What is it all about?

‘Things I desire’ is a website where user can list their favorite products they encounter on the daily web journeys. It doesn’t matter if it is something he/she would like to receive or if it is something the user would like to give away. ‘Things I desire’ only lists the added article in a del.icio.us manner to provide the visitors only with the necessary information. It is not big, fancy, full of advertising, selling centered, company directed. It is a little ajaxy sticky, all time present, usable for grandpas and sisters, fast, shareable, social and completely user-centered.
The focus of this WebApp is its existence in the background. It is not like a Webmail Application where the user has always one browser window in use to simply keep the App running.
top

Add item

To add an item to ‘Things I desire’ people can do this manually through some formfields inside their for free account. Or they can use a bookmarklet similar to the “post to del.icio.us” link. If a user finds a product he/she likes and would like to put on his/her list, he/she clicks on this bookmarklet and will be redirected to the ‘Things I desire’ add item page. This page analyzes the previous product site and tries to guess as much information as it can. The adding process needs to be fast and non distracting. The web is already full of distraction and ‘Things I desire’ should try to make a difference here. Here is a wireframe to get an idea of what I am talking about:

Productname: The ‘Things I desire’ system should analyze the product website and look for headlines. These headlines are suggested towards the user and he/she can instantly pick one from the suggested list. If the provided headlines does not make sense he/she can enter one manually as well.
Rating: Our associative way of thinking is always about relations. I like this more than that, this is better than the one I saw yesterday, …. To allow the user to set these relations he/she can add a rating for it on how much he/she likes this product. One star does not mean it is bad product, it just means that the user liked it but does not really desire it, while five stars means a must-have.
Description: This formfield allows the user to assign even more associative information towards the listed product. The users could add things like ‘I got this link from John’ or ‘I finally found it’ and so on. But there is also the possibility to grab a simple product description from the previous product site. Maybe a little popUp could open up, listing all text from the product site in little boxes and the user selects the appropriate box.
This item is for: These two radiobuttons simply direct the product into the right list. Every user has a wishlist for its own and for his/her friends to adress the common panic right before birthday announcements or Christmas.
Images: This works similar to the productname thing. ‘Things I desire’ takes all images from the previous site and lists them there. The user selects one and this image is going to be displayed on the overview list to make browsing and searching for items easier. If the system can not fetch an image from the products site than the user can select a generic one, where its illustrations are in line with the common product categories. All images are stored on the ‘Things I desire’ server to avoid linking and updating issues. Of course the copyright is still by the owner of the images.
Categories: It has not been added on the wireframe yet since I am not so sure about it. There could also be a system of assigning a category for each product. These could be ‘Friends only’, ‘Xmas list’, …. The assigning of these could work similar to the del.icio.us tagging suggestions. The point I have with these categories is that they are making the system instantly more complex and all of the sudden the user has to make a decision about placing something somewhere. I would like to avoid these kind of decision and let the information be structured in a flat list to make the system as non distracting as possible.
The adding process is monitored by the system. If Person A added the product already and chose Headline3 and Image4 than these are going to be the default values for the Person B who is adding the product right now.
top

Overview list

This page simply shows the simple list of added products. Here is a wireframe again:

On the top the user can switch between his/her own list and the givaway lists. The icons on the top right are for adjusting the view inside the list (single long list or two column layout) The right side shows a list of his/her friends. It might happen that I can’t remember the list from Anna but I know that Amy has Anna on her list. It is again reasoned in our associative way of thinking. For each item the list shows the desired image and the chosen headline. Besides the description text, rating and Added date there is also the del.icio.us sharing feature to provide a more social approach. ‘I really like these cool japanese shirts. Let’s see who likes them as well’.
top

Issues

I have currently two issues in mind which needs to be solved somehow:
Flash: What happens if a product website is made in flash. The information can not be retrieved automatically but more importantly there is no direct linking possibility. The only solution I can think of right now is to add a clickable sticky on the ‘add item’ page saying ‘We could not retrieve the data. Click here to open up the website inside a popup and follow the on screen description to add a path to the product’ When the user clicks on the link, a pop up opens up containing two frames. In frame one the website is displayed and in frame two a list of formfields connected by arrows is provided. The user can then write down the necessary menu items to find the product.
If grandma is clicking on the Florians wishlist and the product is presented on a flashwebsite, the website is going to open up together with a pop up displaying the path Florian has entered before.
This is a pretty distracting thing but I am not aware of a smarter solution right now.
Patent: I am not quite sure if it is going to be problem but I have somehow in mind that the wishlist itself is being a patent of amazon.
top

What else?

