In what category should I look?
You may once have been standing in front of shelves in a shop not knowing in what category to look. You may have been sitting in front of a website in an attempt to find information without getting the idea of where inside the structure of the website to find the desired information. Thus you have been performing the mental puzzle of figuring the logic behind the structure.
Maybe you have also been creating categories, organizing items from the ‘chaotic’ non-sorted world into a system. Just think of how and where you store your files on your computer: which folders do you use for what content? How do you remember where to find it? Maybe you have been in the situation of organizing information intended for others e.g. on a web service. Maybe you have asked yourself: How can I know what categories and what structure the users of this information expect to find at the web service?
Usability and information architecture
Major business- and research areas are occupied with these questions, namely interaction design, information architecture and usability testing. When professionals organize webservices or any other information structure they intend, with help of a variety of methods and tools, to organize content so users find the information they are searching for. Discussions are animated on which methods to use, on how measurements best can be made in the pursuit of creating a better user experience. Methods include those of interviewing real users, prompting them to structure web content with methods like card sorting [1], measuring their actual use of existing webservices with techniques like eye tracking [2]. In some cases construction of fictitious users like Alan Cooper’s “Scenarios and Personas” (Cooper 1999) serves as a tool in the design process.
Behind all these methods lies a problem fundamental to all information architecture: the problem of different people’s different perceptions of the world. What matters to whom and when? Information architects and interaction designers may exercise a categorization of information and services based on their achieved knowledge of best practice and typical user-patterns, but essentially they do not know how each individual user will perceive the designed structure.
Different vocabularies
What seems to lie behind the problem of useful consistent categories may be the problem of different vocabularies. Not only does each person employ different vocabularies for different tasks; but these vocabularies are also expanded and changed as we learn about the world. When searching we develop our perception of structure along with developing knowledge of the topic. We match search results to our preconceived idea about structure. Potential categories bud in different directions as we search and the organizing principle we found useful in the beginning of the search is now hardly a fair representation of the problem.
We are constantly trying out our ideas of structure on the real world objects. This also applies to web search. When we meet categories on e.g. a web portal we try to match our own idea of the website’s structure with the structure we meet. We match our own vocabulary with the vocabulary of the interface.
Companies ponder on how they should represent themselves and their products on the web. How should the web content be organized? Should the website structure mirror the organizational structure or should the website be organized from the customers’ point of view? I believe that interaction designers, information architects, usability experts and all other who are organizing content in categories strive to create the best and most useful systems. However it is difficult; it seems as if content always could have been organized in a better way, categories could always be different. And the vocabulary we are using will never match 100% of the users we want to serve as best as possible.
The Ironist and the Common sense person
I would like to suggest a cause for this unsuccessful Sisyphus-like striving. I will use the term ‘irony’ and ‘common sense’, as used by the American philosopher Richard Rorty. In this article I will, with Rorty’s terms as mental framework, discuss how to deal with the problem of representing information on the web – categories or not - and on how to make information searchable in a way that makes sense to the user.
In Richard Rorty’s book “Contingency, irony, and solidarity” (Rorty 1989) Rorty discusses two types of people – or rather two points of view – regarding the subject of ‘vocabularies’. The ‘common sense’ position is convinced that vocabularies are final, all-inclusive and stable. The ‘ironist’ position is constantly doubting, revising and questioning vocabularies:
“I shall define an “ironist” as someone who fulfills three conditions: (1) She has radical and continuing doubts about the final vocabulary she currently uses, because she has been impressed by other vocabularies, (…) ; (2) she realizes that argument phrased in her present vocabulary can neither underwrite or dissolve these doubts; (3) insofar as she philosophizes about her situation, she does not think that her vocabulary is closer to reality than others, that it is in touch with a power not herself. Ironists who are inclined to philosophize see the choice between vocabularies as made neither within a neutral and universal meta-vocabulary nor by an attempt to fight one’s way past appearances to the real, but simply by playing the new off against the old.” (Rorty 1989 pp. 73-74)
Thus the ironist position reflects the impossibility of uttering anything final and non-contradictory, and Rorty’s term “ironist” has nothing to do with irony in the everyday sense; uttering the opposite of the intended meaning.
