Review of Prototype design – a peice that was cut out the paper

Boardman grouped prototypes in four groups

  • Application specific design
  • Integration of PIM applications
  • Embedding PIM applications
  • Unifying designs
    • Time based designs
    • Contact based designs
    • Project based designs
    • Attribute based designs
    • Other designs

 Karger uses a different approach to grouping tools and ways to unify information space

  • Visual unification
  • Standard data types
  • Metadata
    • Grouping
    • Cross referencing
    • Attribute/value
    • Relations

Rather than grouping PIM and other applications in groups we looked at how new PIM applications and prototype design address drawbacks of present PIM applications in four main PIM activities [20, 19].

  • Acquisition: Mailcat
  • Organization: Presto, Haystack, Lifestream, Umea
  • Mainteinance: Timemachine
  • Retrieval: Stuff I’ve seen, all mentioned in organisation help also retrieving information

Prototype design

Prototypes and new solutions to help users manage information were developed for various pur- poses. Some prototypes were designed to address problems users have in each hierarchy sepa- rately, other prototypes try to merge several information formats to address information fragmen- tation while some propose different management approaches as a solution to hierarchy drawbacks.

***********************************************************

Boardman grouped prototypes in four groups [6]:

• Application specific design: addresses improvement of a design of a specific tool like email client, file manager, etc.
Several prototypes were developed to provide alternative approaches to present PIM applica- tions. The Piles prototype [26] provides a pile metaphor as main organization scheme instead of a hierarchy based on Malone’s observations of piling advantages (faster retrieval, piles tended to store less documents than filers, reminding advantages …). Data Mountain used a 3D spatial environment for managing bookmarks while WebBook used a book metaphor [7]. Multiple classifications (which static hierarchies lack) were provided by a tagging system like Gmail. Email client Mailcat assisted with automatic classification and filing of email [27].

• Integration of PIM applications: integration of two or more information types that can be managed in one PIM application.
Stuff-I’ve-Seen [9], Windows Desktop Search, Apple Mac OS X Spotlight, Google Desktop and similar real time indexing software and search engines try to unify PSI through search capa- bilities across several information formats. Some limited integration is provided by several present email clients where clicking on a sender (in an email) results in opening contact manager or clicking on a date results in opening a calendar. Several tools in Mac OS X use advantage of text parsing and pattern recognition (like names, places, emails …) to easily transfer information from one toll to another (an appointment in an email to calendar).

• Embedding PIM applications: provides a way to embede one information format in applica- tion that manages another information format.
Simple embedding is supported by email systems where files can be a part of email mes- sages. A prototype TaskMaster provides management of emails, to-do items and files in one application [5].

• Unifying designs: provide management facilities to manage several information types in one PIM application.

– Time based designs
Lifestreams orders several information types in a chronological order but it allows to order items based on attributes as well [12]. Milestones explores time representation further with introduction of milestones in cross- tool search engine [25]. Milestones represent personal information like events, photos, etc. and help users retrieving personal information. Mylifebits, an approach to store all user’s information, uses a timeline as one of many data presentations [13].
– Contact based designs
Contactmap prototype uses a spatial social network of an individual and organizes per- sonal information based on this network. From a network an individual can navigate to files, emails and bookmarks that are associated with contacts [32].
– Project based designs
Users often store their personal information based on projects. This approach is used in UMEA prototype [17]. UMEA uses a users defined projects and stores all used informa- tion in the project that is selected at the time.
– Attribute based designs
Dynamically creating PICs based on attributes of information items of various formats is used in Presto [8] and Haystack [15].
– Other designs
One of the ambitious projects to store a lifetime of information is already mentioned Mylifebits which tries to store everything a users comes across like: tv shows, phone calls, photos, maps, documents, emails, etc [13]. The product is still in development.

***********************************************************

Karger uses a different approach to grouping tools and ways to unify information space. He focuses on ways to surpass information fragmentation [18, 34] and divides (PIM) applications in three main groups based on information unification techniques:

• Visual unification: tools that help users organize their visual digital workspace to see more information types at once. This can be done with managing different application windows on a computer screen in such a way that all windows are visible. With several windows opened this can lead to cluttering. A way to group semantically connected windows together are virtual desktops. A product that tried to solve cluttering was called Wincuts which allowed users to select parts of their windows and show only those parts on a desktop [28].