API: API is probably the wrong word for it but it could be sweet to offer somehow tipps and guidelines on how to change the sourcecode of a shopsystem to make the retrieving of information to ‘Things I desire’ more easy and successful. Small shops could use this wishlist site as a wordofmouth advertising system.
Button: Besides a ‘Buy now’ button on some companies website there could also be an add to ‘I desire’ button. These shops would get the amazon feature and could make sure that the product is bought on their website and not on a different one.
Print: I would like to see the possibiltiy to print out a selection of this wishlist. I could see that in the form of a pdf as well as real paper. The shopping tours of parents are always paper centric and thus it could help them finding what they need.
Import amazon list: This could be a handy thing.
top

Money, Money, Money

Affiliate: ‘Things I desire’ could become a member of all big shopping affiliate programs and thus make money when users buy things using the ‘Things I desire’ system.
Commercial maps: I am not quite sure if I like the idea or not but it could be probably pretty easy to draw maps about the users product list. People who have an iPod on their list will most likely also have a nintendo DS in their box. This information might be quite interesting to companies and I am not quite sure if you could sell such statics to them. Maybe offering a subscription to a graph generator frontend or something like this. Dunno about the users privacy issue here.
top

Introduction

One of the most common question in the daily (especially business) life is the question of time: “How long will you need for this?” or “How long did you work on the job?”. The answer is mostly based on approximations. To guess the time for an action somehow accurately he/she needs to have experience in what he/she has to do. Quite a lot of projects fail or users get unsatisfied due to wrong approximations. Frederick Brooks wrote in his popular book ‘The Mythical Man-Month’: “More software projects have gone awry for lack of calendar time than for all other causes combined. Why is this cause of disaster so common? First, our techniques of estimating are poorly developed. More seriously, they reflect an unvoiced assumption which is quite untrue, i.e., that all will go well.[…]”(Frederick P. Brooks, The Mythical Man-Month: Essays on Software Engineering, 20th Anniversary Edition, Addison-Wesley Professional; 1st edition, 1995) Human beings are way too optimistic when it comes to guessing time and this does not only affect software development it is also true for all other creative work.
Todays home computers have the power to composite and cut movies, create music albums with complete virtual orchestration but these computers can not help the user in telling him/her how long he/she worked on a single file? This information could greatly enhance communicating the approximate timing for a specific task and besides that, help the user in getting an overview on his daily productivity.
top

Mechanism

But how to count the time the user have been working on a document? If a document window becomes the front most window, a timer starts in the background. If the window is not active anymore or the computer has been idle for a specified time the timer stops and this new duration will be added to the existing one. That’s all. The ‘Duration’ information plays from a data structure point of view in the same league as the ‘Created’ or ‘Modified’ information. It could be stored as an extended attribute to the file which is associated with the document from the current active window.
“Extended attributes extend the basic attributes associated with files and directories in the file system. They are stored as name:data pairs associated with file system objects (files, directories, symlinks, etc).”(developer.apple.com)
Now that each file has an information on how long the user have been working on that document there is also the topic on how he/she retrieves that information. Logically this information gets displayed when the user look at the file info for a specific file or a selection of files (directory or additive selection). But besides that the duration of files could also be incorporated into different applications i.e. calendar or project management apps. There could also be an interface for retrieving the duration of files for online applications like basecamp which could result in a complete new set of features.
Besides the single display of a duration in the file info properties there should also be the option of resetting the duration of a single or a selection of files. In a template-based working scenario the need for resetting the templates duration emerges sooner or later.
Adding a duration property to a file is a small adjustment to the current man-machine interface but it is one which takes advantage of working digitally(there is no background timer while cutting wood or painting a wall).
top

Private vs Public

But the topic of a files duration is putting up another question on the board: private and public meta data. Right now the operating system does not differ between private and public meta data. With the introduction of duration it could be necessary to create an interface for the user to declare what kind of information he/she would like to share and what kind of information he/she would like to keep in a more or less private frame. Apples iTunes for example counts the amount of time someone listened to a track or when he/she listened to it lately. These informations are stored inside a local iTunes Database and are not saved inside the music file. This solution is not very scalable or flexible but it addresses the just mentioned problem.
Developing a more solid and flexible system/structure on how people classify private and public meta data is going to become a necessity in the near future. The introduction of meta data search engines like spotlight or google desktop puts some pressure on that topic as well.
top

