probably one of the first (major) faults we made in the project.
Words in a document are fine, but do they really convey what we want? Everybody has different ideas, and that really shone trough nearing the deadline. We were not on the same note and the end result was different than, I think, everyone envisioned. If we had sketched the site early on and put it somewhere on the web it would have been much easier to see what we agreed and disagreed on and reach agreements.
So, we had the screens out there in HTML (part of another delivery), we disagreed, agreed and none of them really got updated. The same screens were the basis when we started on the 4th delivery that included the first bit of implementation. The screens where as much worth as the dead document, nothing. So we started coding, all in our own directions, which resulted in confusion and a delivery that consisted of different ideas, not ONE idea.
And, that was pretty much the story for our 5th and last delivery that was finished up some hours ago. We ended up delivering a system that deviated some, or much, from both our personal and common vision.
I think the following would have made the project execution better:
* Interface sketches EARLY on, perhaps just after deciding on our theme for the site, and then meeting up for more iterations and doing an early assessment of feasibility for the different parts of the site. Together with a vision document, and some documentation of elements in the sketch this would serve as the framework for the whole System.
* A live document where we documented our choices, so that they could be found in easily when needed (some code at night, others early in the morning).
* Some more meetings . Although Getting Real really disapproves of them I think they would work in our case (small team) to easily and quickly solve confusion on certain aspects in a quicker way than e-mail and/or IM.
* A centralized, online, meeting place. Campfire could probably have been suitable for this.
Hm, that’s what was on top of my head for now.