Outlook organizing affordances: An analysis

If you flag, you have search for your flags Current PIMs use a contextual flagging model where flagged items stay where they are, but are simply called out or highlighted by some kind of visual flag. Contextual flagging has certain advantages

1. Flagging doesn't change the user's data. It is simply an overlay of visual cues. Items stay where they are and users don't need to learn some new reorganization of their data everytime they flag something.
2. Similar to real world highlighting

However whatever advantages contextual flagging has, it has some significant disadvantages that render it useless. Contextual flagging works well when the user can view the entire "context" on one screen. However, screen real estate being what it is these days, this is impossible for anyone with more than 20 pieces of information in their PIM. As a result, because users can only ever see a tiny fraction of the entire landscape of their personal information at any one time, contextually flagged items are more often than not, out of view, forcing the user to "hunt" down their flagged items.

Text-based displays are in permanently zoomed in mode or absolute row height mode. Because we are constrained by our eyesight, words are unintelligible below a certain font size. As a result, no matter how large your personal information landscape grows, no matter how many hundreds, thousands of emails, appointments, tasks and contacts you collect, the row height of any one item in your PIM remains the same.

Compare this to the way that graphical displays (ie maps) work. USA, Belgium, Modesto, CA can all be viewed in a 600x450 pixel window. All 3 take up the same screen real estate, but each one has a different scale where 1 pixel might represent 300, 30 or 3 miles of the "real world".

• map.gif:
  

If your map were a PIM it would be stuck in the "city block" zoom level. You could never zoom out to pick a new location to zoom into. You could only inch along clicking on N, W, E, or S, to find new places.

Maybe one day we will find a way to make text-based displays intelligible in variable zoom, proportional row height mode just like graphical displays. It would require intelligent, meaningful translations of textual information into graphical visual cues. For now, we are stuck permanently zoomed into Another way to look at is that contextual flagging, like the file system metaphor is a legacy of the physical paper world where physical things can't exist in more than one place at the same time. If you highlight a passage in a textbook, it won't magically deposit itself again into a section in the back of the book called "highlighted passage". Either way you look at it, the flagging user experience is unsatisfactory.

The contextual flag model works best in graphical displays with variable zoom. The model is broken in the permanently zoomed in world of text-based displays where the user can only see a portion of their information landscape at any one time and has to remember to look for flagged items. That's backwards. Flags are supposed to remind you.

Way people hunt for flags:

1. Sort by the flag column
2. Scroll through your view to look for flagged items
3. If you have flagged items in more than one view, you have to a Remember in which views you have flagged items b Periodically click through these views in search of flagged items_

Sarcastic sidenote Of course users could create a rule-based view that automatically agglomerates all flagged items into one place. But if this is such an obvious thing to do, why isn't it already done for the user OOTB? Even so, this is an unsatisfactory solution. It is still an extra view to manage. It is still a view the user has to remember to visit.

If you folder, you have to think too hard and you have to work too hard Flagging being what it is, a lot of people fall back on Outlook's original organizational affordance taken from the file system OS paradigm: foldering.

4 ways to say the same thing Foldering assumes that people process and organize in one fell swoop. Foldering fails to distinguish between processing and organizing. Foldering doesn't assume that users actually pass over the same item multiple times before they can call it fully processed and layed to rest. Foldering fails to address users' need to distinguish between

1. unprocessed items
2. half processed items
3. fully processed items
4. fully organized items and
5. items that don't need to be organized at all

In the foldering world, all folders are created equal. There is no distinction between coarse-grain processing and fine-grain topical filing / organizing. As a result users can never just "store an item as reference". If they want the item out of their face (that is to say, out of the Inbox), a relatively easy to make, coarse-grain, processing decision, they must at the same time make some relatively hard, fine-grain, organizing decisions: Which folder does it go in? AND perform some fine-motor mouse acrobatics to implement the decision. In other words, you can't ever just decide to take an item OUT of the Inbox. You must also have some very precise ideas as to where TO put the item.

Furthermore, deciding where TO put an item assumes that users have managed to successfully create a taxonomy of folders with which to organize their items. But taxonomies are hard to build and most people probably end up with less than ideal taxonomies, conceptually unergonomic folder organizations that that further exacerbate already difficult filing decisions.

Folders are simply the wrong affordance. They launch users off into the difficult and seldom rewarding world of taxonomical complexities when most users just want a simple way to perform coarse-grain triage.

As a result, current PIMS users cope by

1. Don't bother. Don't process. Don't organize. Just use search. Leave everything in the Inbox which can lead to information overload. There are no distinctions between items other than unread and read. Items in various stages of digestion are strewn throughout the Inbox perhaps mysteriously "flagged" for mysterious reasons that the user no longer remembers.
2. Spend your entire life filing everything away neatly into folders. Cry when you go on vacation and come back and there are 500 emails in you Inbox. Click furiously through various folders to search, sort, scan for items.

Sarcastic sidenote Of course, some users have figured out how to coerce the foldering metaphor into something more usable...But "something more usable" should be Outlook's OOTB configuration. Coming up with "something more usable" should be Outlook's job.

Last resort, rule-based views: Today view Bold, daring and committed technology enthusiasts or simple someone with nothing better to do might try to experiment with Outlook's custom rule-based views. Which translates to approximately: not very many people.

Custom, rule-based views like the Today view is another example of how Outlook is simply a feature list, not a design. Let's have a feature that allows people to create views that are automatically populated by rules. Let us count the ways in which this feature is flawed:

1. Today = useful. We all have tasks that were created days, weeks, months ago that are still unfinished, that are still flagged, that are probably buried somewhere in our PIM out of sight, out of mind. What users need is a way to see all things that "impinge" on me Today. NOT all things created / received / due by today. What users need is a hybrid between a rule-populated view and an explicitly created view.
2. Even if this simplistic view on views were somehow useful, the Today view is tacked onto Outlook, yet another view to manage in addition to my Inbox, Calendar, Taskpad, and anywhere between 10 and 50 folders I've created.

Where does this leave Chandler? NaturalWorkflowDesign

-- MimiYin - 12 May 2004