Annotations

There are several functions annotations fulfill. All functions have one thing in common namely that the process of annotating is always done besides an users action. The attention while annotating a text for example resides on the text. Or the focus of annotating in a collaborative situation (reviewing, creating) lies on the talk and not on the annotation action. Annotating is mostly defined as an unconscious act which is done parallel to something. Since computer interaction is mostly done in a serial way a digital annotation system needs to translate a real world parallel workflow into a serialized computer workflow. The key on this translation lies in the amount of attention the annotating process consumes. Thus the digital annotation system should not glance with its rich set of features it should rather focus on providing the right tools in the right moment to keep the users attention on the subject he/she is working on right now. This approach is very different to the existing annotation possibilities.
top

Layer

One property of real world annotation is its all time presence. There is no need to transform a sheet of paper into something different for simply highlighting some words for example. Nearly every kind of pen works on nearly every kind of paper. This is part of the unconscious annotation process. Thus the digital annotation system should not necessarily be an own application which can open all kinds of documents. It needs to be working inside every application with every kind of content inside the document window.
Neon is based on a transparency foil which can slide out of the title bar of every document window. This transparency foil is the the annotation layer where the user can easily work on without touching the original content. The advantages gained by the foils affordances are that each foil

1. has automatically an owner/author,
2. creates a consistent private space to act in,
3. distinct between annotation and original content and
4. communicates a consistent behavior throughout different applications

This layer will play a key role in the task of translating the parallel annotation workflow into the serial digital world.

top

Input Model

As mentioned earlier Neons characteristics are not its rich set of features or its similar look to paper like annotations. Neon focusses on consuming the least amount of attention for controlling the annotation interface.
My first Neon approach included a free form ink pencil which worked similar to real world pencils or popular Adobe Photoshop brushes to allow the user to draw every possible shape. It is a popular approach other digital annotation system follows. This approach tries to imitate the visual appearance of real world annotations to eventually accomplish the same functions and goals of annotations on paper. Obviously this does not work, since drawing with the gross control of the whole hand covering the mouse is not very effective in comparison to the fine control of a pencil with the human fingertips.
Annotations are very individual in their form and thus it is impossible to create palettes or toolbars with all kind of pre-drawn shapes to avoid the pencil approach. The solution lies in an analyzes of the users intention of annotations. This procedure revealed the tools needed to let the user annotate a given text without distracting him/her from the reading process. Basically there are Highlights, Emphasizers and Connectors. By assigning these three categories a visual appearances nearly all kinds of annotations are possible.

 

The basic interaction works as follows: The user selects a portion of text and then assigns an appearance to it. He/she does not need to worry about drawing a ‘question mark’, ‘bracket’, etc. The computer will draw them on the users command.
If there is no text selection possible the user will get a free form ink cursor which lets him draw on images the way a real life pencil would do. Images can not reflow or change their appearance thus the anchoring of annotations is fairly easy.

top

Final words

There are a lot of other features build into this annotations system (i.e. the sharing of annotations) but discussing them all would break the frame. You are welcome to download the complete thesis pdf or take a look at an excerpt of the Neons video presentation. This short video shows the input model mentioned above. You will see how the Neon layer is invoked, how the user assigns appearances to selections, how he/she is connecting text with highlights, how the selecting process works, that each ‘appearance’ allow multilple clicks to increase the level of importance for that annotation and finally the video will show how annotations can be edited.
 

Introduction

As mentioned in previous articles the requirements on which todays GUIs are based on changed a lot since 1984 where Apple Computer introduced the Graphical User Interface (GUI) for the mass market. Document management and handling has been an issue in the computerworld since its beginnings. Most users take the idiom of files and folders for granted and it is always part of their daily computer experience.
As in real life everybody has its own structure on how he/she is going to sort the documents he/she created. These systems are seldom transparent and that is why public environments need to define a mandatory file and folder structure on shared places (i.e. file servers,…). This structure should guarantee that every person who is connected to this ‘network’ will find the appropriate files he/she is looking for. Files and folders proofed to be a very efficient and even more important flexible system to address the individual file structures. Efficient and user friendly software solutions should be adjustable to the users needs and habits and the success of the present file management idiom does obviously a good job there.
But the problem with the systems flexibility is that every simple and repetitive task requires the same amount of user action as more complex and seldom activities. Moving specific files during an installation or adjusting configuration files is for example a task which requires the same actions as sending somebody files or saving current documents at a specific location. The operating system vendors are aware about that issue and provide the user with features like ‘recent place/files’ or tool- and sidebars to place favorite locations and frequently accessed files into. That solves the issue of the required speed, but the basic interaction requires still the same steps.
 top

