Managing complexity

Working as an HCI specialist in a development team is working in a different culture having it’s own history, language, beliefs, powers and so on. Modern software systems are the most complex artefacts mankind is able to build. Accordingly, dealing with complexity has ever been a major issue in Software Engineering.

A central idea in software development is modularization, in general: How can we divide the total project into pieces that are manageable by small teams, but can be easily integrated again later on. This basically is what software engineers mean when they talk about architecture: dividing the system into components.

For independent development it is required that the components are loosely coupled, this means: Changing the internals of one component does not affect the development of any other component.

Three-tier architecture

A famous example of such an architecture is the three-tier architecture. It divides the system into the following components:

1. Data layer: Responsible for proper delivery and storage of data

2. Application layer: Here resides the functional part of the system. Real world processes are represented by software functions.

3. Presentation layer: This is where functions and data objects are finally presented to the user.

Such an architecture makes it much easier to handle complexity. It also adds a great deal of flexibility. For example, new user devices can be supported by replacing the current presentation layer with a different one, which may for example make the application accessible on smartphones. Also, the data layer can support different applications. For example, a webshop may (and typically does) use the same data layer for handling the incoming orders and for managing the warehouse.

Architecture and usability

You may ask: If architectures are so marvellous, why shouldn’t we get fooled? Here it goes: Remember the quote:

“The product is nearly finished, except for the usability”
– anonymous developer

In user-centered design it is common sense to do usability first, for example: Know your users, know their tasks. But, if the presentation layer can be developed independent of the other components, why not doing usability last?

The answer is: Independent development, i.e. coding, does not mean that the components are independent wrt. requirements. The presentation layer can only display, what is supported by the application layer and the data layer. One may be able to quickly change colors, fonts, type of charts or even the device on which data is represented, but, if a data object is not available in the lower layer, there’s no way to display it.

Not so obvious: Several functions that are demanded by well-known usability guidelines need serious implementation on the two lower layers. Just to name a three:

Speak the user’s language

Imagine you have a nicely designed, successful website and want to make it available in a second language. This requires every data object to be held in more than one language instance. If this hasn’t been foreseen in the data model, you’ll have a hard time.

User control and freedom – Undo

Enabling the user to make every action undone is sth. we would ever demand for virtually any application. It helps so much, to safely explore the system and to recover from errors. But, an Undo function imposes serious requirements on the application and data layers: The application layer has to keep a record of every action the user has done and also has to “know” the reverse of each action.The data layer has to make sure that any “historical” state of the data is kept in order to recover it.

User control and freedom – Cancel

Setting the user in control also means to be able to cancel any long-running process at any time. Here, application and data layer together have to make sure that every incomplete (i.e. cancelled) action leaves the system in a well-defined state. This is often referred to as a rollback, and it is far from trivial.

Conclusion

Architectures are a wonderful thing and developers usually present them with pride. But, as long as some software developers think of usability as a skin-deep property, the three-tier architecture is harmful. Whenever you hear a software developer say something like:

We have a three-tier architecture, we can make the user interface look as we want at any time.

Ask him oder her:

• Can we make this website multilingual, once it has become a success?

• Do we have an Undo function for every filter the user applies to the image?

• Can the user at any time cancel the defragmentation of his hard disk?

1. A study exploring an innovative technology

2. Applying a psychological theory

3. Evaluating an artefact or a method

I will now summarize the solutions to this homework:

Exploring innovative technology

Originally it was the idea to find studies that explore how people are using innovative technology, like it was the case with the Thatcher study on search behaviour. Such studies typically employ relatively open experimental settings and use qualitative data analysis in order to describe the variety of behaviour.