Redescription as a concept is central to the ironist. Redescription means the different possible ways a topic could be described. The idea of a universal meta-vocabulary – a final way to describe the world – is rejected by the ironist. Essentially communication is to the ironist just different vocabularies clashing and thus keeping the ironist entertained or at least pondering over new emerging meanings caused by clashes between vocabularies. To the ironist there is no ‘real’ description of the world; there are only different descriptions of the world.
In opposition to the ironist position is the common sense position:
“The opposite of irony is common sense. For that is the watchword of those who unselfconsciously describe everything important in terms of the final vocabulary to which they and those around them are habituated. To be commonsensical is to take for granted that statements formulated in that final vocabulary suffice to describe and judge the beliefs, actions and lives of those who employ alternative final vocabularies.” (Rorty 1989 pp. 73-74)
The common sense position believes that it is possible to describe the world in a final vocabulary. If the common sense person apply enough hard work, research long enough (or read any sort of bible thoroughly) he will finally achieve a consistent and universal description of everything in the universe.
Rorty and web structures
If we apply Rorty’s philosophical distinction between the ironist and the common sense position to web users’ activity of browsing and searching for information and services we can imagine two stereotypes of users: The ‘common sense’ – user and the ‘ironist’- user. Both of them are the most extreme positions, but might be encountered in the real world in more blended forms.
The desired information can either be a well defined concept in the mind of the user or it can be ambiguous and ill-defined. In the last case the user has a general wish for knowledge within a topic or related topics but search is more like browsing and getting an overview. Cognitive psychologist David Perkins calls this last kind of search process a search in the ‘Klondike Possibility Space’ (Perkins 2000). Perkins apply the metaphor of the late 19th century search for gold in the endless wilderness of Klondike, Alaska, USA on the solving of ‘ill-framed’ problems where the very formulation of the problem is part of the problem. Even though Perkins considers matters as innovation and breakthrough thinking I assume that the confusion the first-time visitor may experience at a webservice with an - to her - opaque and unknown structure may be similar to breakthrough thinking - the answer may be right there but is invisible.
Common sense computer data and the ironist user
One may state that computer technology itself works against the interests of the ironist web-user and pro the interests of the common sense user. Computers are logic symbol-handling machines which distinctly do not understand their surroundings. They may ‘sense’ it but cannot understand it. The Turing test [3] of making a machine’s response so human that no one can distinguish has still to be passed.
Computers are not only insensitive to context but are also essentially following a rigid hierarchic structure. Due to the library structures of early operating systems like MS-DOS or Windows they insists in categories and in a final consistent vocabulary at their hard disks and in their applications.
The invention of the bookmark, the hyperlink and Worldwide Web changed the hierarchic situation and made the ironist partly happy. With hyperlinks and bookmarks hierarchic hard disc structures were by-passed by the user. To the ironist user web search became a matter of finding a good link and keeping it in a safe place.
The search algorithms of e.g. Goggle made the ironist user even happier providing her with a non-hierarchic search possibility. The common sense user was despairing at the non-contextual chaos of high and low presented to him by Google. He preferred someone to organize the information for him.
The future ‘data dream’ of the common sense user is that all data is meta-data coded and described in categories and stored somewhere in a universal well-organized database. The future ‘data dream’ of the ironist user is to combine information she had not thought of but which intelligently challenge her current knowledge. She is looking for yet another comment on her thoughts like an intelligent conversation with a colleague. The ironist loves Amazon’s “Customers who bought this book also bought…” service.
Designing webservices for the ironist user and common sense user
Information architects and interaction designers have to relate the two kind of users when designing or redesigning webservices. They have to answer the question: How to create information architecture for the common sense and the ironist user? Is it possible to satisfy them both? What kind of research methods and techniques should be applied?
I believe the ironist user is the most difficult to satisfy. She attempts constantly to discuss or challenge any categories. She likes to reorganize things herself. The common sense user will be happy with a consistent never changing interface and he will follow the hierarchic paths already build. He will be satisfied if usability expert Jacob Nielsen’s [4] requirements of consistent and universal user feedback and interface conventions are met.