Requirements

This article would like to address the issue on how people organize their incoming and outgoing files. In real life nearly every person needs to document when he/she sends somebody files or when he/she receives files. This is mostly an interaction which involves several steps and applications. The interaction of documenting a received file requires basically the same steps as the documenting of a sent file. If the user A sends a file to user B, this file is going to be part of the same sending and receiving interaction. It just depends on the viewpoint. Thus the interaction improvement should solve both documenting processes without any kind of transformation.
By looking what information are necessary for the users internal documentation of sending and receiving files it might get more clear why the current file/folder interaction is not sufficient. The user needs to know:

1. From whom he/she received the file.
2. When he/she received the file.
3. What kind of files are part of the ‘package’.
4. If he/she needs to consider something while ‘using’ the package.

How are these informations provided in the current files and folder paradigm?

1. The user could easily name a folder by the senders/receivers name.
2. Every file has the property on when it was created and when it was modified. But this does not automatically imply when the user received the file. So again he could create another (sub-)folder which holds the receiving date as its name
3. For every new ‘package’ he/she could create an own folder to document which files where part of the same package
4. Every folder could contain an extra file with some text information in it to provide extra information associated with that package.

These step will be necessary for the sender as well as for the receiver. Both need to do the same interaction and both might suffer from the non transparant file organization of the counterpart.
 top

Inspiration

The question that arises is why this issue has never been solved or if there is no solution to this issue without loosing the so well known flexible and established idiom of files and folders.
By looking at the way email gets handled it is obvious that the issue is known and that engineers and software designers solved the issue already. Most popular email applications borrow terms and systems from the real worlds snail mail. Emails are organized by mailboxes and there is mostly one inbox, one sent box and one draft box.
I am not going to suggest that the file system should be organized like current email applications since that would definitly ignore the flexibility the files/folders idiom is offering now.
But let’s take a look at the snails mail system of package delivery. It has been quite stable through out the time and by inspecting that system in more detail it reveals the question why this system has not found its way into the computer systems.
Everything somebody wants to send with snail mail needs to be put somehow into a ‘box’.

1. A box has a sender and a receiver.
2. A box has a delivery date
3. A box has a specific content which can not be changed without ‘destroying’ the box itself.
4. A box can have little notes you put on its cover to explain something about the content (fragile, urgent, latest version, only for John, …).

This is very similar to the requirements that has been written down before and thus I would like to claim:
Everything that leaves or enters a computer is packed in a box.
 top

Boxes

Boxes are kind of special folders with some extra informations (meta data) added to it. If the user would like to send files to someone/something, he/she is going to pack his/her files and folders into a special folder which he/she declares as a box. He/she adds the destination address as well as some notes about its content. If the desired destination is an email address the box is send via email. If the sender is a location on a file server or just a colleagues computer this box gets then transferred over there. The receiver will know exactly when he received what from whom. To work with files inside a box, these things needs to be put outside a box (similar to the ‘stationary pad’ behavior in Mac OS X). Boxes will always stay the way the original user packed them. To look into a box the user doubleclicks on it which is opening a new window displaying the appropriate files (read-only).
The box system provides a much more efficient delivery and documenting system than the manual non-transparent private file structures present today. By just being a special tagged folder boxes are fully backward compatible. This technology exists next to the current files and folders idiom which let the user decide wether to use boxes or their manual documenting system.
The use of present technology guarantees on the one hand backwards compatibility but on the other hand integrates the box into the current UI as well. If the user needs to deliver several screens from one layout for example he/she does not need to save the jpgs for the customer on his desktop to attach them later to an email. He/she can save directly into the box which he/she wants to be delivered. It is basically the same interaction as saving them on the desktop but saves the time for organizing and sending the content and documenting the delivery. It is simply a system to reduce the redundancy in the current file and folder interaction for documenting the transfer of files.
 top
 

Introduction

Apples iTunes is one of the most successful music player applications in the current computer history. This application changed the way people organize and listen to their music since it is one of the first music player applications which used the meta data inside the music files quite extensively. This system inspired a lot of different software vendors to follow the same path and that is likely to be one reason why the term meta data became so much more popular lately.
top

