This is my second update on my Summer of Code project. In these weeks, I’ve been pretty busy because of Univ. exams, but I’ve hacked as much as possible. Before I’ll demonstrate my prototyping progress, let’s first consider some picks of our current design decisions. Note, that I’m not alone! :-) … quite some people commented on early drafts of a design document and some even provided designs. Big Thanks! … to all of you … more detailed acknowledgements will follow …
I’ll cover some of the rationales for the most important elements of the Journal, in no particular order, starting out with the most obvious one.
The central presentation metaphor of the main Journal view remains a timeline, just as holds for many of the systems, which share some of the goals with my Journal revamp (such as Forget-me-not, Lifestreams, LifeLines, Milestones / Stuff I’ve Seen, PostHistory, the OLPC Journal as well as the ReflAction Journal).
Continuous scrolling of a timeline appears to be the approach to local exploration of a temporal environment, which is most easy to understand and to handle. So, as compared to the ‘old GAJ’, it was decided to move to a continuous timeline with continuous scrolling. Horizontal and vertical timeline layouts have been discussed and designed within the past weeks and I started out with the vertical version because, along with a vertical scrollbar, it’s pretty much known to everyone, as the majority of us is still used to something A4 portrait, right? However, there will be choice, one day.
“The Mind so easily gets Bored!”: Uniform Boxes in Columns vs. Variably-Sized Activity Bubbles on a Timeline … “and the Winner is?”
As already indicated above, the ‘timeline plus columns’ layout of the old GAJ was abandoned, also because much regularity in layout rather suits visual comparison of the boxes’ contents than it is supportive to re-finding and recognizing specific activities and therein involved information and people. In our opinion, an ‘overly regular’ layout prevents visually reflecting, for example, both clusters of information touched and particular activity distributions occurred. These observations, which made us dismiss structures such as lists and homogeneous thumb-grids as well, are in line with what we refer to as Federico’s ‘irregular grid’ note.
The need to adhere to a ‘too strictly predefined structure’ simply provides too little room and flexibility for adapting the boxes to the contained different information types and relative representation forms. Besides that, the rather small boxes didn’t really allow for adding visually effective reminding support.
Our assumption that the chosen design move to ‘activity bubbles’ on a timeline (as presented below) can provide more overview, than what was achievable with the old GAJ user interface, remains to be proven correct. By contrast, enabling more condense search-result presentation by the new Journal user interface seems evident to us.
The appearance of the activity bubbles is on purpose similar to the design of the GNOME popups: a triangle for visual anchorage plus rounded corners. All nice.
Regarding the two-column layout which contains the activity bubbles line-up, we consulted a numerologist :-) Additionally, we opened our ears to rumors about the design rationales, which led to the two-column layout of most ACM conference submission templates: while this template might not be perfectly supportive to the production of easy to read pages, it certainly allows for a condense presentation of easy to scan and overview content.
Logarithmic timelines with their focus on the more recent periods of time were invented in the 1930’s and are today widely used for historical overviews and scientific diagrams: a logarithmic scale ensures increasing resolution when approaching the present time, which is useful in cases, where recent times hold the most interest.
The advantages of a ‘logarithmic timebar navigation widget’, complementing the Journal’s timeline are obvious, considering that the information touched today, within the last week, and within the last month, respectively, was shown to account for 5%, ~20%, and ~50% of a user’s overall information re-use. In fact, even without awareness of these numbers, some designs of the old GAJ (soon after the project had started) favored a similarly distorted timebar navigation widget; but this was only until the decision to merge a histogram navigation widget (for jumping to specific days within the Journal) with visual indications of the amounts of activity per day. So, now, at the occasion of the Journal revamp, these ideas surfaced again, and I decided to give experimenting with them a try. In general, I am not really convinced of distorted information visualizations, but I like this distorted widget thing.
So far, my work mainly focused on reaching with the new Journal about the same feature set as the old GAJ provides. This involved fixing a lot of nasty bugs and the Journal is day-by-day becoming more smooth. I didn’t surface the implemented support for searching and filtering, yet. Another thing on my TODO list is continue prototyping of the ‘detail view’ of activity bubbles as well as of the transitioning back and forth between the Journal’s main view with the bubbles and the expanded detail view of a single bubble, which will allow to focus on its content.
Finally, to wrap-up of this post, I’ll give you a screenie for take-away.