From Nielsen’s point of view a ‘best practice’ exists, consisting of web content which fulfil all usability criteria of efficiency and ‘ease of use’. The term ‘best practice’ could be seen as just another word for ‘common sense’ as the idea of ‘best practice’ essentially suggest a single universal solution to what an ironist may call a ‘complex problem’.
User Centred design and the Ironist and Common Sense User
According to tradition of user-centered IT-design it is crucial to involve users and user representatives in the process of creating a webservice. The tradition emphasizes designers’ social responsibility as the designed product has an impact on the life and opportunities of the users. See e.g. Jonas Löwgren and Erik Stolterman (Löwgren & Stolterman 2004). How would the two kind of users react to the methods and techniques applied by the professionals.
Card Sorting
With the Card Sorting technique the menu hierarchy of the webservice is being discussed and developed with users. Existing and proposed menu items are presented on cardboard cards and the users are asked to sort the cards in a hierarchy. The method is e.g. described at the webforum for information architects: “Boxes and Arrows” [5]. While repeating this exercise with different users a statistical pattern evolves informing the information architects about ‘how users structure the content’.
The ironist is unhappy with card sorting as her special way of sorting the cards is disappearing into the statistical uniformity of the average user. The ironist user will also ask: How was the test persons recruited? The common sense user is glad as finally a menu hierarchy finally is established.
Company taxonomies and controlled vocabularies
On a corporate or organizational level another response is to create a so-called ‘controlled vocabulary’ to be used by all content producers. According to this strategy all web information published should be codified according to the taxonomy the company applies. Within the web system it should not be possible to describe e.g. the same product or service with two different names. In an article [5] at “Boxes and Arrows”, Karl Fast, Fred Leise and Mike Steckel explain the practical process of creating a controlled vocabulary for a company. In the end of the article however they admit that keeping a controlled vocabulary up-and-running requires substantial man-power:
“However, maintenance is required to keep your controlled vocabulary viable and usable. Constant monitoring, evaluation, and tweaking are critical. This may require daily reviews of search logs, regular testing with users, regular conversations with subject specialists, or other analysis.” [5]
An editor is needed to maintain, control and update the “Controlled Vocabulary”. The task facing this person is however tremendous; he has to keep up with and analyze all usage, get into the deepest cognitive processes of all users and figure out how they perceive the world. Furthermore he will be confronted with the internal company battle about influence and dominance as the ranking and taxonomy established by users may not match the taxonomy wished by management board.
The ironist user is not very happy about the idea of a controlled centralized taxonomy. She thinks it is waste of time and effort, as she knows that no system or vocabulary will be final. She asks herself: How do contributors initially figure out to which category an item belongs? The common sense user asks: Is it being done ‘in the right way’ as seen from his point of view.
The semantic web
Another response to the problem on a truly global level is the large research project ‘The semantic web’ [6] & [7]. The objective is to create a world wide shared and unified description of web content – a “Web Ontology Language”. According to Tim Berners-Lee, the inventor of the Worldwide Web and engaged with the Semantic Web project, the goal is “enabling computers and people to work better in cooperation” [8]. The idea is to mark up all information on the web with descriptions of type and category – so called metadata. By describing and thus contextualizing the information on the web computers would be able to ‘understand’ what they find when searching.
They may filter information according to the need of the user – instead of today’s e.g. Google search where concepts from very different topics may be presented side by side. With the semantic web computers will be able to search the web autonomously and they can be ordered to perform trivial tasks efficiently. A ‘trivial task’ could be the one of finding the nearest cinema where your favorite film plays at an evening when you and your friends are off-duty and where tickets are not sold out. In order to realize this vision, computers need to talk to each other about content. If machines should talk to each other information should be uniform and consistent, non-ironic.
The ironic web-user is not very happy about the project. Not only does she encounter common sense structures on a website, the very search process is also governed by a web ontology language. A nightmare! Can she escape the structure, the grid? In pragmatic real life the ironist web-user would probably have fun pondering over how those people who once build the web ontology structure the world; the ironist web-user would probably just look at the behavior of the web ontology language as another vocabulary at the playground.