Apparently, most of the suggestions focussed on the term „innovative“. However, a study that proves a new technology to work as expected I would classify as exploratory with respect to feasibility („Can we make this work?“). Comparing two artefacts in terms of performance measures instead is clearly evaluative. It is also something that is seen critical by at least some researchers in HCI. I remember well the provoking talk of Greenberg and Buxton at CH2008. They claimed that usability evaluation can “be ineffective and even harmful if naively done ‘by rule’ rather than ‘by thought”, in particular if “done during early stage design” (Greenberg and Buxton, 2008).

Below I list my three favorite examples for exploration studies from your results:

• Bailenson et.al. (2007), Virtual interpersonal touch expressing and recognizing emotions through haptic interfaces, Human-Computer Interaction, 22, 325-353

• Gaver et.al. (2007), Electronic Furniture for the Curious Home: Assessing Ludic Designs in the Field, IJHCI, 22(1), 119-152.

• Kuber & Yu (2010), Feasibility study of tactile-based authentication. IJHCS 68(3), 158-181

Applying psychological theories

Compelling examples of applying genuine psychological theories are comparably rare in the HCI literature. On the other hands these articles often have some kind of profoundness and timelessness compared to the many evaluation studies. Thus, authors taking the challenge of applying a theory most likely will name the theory in the title or at least in the abstract. Some of you only found some psychology-sounding phrases in the abstracts and merely assumed that there must be some kind of theory behind it. Without at least naming this theory, these solutions didn’t meet the requirements.

My top three examples from your results were:

• Shin (2009), The Evaluation of User Experience in the Virtual World in Relation to Intrinsic and Extrinsic Emotion. IJHCI, 25(6), 530-553.

• Anderson, J. R., Matessa, M., and Lebiere, C. 1997. ACT-R: a theory of higher level cognition and its relation to visual attention. Hum.-Comput. Interact. 12, 4 (Dec. 1997), 439-462. DOI= http://dx.doi.org/10.1207/s15327051hci1204_5

• Xu, S. 2005 Development of Dual-Modal Presentations of Textual Information. Doctoral Thesis. UMI Order Number: AAI3202828., DePaul University, School of Computer Science, Telecommunications, and Information Systems.

Evaluation studies

Evaluation studies are probably the most common type of research. Typically, these studies measure outcomes in a lab situation or even a controlled experiment and argue about the benefit of doing a design this way or the other.

My personal top three were:

• Wilson, S., Galliers, J. & Fone, J. (2007). Cognitive Artifacts in Support of Medical Shift Handover: An In Use, In Situ Evaluation. International Journal of Human-Computer Interaction, 22(1), 59-80. doi:10.1080/10447310709336955

• Blandford, A. E.; Hyde, J. K.; Green, T. R. G. & Connell, I. Scoping Analytical Usability Evaluation Methods: A Case Study. Human-Computer Interaction, 2008, 23, 278 – 327

• Lee, S. and Zhai, S. 2009. The performance of touch screen soft buttons. In Proceedings of the 27th international Conference on Human Factors in Computing Systems (Boston, MA, USA, April 04 - 09, 2009). CHI ’09. ACM, New York, NY, 309-318. DOI= http://doi.acm.org/10.1145/1518701.1518750

A good joke

There is still sth. to add to my personal list of “honorable mention”: One of you found the silliest evaluation study I ever heard of. Btw., here we see such a vague relation to sth. being “psychological”. I wouldn’t judge this to be about applying a psychological theory, because no such theory is mentioned in the abstract. Still, a really good joke! Enjoy!

Suzuki, N., Kakehi, K., Takeuchi, Y., and Okada, M. 2004. Social effects of the speed of hummed sounds on human-computer interaction. Int. J. Hum.-Comput. Stud. 60, 4 (Apr. 2004), 455-468. DOI= http://dx.doi.org/10.1016/j.ijhcs.2003.09.007

