This is how you write down Matjaž in English

Still better than mad jazz. How was I supposed to find the taxi driver that was ordered online (with my name written correctly) :)?

My juice is best before 2076

I could leave the juice for my grandchildren to try :).

There's no other date around the box


This is probably a type and I understand that there might be severe consequences following this mistake.

"Information about me" and "information for me" - what is personal in PIM

There is a nice blog post by David Karger about information "about me" and information "for me".

"About me" is information about us and we might have control over it (e.g. our Facebook, Twitter streams) or might not (e.g. information other people post about us on-line or in public space).
"For me" is information that we can manage, it fulfils our needs and it might be about us (eg. appointments in our calendars) or not (my mp3 files in our music library).

I just had a discussion on this issue with Alan when I remembered this post. David pretty much covers it all, but if focusing on the management, several instances between "information about me" and "information for me" in the social context come up. With the social context I mean sharing personal information as defined by Jones [1]. 

(I) Controlled or owned by me
(II) About me without my control over it e.g. bank records
(III) Directed toward me e.g. phone calls
(IV) Sent or posted by me for others e.g. emails, web pages
(V) Experienced by me e.g. a book in the library and
(VI) Relevant to me but I haven't discovered it yet

If we stick to these 6 types of personal information, "about me" strictly falls into the group (II) while "for me" in the group (I). Although (I), (III), (IV), (V) and even (VI) can or not be about me, while (II) is about me only. Examples of info about me (I) my bank statement I keep in a folder, (III) a sent passport by the government to my home address, (IV) photos about me I post on-line, (V) an autobiography book I have(n't) authored in the library and (VI) something about me I don't know yet. Likewise the are examples of information not about me under each of these categories.

So what are these in-between instances of personal information in the social context (for others to consume) I mentioned before?

1. Information that we have (some) control over it - we manage it. This can be either information about me and/or information for me (which in the social context could also reveal a lot about me as well). There are three different types of such information:

  • Information that we manage primarily for ourselves, but also make it publicly available for others to consume (this information is not necessarily about us but can reveal a lot if information about us). Examples include Delicious, Mendeley, Amazon wish list, etc.
  • Information that we usually manage for ourselves only, but individual items or even collections are shared with known end-users (information could be about us or not). Examples are Google Docs, Dropbox, Sharepoint, etc. (if publicly available can also fall under previous group).
  • Information that we publish (sometimes even manage in categories) with the intention for others to consume and is in general about us (but not only). This includes blogs, social networks, personal web pages, etc.

2. Information that we have no control over it (or maybe, if we stop socializing :)).

  • Information that we do not publish but is about us and we have not control over it such as our photos online posted by other people (company, friends ...).

This second one (2.) is problematic. By definition, this is still personal information from the (II), although Jones mentions only medical, school, police, justice, bank or state records about us, which are usually not disclosed to the general public. We in general trust these institutions to take good care of information "about me". However, when information "about me" gets disclosed. we have no control over it (just imagine our credit card details sold on the black market or just the photo of us in the newspaper). Potentially can such disclosure cause real harm, but most of the time it might be just annoying (our friend uploading photos of our holidays, parties, etc.).

While the first one (1.) is still information "for me" even though we make it available to others as well. But it can be also information "about me" either directly (posting our photo on Flicker) or indirectly (posting our reading taste on Amazon wish list or hobbies on Delicious). Karger proposes a new term

  • End User Information Management

which better captures who is managing information - but IT DOESN'T equal to information "for me" only as we could manage it for others as well. This end user information management could be done to satisfy our needs (e.g. the need or better the wish to post holiday photos) or needs of others (e.g. our parents and friends who wish to see these photos). David for example mentions his management of music (old style PIM) which hopefully satisfies needs of others when he's DJing. While the second in general (2.) satisfies just the needs of others!

To distinguish the type of management maybe these two axes should be considered:

  • who is managing personal information (we or others) and
  • whose needs this management satisfies (our needs, also needs of other or just needs of others).

And we would end up with six different categories of information like this (while each information could span over several cells):

Management of info \ Satisfying our needs our and/or needs of others needs of others
by me



by others



Each category of personal information should be then be weighted against privacy, policy controls, ethics and possibly other issues concerning the information in question. So there are not only "about me" and "for me". The issue is definitely more complex.


[1] Jones and Teevav, editors. Personal Information Management. University of Washington press. 2007

Suggestions on PIM applications design

PIM research produced a significant amount of work. Many researchers gave guidelines on how future PIM applications should be designed or what should be taken into account when designing them. These guidelines are based on observations during the PIM studies. Let's look at some of them.