• Standard data types: this type of unification is based on presentation of information items with text and files which can be copied or cut from one application and pasted to another. The main weakness of this approach is that reach information meaning gets lost while chunks of information lose semantics links with original source.

• Metadata: based on metadata information items can be grouped, annotated and linked. This groups of PSI unification is further divided

– Grouping: simple grouping cane be done with folders. But other approaches have been developed like already mentioned Presto and TaskMaster [5]. While Presto groups items based on tags, TaskMaster groups emails and to-do notes into tasks based on metadata.
– Cross referencing: one of the example of cross referencing is when clicking on an address of a sender in an email message the address book opens with contact information about the sender. This enables orienteering from one object to another. But it is still not possible for example link an e-book file to his authors’ details in an address book. Cross referencing is a main WWW navigation form.
– Attribute/value: we can represent (a part of) each information item in a structured form with e.g. XML. Every application could use some of this structured information for its purposes if needed. The main problem of this kind of unification would be a definition of a standard schema that all applications would understand. Already mentioned prototype Presto uses attribute/value to dynamically structure PICs.
– Relations: relations can be represented with for example relational databases or Re- source Description Framework (RDF). Bento is an example of a (commercial) product that uses database for managing personal information [10]. While RDF is used in a prototype Haystack [15]. Haystack provides a web like navigation interface and it uses metadata to (dynamically) create collections.

***********************************************************

Rather than grouping PIM and other applications in groups we looked at how new PIM applications and prototype design address drawbacks of present PIM applications in four main PIM activities [20, 19].

NOT FINISHED
* Acquisition Mailcat
* Organization Presto, Haystack, Lifestream, Umea
* Mainteinance Timemachine
* Retrieval Stuff I’ve seen, all mentioned in organisation help also retrieving information

 _________________________________________________________________________________

 LATEX CODE

 %************************************
\subsection{Prototype design}

Prototypes and new solutions to help users manage information were developed for various purposes. Some prototypes were designed to address problems users have in each hierarchy separately, other prototypes try to merge several information formats to address information fragmentation while some propose different management approaches as a solution to hierarchy drawbacks. Boardman grouped prototypes in four groups \cite{boardman04}:

\begin{itemize}
\item Application specific design: addresses improvement of a design of a specific tool like email client, file manager, etc.\\

        Several prototypes were developed to provide alternative approaches to present PIM applications. The Piles prototype \cite{mander93}
        provides a pile metaphor as main organization scheme instead of a hierarchy based on Malone’s observations of piling advantages
        (faster retrieval, piles tended to store less documents than filers, reminding advantages …). Data Mountain used a 3D spatial
        environment for managing bookmarks while WebBook used a book metaphor \cite{webbook96}. Multiple classifications (which static
        hierarchies lack) were provided by a tagging system like Gmail. Email client Mailcat assisted with automatic classification and filing of email
        \cite{mailcat99}.

\item Integration of PIM applications: integration of two or more information types that can be managed in one PIM application. \\

       Stuff-I’ve-Seen \cite{dumais03}, Windows Desktop Search, Apple Mac OS X Spotlight, Google Desktop and similar real time indexing
       software and search engines try to unify PSI through search capabilities across several information formats. Some limited integration
       is provided by several present email clients where clicking on a sender (in an email) results in opening contact manager or clicking
       on a date results in opening a calendar. Several tools in Mac OS X use advantage of text parsing and pattern recognition (like
       names, places, emails …) to easily transfer information from one toll to another (an appointment in an email to calendar).

\item Embedding PIM applications: provides a way to embede one information format in application that manages another
         information format.\\

       Simple embedding is supported by email systems where files can be a part of email messages. A prototype TaskMaster provides
       management of emails, to-do items and files in one application \cite{bellotti03}.