Playlists

A basic part in the daily music experience is the order of tracks the users listen to. The order of tracks is normally collected in so called playlists. These are generally linear lists of a music collection which are generated by the user. There are basically only two ways on he/she can listen to the music inside playlist–linear or shuffled. The amount and the kind of playlists vary from user to user since the choice of playlist is for example influenced by the musical taste, current mood, amount of music files, characteristics of the user, etc… . Since playlist are an important part in the daily music experience it is safe to assume that playlist have a personal value for the music consumer. Playlist are normally generated manually and thus took the user some time to create them. Playlist are not the kind of digital information/collection the user would easily throw away or can easily be recreated. Playlists are mostly created thoroughly and as mentioned earlier have a personal value.
This fact makes dealing with playlist rather static and clumsy. But there are several scenarios where people would like to interact in the linear and static field of playlist:

1. My computer should go to sleep after the last track from Pink Floyd.
2. When the music gets faster on my party playlist the visualizer should change
3. The computer should change the playlist automatically when this playlist is over
4. …

There is a need for interrupting the static behavior of playlist with actions defined by the user.
top

Smart Tracks

Smart Tracks are items which can be manually inserted at any point inside the static playlist. They behave exactly like standard audiofiles in this list namely when the iTunes application hits the smart track in the playlist order the track gets played/executed. But what makes a smart track smart? Smart tracks are similar to smart playlist-semantically descriptive actions. These tracks allow the user to define an action on what should happen when the track get hit by iTunes.
Smart tracks are based on AppleScript and thus can be easily extended/modified by the users demands. There is also the option to connect an Automator workflow with a smart track, which allows the user to define actions which includes applications outside from iTunes. For example: “When this track gets played pause this playlist for 7:13 minutes and open the WindowsMediaPlayer with the Kylie Minogue music video in fullscreen. When the 7:13 minutes are over quit WindowsMediaPlayer again and start the iTunes Visualizer.”
Predefined actions like Smart Tracks let the user enjoy/focus on the music itself instead of interrupting the music experience for controlling and commanding the iTunes music application. This could have a great impact on setting the computer yet another step into the background and the users goals into the foreground.
top

Introduction

The internet is popular for its presentation of different kinds of media. The users can watch movies, listen to music or simply view pictures on websites. There is a variety of different scenarios on how the user interacts with the media presented on a single webpage.
top

Situation

To provide the enduser with some information about the item he/she is interacting with right now, the author of the webpage is adding additional information next to the media object. This includes name, format, sometimes size, download options, copyright info, etc. In more detail it could also be EXIF or IPTC information on photo objects or frames per second on movies or simply additional information/licence about the use of this object. To provide the visitors of a given site with these informations the original author has to integrate it into the site design and implement them manually into the website content. The manual placement of related information increases the risk of misspelling or the display of information the visitor is not seeking for. One solution to both problems would be that these information are retrieved and listed automatically and that the visitor can decide on his own which meta information he/she would like to look at.
top

Solution

By looking at Apples Dashboard technology it becomes obvious that there are similiar problems for every dashboard widget. Each widget has some sort of meta information (copyright, author, company,…) and even preferences to display. Apple solves the issue by simply flipping the widget on its back so the user looks visually on the backside of the object to retrieve/set the necessary informations.
I would like to suggest to use the exact same mechanism on every media object in the Apple Safari browser. If somebody would like to gather more information about a photo on a website, he/she could simply flip the image and see the EXIF Informations or the size or the copyright and maybe a license on how this image is allowed to be used on its back. This little refinement would put the information directly in the right context and allows the user to decide when to see what where.
top
 

Goals-Tasks-Actions

A users action independently from the goal always needs to take place in some kind of space. It really does not make so much difference if it is a virtual or a real space. The reason why somebody uses some things or act within a system is the presence of his/her goal. This goal gets divided into several tasks which then forms the goal. Thus a goal consists out of tasks.
Since the amount of time people are spending in front of their computers keeps on increasing it is obvious that more and more tasks are transferred into the computer to accomplish finally a goal. To fulfill a task the user needs to interact with applications. One application can fulfill more than one task.
Most of the time single applications are part in a chain of tasks. A chain is defined by having a connection between each chain link. In an interaction this means that every item in that chain has an input port and an output port to connect all task to one chain and thus goal.
top