Henderson [1] argues that there are three groups of users: pilers, filers and structurers. PIM tools should accommodate the following:

  • Piling strategy
    Do not require containment
    Support a time based interface
       Provide optional tagging
  • Filing strategy
        Support containment
        Provide a cleanup interface
        Support reminders
  • Structuring strategy
        Support containment with multiple classification/dynamic containers
        Provide optional relationships between items
        Provide optional tagging and colour coding
        Provide optional custom metadata

Ravasio et. al. [2] looked at the problems users encounter when managing files on computers. She and her colleagues addressed several issues that need to be addressed quickly

(1) ‘Annotations’ as a new type of information. Suitable handling—in programs, storage and retrieval facilities, and so forth—are subsequent requirements.
(2) Store documents in publicly available default formats in order to avoid difficulties when exchanging documents.
(3) Provide a ‘search for content’ feature that includes all file formats that exist in a system.
(4) Simplify the UI of the existing built-in search tools. Do not require users to specify more information than they can possibly know.

In general, She claims that PIM systems should accommodate 3 aspects of file management:

(1) Task oriented—focused on the task to be accomplished,
(2) context oriented—focused on other documents, programs and tasks at hand con- currently, and
(3) content oriented—focusing on the actual information encapsulated in a specific document.

Bergman [3] identified 3 subjective principles that should be taken into account in PIM

  • The subjective classification principle: information items related to the same subjective topic should classified together despite technological format.
  • The subjective importance principle: this should provide a degree of visibility or salience and accessibility. When exposed to new information users determine how important that information is and systems should take in to account the this importance as perceived by the user.
    - High subjective importance: what was dealt with most recently and what was frequently dealt with in last xy time is the most important.
    - Low subjective importance: a cognitive problem or irrelevant information being in a way while searching for important information, users do not archive old information, graphically change items not being accessed for a long time specified by user.

  • The subjective context principle: information should be retrieved and viewed in the same context in which it was previously used. Context should be captured and added to information. - External context: items that users dealt with (activated) while interacting with a specific item: like coping and pasting text should create a link.
    - Internal context: relates to users thoughts while interacting with information item, users might write annotations of what they read and search these.
    - Temporal context: this pertains the state in which the user left information item when last interacting with it and working plans regarding that information. Similar happens with unread email that are in bold and web links (change color when visited) so the user knows what has been already done - trace past activities. Users could mark information they plan to work on in the future.

Boardman [4] highlights two perspectives that need to be taken into account (he also defends the idea of separated information by type):

  • From a tool-specific perspective, each PIM-tool is a distinct sub-system to be optimized independently.
    - consider how that tool is employed along with other tools in supporting cross-tool production activities
    - be aware that anything new is added to a set of existing PIM sub-systems
    - careful attention should be paid to potential side-effects in other tool contexts
  • On the other hand, a cross-tool perspective emphasises the need to optimize the combined sub-systems. In other words, the designer is more concerned about how well the PIM sub-systems work together.
    - when investigating user needs, designers should pay attention to current user behaviour across all the tools that will be affected
    - a key implementation challenge is dealing with the sheer range of PIM-tools in use

Voit et. al. [5] addressed different aspects that need to be taken into account when designing PIM tools. They claim that PIM tools should comply with this set of requirements:

  • Be Compatible with Current User Habits: Users are comfortable with their application environment and want to keep it that way. Any new software solution has to integrate into the current environment as smoothly as possible.
  • Minimal Interference: Any new software solution requires some kind of additional user interface. It is essential to keep the learning effort as small as possible. Any interaction step which the users have to make should be absolutely necessary to the process.
  • Support Multiple Contexts: Studies show that over the years users still prefer browsing over teleporting. When browsing a classification hierarchy, users can see the choices available at each level and choose the most promising.
  • No Unnecessary Limitations: Since large numbers of computer files define our everyday lives, any PIM software solution should scale well to a large number of files and should not affect the efficiency of the browsing process.
  • Transparency: User have built up knowledge of their software environment: a set of experiences, expectations, and standard processes concerning file storage and retrieval. For example, an existing backup process should not be affected by a new PIM system. Users should know where their files are located and what happens to them.
  • Provide for Expiry Dates: Providing an expiry date offers the user to explicitly define information as ephemeral, which is an important need as user studies suggest.
  • Add Metadata While Storing: When a file is stored the user should be given the option to manually add metadata and contextual information to the file. Manual and semi-manual tagging can offer an effective solution for a better retrieval method.

[1] Henderson, Sarah, Personal document management strategies, CHINZ '09
[2] Pamela Ravasio and Sissel Guttormsen Schr and Helmut Krueger, In pursuit of desktop evolution: User problems and practices with modern desktop systems, TOCHI, 2004
[3] Bergman, Ofer and Beyth-Marom, Ruth and Nachmias, Rafi, The user-subjective approach to personal information management systems design: Evidence and implementations, JASIST, 2008
[4] Richard Boardman, Improving tool support for personal information management, PhD thesis, mperial college London, University of London, 2004
[5] Voit, K. and Andrews, K. and Slany, W., Why Personal Information Management (PIM) Technologies Are Not Widespread, ASIST '09: Personal Information Management (PIM) Workshop

Filers, pilers and other strategies of document management - do your recognize yourself?

Many researchers tried to classify people in groups depending on PIM strategies. Here's a short overview of classifications.

Paper documents

Malone 1983 [2]: files and piles are the basic building blocks of paper document management.