Our research focuses on the nature of voice interaction and activation of psychological tendencies in humans by the power of prosody sounds. This study examines whether people’s impressions and behaviours are affected by variations in the speed of hummed sounds. The sounds consist of just prosodic components similar to continuous humming on the open vowel /a/ or /o/ without any language information. In interaction between individuals as well as among animals, temporal structures including voice speed or duration contribute rhythmic interaction and are closely connected to the participants’ dynamics of mental or emotional states. We think that this phenomenon can be applied to human-computer interaction, even through the variation in temporal structures of hummed sounds used to reduce the influence of content or meaning in language. Our interactive system mimics the prosodic features of the human voice by using humming-only sounds under three different voice speed conditions: (a) faster, (b) normal, and (c) slower than the original speaker. We examine whether the variation in the sound’s speed gives rise to both psychological and behavioural influences in the relationship between the computer and the subject through interaction. Subjects tend to prefer a computer with a normal or faster speaking rate to that with a slower rate on both usefulness and familiarity. Moreover, the speech rate of the subjects changed inversely to the variation in a computer’s hummed sounds. This study demonstrates the importance of temporal structure and emphasizes the need for an investigation of the fundamentals at work in interaction.

We may use this to predict the time a user needs to make a choice at each level of a hierarchically organized website. First, lets assume there be only one level with 5 choices. The factor c is a constant for processing speed, for the following examples we take this to be 1.

Given that A is the correct choice, the time for decision is

In the second example, there are two levels of navigation, which are not equally balanced (one subtree has 3 entries, the second has 2).

In order to reach A, two decisions have to be made and the decision time is

For reaching B the total decision time is

Thus, we can already make two observations on this small example:

1. Comparing the effort of reaching A in both examples, flat hierarchies (i.e. fewer levels of the hierarchy) are preferable.
2. Comparing the effort to reach A vs B in the second example, fewer entries in a navigation level make a faster decision, meaning less effort.

Decision time is effort and remember that in the card sorting method the distance measure between two topics can be defined as effort – the length of the path between either two topics. In the example below the distance between A and D is 5. But, according to Hicks law, effort for reaching A and D is not the same. Instead we get:

However, in total it does not make a difference whether a user wants to reach A but mistakenly arrives at D, or the other way round. He or she has to take both paths anyways.

In contrast, the distance measure of length of paths is questionable when comparing C and D. The distance from A to D and A to C are equal in terms of path length (5). But, according to Hick’s law, the effort for reaching C is lower than for reaching A, because on the second level there are four choices in the group containing D, but only 3 for C. This is neglected in the distance matrix of card sorting.

The third tool in our inventory for analyzing navigation structures is the SNIF-ACT model that is based on latent semantic analysis (LSA) and information scent theory. This model differs from Hick’s law in that it does not take the options to be of equal probability. Instead it predicts the salience of a term based on the semantic distance to the users goal.

An interesting obervation is that all three models are based on different measures:

• Hicks law: number of choices
• Card Sorting: length of path
• SNIF-ACT: semantic distance

But, as far as I see, they would all favor flat hierarchies, which is in contrast to the common practice in web design.

To wrap it up, there are two interesting research questions:

1. By combining the three approaches, can we increase the predictive power of how users travel the web and better advice of what makes an optimal design? For example, should we account for the number of choicesin the distance matrix of a card sorting study?
2. Is a flat hierarchy (e.g. a site map) truly superior. Most radically this means to abandon any hierarchical structures and present a list of all topics to the user. Intuitively this does not make sense, but why?

Anybody?!

1. helps to understand the problems
2. guides solutions

One of the students suggested that applied science also is responsible for transferring empirical methods (e.g. for measurement) from fundamental to engineering. This is absolutely the case. Consider eye tracking: First used to study human visual attention, today it is found in many usability labs. Below find an updated diagram.

Furtheron this is the place, where I will

• summarize results from your homeworks
• tell you about further ressources you might find interesting
• Answer questions you post here

Feel free to visit this blog and leave your comments or questions regarding the lecture! Remind, that this is a very public place, so please respect netiquette. Also, I strongly recommend that you maintain your and your collegues anonymity.