\item Unifying designs: provide management facilities to manage several information types in one PIM application.
          \begin{itemize}
             \item Time based designs\\

             Lifestreams orders several information types in a chronological order but it allows to order items based on attributes as well
             \cite{lifestreams96}.

             Milestones explores time representation further with introduction of milestones in cross-tool search engine \cite{ringel03}. Milestones
             represent personal information like events, photos, etc. and help users retrieving personal information.
             Mylifebits, an approach to store all user’s information, uses a timeline as one of many data presentations \cite{gemmell02}.\\

             \item Contact based designs\\

             Contactmap prototype uses a spatial social network of an individual and organizes personal information based on this network. From
             a network an individual can navigate to files, emails and bookmarks that are associated with contacts \cite{whittaker04}.\\

             \item Project based designs\\

             Users often store their personal information based on projects. This approach is used in UMEA prototype \cite{kaptelinin03}.
             UMEA uses a users
             defined projects and stores all used information in the project that is selected at the time.\\

             \item Attribute based designs\\               

             Dynamically creating PICs based on attributes of information items of various formats is used in Presto \cite{dourish99} and Haystack
             \cite{huynh02haystack}.

             \item Other designs\\   

             One of the ambitious projects to store a lifetime of information is already mentioned Mylifebits which tries to store everything a
             users comes across like: tv shows, phone calls, photos, maps, documents, emails, etc \cite{gemmell02}. The product is still in development.

          \end{itemize}
\end{itemize}

Karger uses a different approach to grouping tools and ways to unify information space. He focuses on ways to surpass information fragmentation \cite{karger06, jones_teevan07} and divides (PIM) applications in three main groups based on information unification techniques:

\begin{itemize}
\item Visual unification: tools that help users organize their visual digital workspace to see more information types at once. This can be done with managing different application windows on a computer screen in such a way that all windows are visible. With several windows opened this can lead to cluttering. A way to group semantically connected windows together are virtual desktops. A product that tried to solve cluttering was called \textit{Wincuts} which allowed users to select parts of their windows and show only those parts on a desktop \cite{tan_wincuts04}.
\item Standard data types: this type of unification is based on presentation of information items with text and files which can be copied or cut from one application and pasted to another. The main weakness of this approach is that reach information meaning gets lost while chunks of information lose semantics links with original source.
\item Metadata: based on metadata information items can be grouped,  annotated and linked. This groups of PSI unification is further divided
    \begin{itemize}
    \item Grouping: simple grouping cane be done with folders. But other approaches have been developed like already mentioned
             Presto and TaskMaster \cite  {bellotti03}. While Presto groups items based on tags, TaskMaster groups emails and to-do notes into
             tasks based on metadata.
    \item Cross referencing: one of the example of cross referencing is when clicking on an address of a sender in an email message the address
             book  opens with contact information about the sender. This enables orienteering from one object to another. But it is still not possible for
             example link an e-book file to his authors’ details
             in an address book. Cross referencing is a main WWW navigation form.
    \item Attribute/value: we can represent (a part of) each information item in a structured form with e.g. XML. Every application could
               use some of this structured information for its purposes if needed. The main problem of this kind of unification would be a
             definition of a standard schema that all applications would understand. Already mentioned prototype Presto uses attribute/value
             to dynamically structure PICs.
    \item Relations: relations can be represented with for example relational databases or Resource Description Framework (RDF).
             Bento is an example of a (commercial) product that uses database for managing personal information \cite{bento08}.
             While RDF is used in a prototype Haystack \cite{huynh02haystack}. Haystack provides a web like navigation interface
             and it uses metadata to (dynamically) create collections.
    \end{itemize}
\end{itemize}

Rather than grouping PIM and other applications in groups we looked at how new PIM applications and prototype design address drawbacks of present PIM applications in four main PIM activities \cite{kljun05, kljun08}.

NOT FINISHED
\begin{itemize}
\item Acquisition: Mailcat
\item Organization: Presto, Haystack, Lifestream, Umea
\item Mainteinance: Timemachine
\item Retrieval: Stuff I’ve seen, all mentioned in organisation help also retrieving information
\end{itemize}

%*************************************************************************************************************