Tasks in software

In software these inputs and outputs are mostly realized through fileformats and import/export commands.
An application can take part in a chain of tasks more than one time. Taking a layout for example, parts of that might be graphics and the other part might be text. The graphics might be even split into vectors and pixels. So there are already three different application(text, vector, pixel) necessary to create a layout. Even though the task chain is pretty linear this does not mean that each application will be used only once. So there is a frequent interaction between the different applications and each input and output port will be used more than one time only. Switching between applications and transferring the content between these application are main interactions in the daily computer life.
The import and export between these different applications in the chain of tasks are always a break in the users workflow. The output of one task should be easily integrated into the input of the following task so that the user can focus on the outcome of a single task as well as keep an overview of the final goal. That means that if the distraction from switching between applications could be reduced, the efficiency in the chain of tasks could be increased.
top

Problem

One reason why the switching of applications can cause quite a distraction is on the one hand the different behavior and look of applications but on the other hand the problem that each window can only belong to one file and one application. The user can not drag a window into other applications to continue his work. He/she needs to switch the window/workspace to complete the next task.
top

Solution

Looking at the layer paradigm of the digital annotation system reveals a very promising solution. If a window could consist out of several layers then each layer could be associated with an application. This means that the user will still work within the same window but only on a different layer. Users are not switching between windows, they will switch between layers of one window. The windows title bar will become a clip for application layers. The user will not be distracted by moving content between applications. He/she will reside inside the same spatial environment while the computer takes care of the application switching.
This could decrease the distraction the application switching provides and thus increase the overall efficiency in the chain of tasks to accomplish the given goal.
top
 

Introduction

Scrolling is a common task in the everyday computer life. It is performed several times a day in completely different scenarios. The WIMP paradigm affords the use of scrolling in the popular computer interfaces since content and display space are nearly always counterparts. Mobile devices like pda, cell phones or ebook readers encounter the same problems. The idiom of scrolling has become a main activity in daily computer interaction similar to the doubleclick. The rise of the popular scrolling wheel demonstrates the importance of this basic interaction.
top

Distraction

One main issue in current computer systems is distraction. The computer does a very good job in doing a lot of thinks parallel and some of these things even give feedback to the user (’You got a new mail‘, webpage could not be opened, new software updates are offered, Instant Messaging pops up, etc.) The user is easily distracted from his main work. The interaction design of systems should focus on the outcome of the users work since that is the reason why the he/she is using the system intentionally. Distraction is thus something that should be avoided for guaranteeing a fluid workflow.
Reading long texts on the computer screens has always been a hassle and there are several different solution out there on how to improve screen reading in general. As mentioned in the introduction scrolling is very much part of the daily computer experience but scrolling is very much part of distracting the user from a reading flow as well. Victor Kaptelinin concludes in his study ‘Transient Visual Cues for Scrolling: An Empirical Study’ that “first, the [above] results suggest that scrolling can be a significant source of reading distractions. Disruptions occurred immediately after scrolling accounted for almost 30 percent of all disruptions observed in the experiment.” This distraction has one of its reason in the way the computer provides feedback to the scrolling action of a user.
top

Feedback

Feedback of a users action is one of the most important topics when human beings interact with different systems. Victor Kaptelinin did a study on the feedback of scrolling and he suggests that the “the window display ‘old’ lines as temporarily “dimmed” (but still readable) immediately after scrolling.”

Kaptelinins coded for his study a prototype in VisualBasic which resulted in a significant decrease of the reading disruption by the scrolling mechanism. This shows very well how much more space of improvement the current scrolling feedback implementation provides. I am not too convinced of his solution, because:

1. I see a discrepancy between dimming of specific letters in the scrolling feedback and the visual path the eyes of the reader follows. If the computer could recognize which word the user is reading right now and only dimm the words that he/she has been reading already then I could see a positive effect on the reading experience with this mechanism.
2. Dimming lines for showing that these had been present on the screen before ignores the complex formatting possibilities of text. How is Kaptelinins system going to work on a column layout text?

But besides these two issues his study shows that a rich visual feedback on the scrolling mechanism could increase the overall reading attention.

