temporal context and privacy in blogs Abstract This paper describes the current evolution of online diaries and the convergence of these information sharing strategies towards a personal publishing habitat, and begins to address the further steps towards managing and inhabiting such an environment. It is argued that by addressing privacy concerns and reducing the cost in time of organising audience access and publishing online, many people will be keen to locate the content they post in a rich personal landscape of physical, temporal and social context. An example prototype that compliments existing weblogs with temporal information, and the methods for collecting and appending this metadata to posts using current and furture mobile technology is outlined. CHI style paper [.ps] CHI style paper [.doc] contextualising online diaries - notes and links Project Design My paper describes the motivation for adding context to blogs and for visualising temporal information. It discusses the development of diary-like blogs and moblogs and outlines the need for core management of privacy and audience access. My project design intends to effectively visualise the temporal context of weblog posts. The idea is to collect raw measurements of any avaliable personal time varying data, and offer a blogger the oportunity to present this as landscape onto which temporal posts can be located. Much of the rhythm of posts in time is lost in current blog formats which have been historically dominated by design decisions restricted by html page layout, low screen real-estate and low download bandwidth. By organising temporal environments visually, it is argued that comparison and feature evaluation become simple, obvious tasks. A prototype is proposed that presents life rhythms obtained from computer activity and mobile device use as a series of overlaid time-plots. For those willing to share indicators of activity and mood, and similar temporally varying measurements, this interface would allow the presentation of this data. Comparisons between various metrics published by a blogger can be easily drawn, and activity and 'state' can be inferred at a glance from visual clues. By re-scaling curves as a visitor browses over varying time periods, patterns in short and long term activity can be revealed by a blogger. Examples include increasing e-mail load nearing a project deadline, personal weather measurements or even galvanic skin response. Further to restricting access to certain measurements, privacy can be effectively managed by varying the maximum resolution viewable by different audiences. So, for example, someone may choose to reveal up to the minute information to a few close loved ones, daily averages to some friends and weekly approximations to the general public. Comments and pictures can then be attached to this ambient temporal context as they are uploaded. Posts will form perceived groups according to their proximity in time. Just as bloggers currently draw on others in their community for content and commentary, so comparisons can be made between activity patterns in these communities. This system also offers the opportunity to present this timeline data back to the user in a handheld device interface which could possibly trigger further narrative as they comment on measured levels. As rhythm is more important for this application than the measurement of particular events, the processing of data into final curves is focused on smoothing, averaging and curve fitting to output clean lines. The four curves shown in this example demonstrate the formatting styles available. The ribbon curve describes the variation and difference between two closely linked variables, such as incoming and answered e-mail. The flat line shows a measurement for which the current viewer has access, but cannot see in detail at this resolution. The two smaller sections of curve demonstrate examples of author editing of data. The user has chosen here to include only some periods of this measurement either as a privacy choice or to illustrate a specific point (e.g. "check out the snow fall last night!!"). The final curve shows a measurement with many variations over time. The line width is tied to the maximum and minimum variations at a point. The curve intensity is based on a measuring device's estimation of 'uncertainty' or 'error' where appropriate. The curves are designed to sit behind the site's content, and provide ambient at-a-glance information. On the right of the screen we can see the user's standard blog layout. As a reader scrolls down the page of content, the currently visible posts are indicated on the thumbnail view. Above the curves is the timeline navigation bar. The vertical strokes mark the current display width. The central toggle provides navigation of the timeline. Mousing-over the toggle shows the time period and date of the current section. Pulling the toggle to the left or right, scrolls the display backward or forward in time respectively. Pulling the toggle up or down zooms the view out or in respectively. Zooming is performed in steps (hour, day, 3 day, week, fortnight, month, etc..) rather than smoothly, with the curves redrawing to smooth small variations at each level. Mousing over a post displays it's time, date and title. Double clicking a post slides the current view region over to the correct time section, and scrolls the standard view layout to the correct place. Clicking and dragging a thumbnail enlarges the image or text as it is dragged, and smoothly slides back to a small image when released. Mousing over a curve displays a label of the data it describes and hovering scale markings. The layout is intended to be a flexible toolkit, with background, colours, and screen position all specify-able by the user. Space above or below the curves is left available for site navigation, blog-roll, other links or decoration.