  • neat: designate a category for every document and place it the location corresponding to that category; the location may have been a folder inside a filing cabinet, a paper tray, or a named pile.
  • messy: pile up documents over time, in a less structured way.
Whittaker and Hirschberg, 2001 [10]
  • filers: filers amassed more information, and accessed it less frequently than pilers;  filers engage in premature filing: to clear their workspace, they archive information that later turns out to be of low value; given the effort involved in organizing data, they are also loath to discard filed information, even when its value is uncertain
  • pilers: tend to amass less documents and access it frequently than filers

Computer files

Boardman and Sasse 2004 [9]:

  • total filers: file majority of items on creation.
  • extensive filers: file extensively, but leave many items unfiled.
  • occasional filers: file occasionally, leave most items unfiled, have few folders.

Henderson 2009 [3]:

  • filing strategy: variants of the pro-organizing, frequent-filer and keeper categories identified by others - filer is organized, with just in time folder creation, combination of browsing and searching only as a last resort, the hierarchy structure is medium in depth and width and has a moderate number of unclassified top level folders.
  • piling strategy: analogous to messy, no-filers, keepers, and organizing neutral strategies identified by other researchers
  • structuring strategy:variants of the pro-organizing, frequent-filer and keeper categories identified by others - high depth, low level of unclassified files, folders created in advance or just in time creation and consider themselves to be fairly organized.

Email

Mackay 1988 [1]

  • prioritizers: concentrate on managing incoming messages
  • archivers: use email to archive information for future use
  • requesters and responders: use email for task delegation

Whittaker and Sidner 1996 [6]

  • no filers: allowing all their email to pile up in the inbox using search to retrieve emails
  • frequent filers: attempted to place all their emails into folders
  • spring cleaners: periodically attempting to put their emails into folders

Balter 1997 [5]

  • no filers: allowing all their email to pile up in the inbox using search to retrieve emails
    • folderless cleaners
    • folderless spring-cleaners
  • frequent filers: attempted to place all their emails into folders
  • spring cleaners: periodically attempting to put their emails into folders

Boardman and Sasse 2004 [9]:

  • frequent filers: file or delete most incoming messages everyday
  • extensive filers: try to file many messages everyday (employ different above mentioned strategies e.g. a combination of frequent filer, spring cleaner, and no-filer)
  • partial filers: file only a few (<5) messages everyday (employ different above mentioned strategies e.g. a combination of frequent filer, spring cleaner, and no-filer)
  • no-filers: do not file any messages.

Gwizdka 2004 [7]

  • cleaners: have specific times for dealing with email, and don‟t keep events or to-do items in their email
  • keepers: read email constantly, allowing tasks to be interrupted by email; they keep events and to-do items, and search their email archives.

Web bookmarks

Abrams at al 1998 [8], depending on whether and when the user saved web bookmarks during a browsing session

  • no-filer
  • creation-time filer
  • end-of-session filer
  • sporadic filer

Boardman and Sasse 2004 [9]:

  • extensive filing: file many bookmarks as they are created or at the end of browsing session (employ different above mentioned strategies)
  • partial filing: file bookmarks sporadically (employ different above mentioned strategies)
  • no-filers: never file, all folders abandoned
  • no collectors: not collecting bookmarks

Boardman and Sasse also found out that people use different strategies in different hierarchies. People filed more files than they did email and bookmarks. Kamaruddin [4] studied folder creation and found that people either plan in advance or retrospectively create folders and employ both strategies depending on the project. Note that she classified folder creation and not people into groups. This is contrary to Henderson's study where she clearly divided people who create folders in advance and people who create folders retrospectively. Our study (see Publications) also showed that people employ different strategies depending on project/task evolution. Some tasks are well defined in advance, while others 'become tasks' after a while (when users see them as tasks) and might or might even not get organized in folders (cost/benefit approach).

[1] Mackay, W.E., More than just a communication system: diversity in the use of electronic mail. in CSCW'88
[2] Malone, T.W. How do people organize their desks? Implications for the design of office information systems. ACM Transactions on Office Information Systems, 1
[3] Henderson, Sarah, Personal document management strategies, CHINZ '09
[4] Azrina Kamaruddin and Alan Dix and David Martin, Why Do You Make A Folder?, BCS HCI '06: Poster
[5] Balter, O. Strategies for Organising Email Proceedings of HCI on People and Computers XII, Springer-Verlag, 1997.
[6] Whittaker, S. and Sidner, C., Email Overload: exploring personal information management of email. in CHI'96
[7] Gwizdka, J., Email Task Management Styles. in CHI'2004
[8] Abrams, D., Baecker, R. and Chignell, M., Information Archiving with Bookmarks: Personal Web Space Construction and Organization. in CHI'98
[9] Boardman, Richard and Sasse, M. Angela, "Stuff goes into the computer and doesn't come out": a cross-tool study of personal information management , CHO '04', 2004
[10] Steve Whittaker and Julia Hirschberg, The character, value, and management of personal paper archives, TOCHI, 2001