Suggestion

The distraction of scrolling a page by hitting the page down/up key or clicking on the scroll bar or hitting spacebar is reasoned in loosing the orientation of where the eyes rested before the scrolling command was invoked. There is a short amount of time where the users eyes needs to relocate the position on where to continue reading. This relocating time could be minimized if the scrolling process would be accompanied by a visual feedback, which slowly fades out when the scrolling process is over. There are several ways on how this feedback could be visualized. In the following demo I used a block and a single bar as an indicator on where the visual end of the page was.
The advantage of this simple and non-textual approach is that it can be used in various scenarios. It is in no way connected to the layout of the content displayed in the window.
top
 

This idea is not about bars or blocks it is about suggesting that the feedback of scrolling should be improved by providing a visual hint on where the end of the page resided before the scrolling was invoked.

top

Introduction

The customer is always right. This well known principle is mostly the right way how to communicate with a given user. This is especially important when in a chain of interactions something goes wrong. Errors inside a computer system can not be avoided since software is still written by human beings. Donald A. Norman wrote: “Of course, people do make errors. Complex devices will always require some instruction and someone using them without instruction should expect to make errors and to be confused. But designers should take special pain to make errors as cost-free as possible.” (Donald A. Norman, The design of everyday things[The psychology of everyday things], Basic Books, New York 1988) Thus an important part in the design of systems is how the situation is solved when an error occured. This article would like to adress some issues with the current design of error messages.
top

Goals, goals and goals

The user does an action because he/she has a given goal. If an error message pops up the user obviously can not fulfill his/her goal. That is without any doubt a very good indicator of a bad system design. The interruption of the current interaction is rarely something that is considered pleasant. That is why nearly every book/article/guideline will tell the system designer that error messages should be polite, illuminating and to some degree helpful. A lot of applications especially the major ones are aware of that fact and do their best to avoid bad error messages. But by going back to the beginning of this paragraph the users main focus is his/her goal – that is in the end all that counts.
top

Categories

There are undoubtly situations where the error is caused by the user since he/she did something which obviously can not work in that situation. Let’s call that user error. Than there are also these situations where it is definitely not the users fault that caused the problem – system errors. For sure there are also situations where these both categories can not be clearly separated but let’s ignore this situation for the moment.
If the most important thing is that the user accomplishes his/her goal how could these error messages be improved to help the user getting his/her work done?
top

Suggestion

System errors are by their definition not caused by the user and thus not their fault. The current implementation of error messages does not divide between user and system error and thus the user does not know if he/she did something wrong or if the system was the reason. If it is by no means the users fault that the error occured then I would like to suggest to tell this the user to let him/her search for a workaround to that problem. It is all about getting the work done. There are several ways on how to implement the difference between system and user errors. System errors could be red and user errors orange. System errors could use frog as an alert sound while user errors will play the submarine tone. This article would only like to adress the issue of cause and fault and not its visual or audible presentation. It is still important that error messages are polite and illuminating and it is good that customers are always right but at the end of the day all that counts is wether the user could get his/her work done with the system or not. That is the main focus and this should be never forgotten. This refinement could be very helpful for novice users since their knowledge of the inner workings of a system are rarely sufficient to determine who or what is the cause of the current error.
top

Example

If the user starts for example his/her FTP Client to upload some files to a different computer and he/she gets the error message “Could not connect to server xxx.xxxxxx.xxx” then how should he/she upload the files? Donald A. Norman writes in his book: “…if the task appears simple or trivial, then people blame themselves. It is as if they take perverse pride in thinking of themselves as mechanically incompetent.”(Donald A. Norman, The design of everyday things[The psychology of everyday things], Basic Books, New York 1988) Thus he/she might start to check Network Preferences and so on to figure out what is wrong. If the server is simply down but the rest of the network is working why not making it clear that it is by no means the users fault that the FTP Client could not connect to the server. If the user would know that it is not his/her fault than the user has only two choices left

1. find another server
2. wait until the server is up again.

That makes the whole situation clear and the user knows what to do to get his job done or in other words to reach his/her goal.
top
 

Introduction

When the computers appeared on the commercial platform they were marketed as a saving time and allowing new possibilities machines. But looking at todays offices reveals that quite an amount of time goes into the organization and maintenance of the machine, the operating system and file structure. This article is about a new idiom called projects which tries to reduce the administration amount and let the user focusses on his/her work again.
In 1975 Frederick P. Brooks published in his book ‘The Mythical Man-Month’ the following quote “Ease of use is enhanced only if the time gained in functional specification exceeds the time lost in learning, remembering, and searching manuals[…]Because ease of use is the purpose, this ratio of function to conceptual complexity is the ultimate test of system design. Neither function alone nor simplicity alone defines a good design.”(Frederick P. Brooks, The Mythical Man-Month, Chapter 4) A systems design affects the user productivity to quite an amount. Todays files and folders are used in much more different context than in 1984 where Apple Computer introduced the Graphical User Interface for the mass market with their product Macintosh.
top

Model of human thought

To optimize organization and handling of files and folders it is necessary to know habits and thoughts of the people using the system. There are several theories which studies the behaviour and thought of human beings. This is not the space to go into detail on this topic but there is one theory which is quite popular today and which fits very well into the critic of the current file and folder implementation. The theory is based on the filing cabinet model wherein there are lots of cross reference and pointers to other records. This hypothesis is more popular known under names like ‘schema theory’, ‘frame theory’ or ‘semantic networks’ and ‘propositional encoding’. “The essence of the theory consists of three beliefs, all reasonable and supported by considerable evidence:

1. that there is a logic and order to the individual structures(this is what the schema or frame is about);
2. that human memory is associative, with each schema pointing and referring to multiple others to which it is related or that help define the components (thus the term ‘network’)
3. […]“

Found in: Donald A. Norman, The design of everyday things[The psychology of everyday things], Basic Books, New York 1988, p.115 .
The project idiom that is reasoned in the associative human memory architecture. In the current implementation of files and folders the user needs to create the asscociation between different files on his own by creating specific folder structures to save the associations for a longer time. But the important difference in the computer interaction in comparison to 1984 is that users are dealing with much more different applications and platforms (Web vs. Desktop, Email vs Disks vs iDisk) today. The file and folder structure does not save any association with emails related to specific files for example. These relations are saved in the users brain or he/she needs to document these manually. That is where projects are coming into the game.
top

Concept

First of all the project idiom is about clustering information by its context. Right now this clustering is done manually through the user by creating his/her own file and folder structure.
A project is based on context. Files that belong together in one context build a project. To organize the files inside these context boxes named projects the file types of the contained files could be used. To relate all the organized files to create the necessary context, users activities(When was the file created, when was it modified, was it send together with other files, who received the files…) are taken into account. That means:

By taking the above figure into account the result would be a folder with contextual files who can be organized by its file types for example. Context files are collected and through the relation to each other they are put into a (new) context (again). This system does not imply any file structuring. Thus the relation of context – organization – relation does not directly result in a sorting mechanism but some of the relation and organization facts could be used to create a folder structure. top

Behavior

Now the question that arises how is this project idiom is going to be behave from a users point of view? Projects are basically just special tagged folders. They are somehow similar to drop boxes. If the user wants to save a file which is part of a project he/she simply clicks on the project folder and hit save. No subfolder creation or sorting mechanism. He/she is simply assigning the file to a specific project.
But how to get the files out of the project folder again? There are two different scenarios on how to retrieve files

1. dynamic properties
   This involves user demands like: ‘show me the files from today’ or ‘show me all clipboard files’ or ‘show me the fonts used inside the documents’ and so on. The organization by dynamic properties will filter the files on the fly.
2. static properties
   That means that the user view the files in a more solid file and folder structure. This structure is created dynamically too but from the users point of view it appears static.

There are quite a lot of studies out there who tried already to find an universal folder structure for automatic sorting mechanisms. I personally doubt that such a structure could be found and that is why a categorization scheme is necessary.
top

Categorization

The topic of categorization (project categories like code, design, school,…) and appropriate organizing methods (folder structure templates, file types,…) are closely connected. The folder structure for a student project is quite different to a designer structure or a coder structure. So the general computer approach of 0-1-∞ does not work here. It is impossible to build a folder structure which is generally applicable for all situations but at the same time not too generic/abstract so the user still finds his files fast and convenient.
Projects do not only contain files which the user saved into them manually. The relation aspect in the flow diagram visualizes this. The user starts with context and organization. Than the computer monitors the relation and action of users which will then result in new contexts. The project idiom automatically saves all the clipboard data and fonts used inside the projects into the material folder. Emails and Websites will be part of these folders as well.
The project idiom is about association. The user focusses on his work/goals while the computer monitors his/her tasks to support and help the user.
top

Final Words

Apple computers follows already the approach of project folders with software like iTunes, iPhoto or Aperture. All these apps have an own folder where all files used inside the app resides. The user never worries about where to save what. They focus on their work/outcome. But the difference to the project idiom right now is that on the first hand it only works together with these specialized apps and on the second hand the retrieval of files always requires the app.

I am not going to present screenshots here since the visualization on this topic is secondary. There are a lot of different scenarios on how the project idiom could be visualized. The main essence of this idea is the structure behind it: Users save files into one folder without organizing or worrying. The computer monitors his/her actions and creates special views for the users demands to retrieve more information even faster.