On the contrary the common sense user is overall happy – as long as he agrees in the vocabulary applied in the web ontology.
Metadata is fun!
In a talk in November 2004 at the Danish association for professionals in usability, information architecture and interaction design Sigchi.dk [9], information architect Christina Wodtke [10] identified trends in future information architecture. She argued that one reason why information architects cannot proceed with the usual methods of structuring website content is the existence of “The Magic Search box”. Users now are accustomed to Google’s and other web portal’s search boxes and users expect the search results to be context sensitive.
The menu structure which presented the classic ‘webportal’ format is now more a presentation of the company or organization’s own idea about its structure, not anything being used by the goal-oriented web-visitor.
Wodtke imagines that in a near future a link will not be a link to another webpage but to a search for new web content within a certain context. Google Scholar [11] may be an example of this. Wodtke uses the word ‘content’ instead of ‘page’ as she declares the static web page dead, the dynamic composed customized webpage is reality on today’s web.
One way to qualify search and making it context sensitive is to let the users search on metadata; e.g. what kind of information, media type etc. as supported by the semantic web project. Metadata makes the user view information in many different perspectives emphasizing properties the user finds interesting. In this way the project of creating a semantic web actually meet the ironist’s longing for getting rid of categories. The ironist is hilarious enthusiastic about metadata – especially if it is being added by other users and could be a subject of a good discussion.
Customization and personalization
The future task of interaction designers could be one of managing meta-tags and controlled vocabularies in never ending striving to create the perfect sitemap or make all information machine-searchable and interconnected. Data-mining analysis of real life usage of data could pave the way for this work; digging in the data and analyzing how users search and treat information. This statistical work is already a big business in web research.
The future task of interaction designers and information architects could also be to facilitate users’ own creation of structures. One way to meet the demand from the ironist user is to let her customize the website herself. The possibility of making different ‘skins’ as seen in many consumer applications, does to some extend meet the demands of ironist user but when she also gets control over what kind of functionalities she can get on her interface she’ll become even happier. If she is a first time visitor she would like the website to ‘know’ her preferences; it should play with her, in her own vocabulary, surprising her in exactly that funny way she had not thought of.
The common sense user thinks customization is a useless waste of resources.
‘My own tag’
Tim O’Reilly, one of the founding fathers to the movement around the new concept of “Web 2.0” – the next version of the internet – writes in an explanatory article [12] on Web 2.0 how “Network effects from user contributions are the key to market dominance in the Web 2.0 Era”. The idea is that a web service gets better the more people use it. Users do to not only use the web service but also contribute to it. On example is the photo sharing / blogging service Flickr (http://www.flickr.com/) where the words users apply on there own and other people’s photos makes the collective shared photo database searchable in another way than e.g. Google’s image search or any professional image databases where categories are fixed and metadata on the pictures remain unchanged. To an ironist Flickr provides hours of social fun; what photos are e.g. tagged with what words – and why?
The online encyclopedia Wikipedia (http://www.wikipedia.org/) not only profit from the same kind of user contribution but has its whole existence depending on it. Wikipedia actually strive for dynamic common sense description of the world by incorporating the ironist user; as everybody is free to change the common sense definitions, these definitions only stay common sense until another ironist takes up the discussion. If no one wishes to discuss a topic it rests in common sense or silence. Lately, however, the total ironist freedom of Wikipedia has been compromised by certain incidents of removing or rewriting statements which to some were ‘unpleasant’. The concept of common sense consensus truth was thus challenged by the concept of power.
These kind of web services only live through committed users performing a social control by tagging or commenting and thus changing the web-service’ link structure. It is an emerging social control as the popularity of the message which gives it importance. No information architect involved in anything but overall structural decisions. Hierarchies are dynamically changing according to ‘market’ mechanisms. The ironist web-user is happy; finally someone as changing and indecisive as her self!
Conclusion
In this article I have been looking at different methods and paradigms for presenting web content to users. I applied two fictitious and stereotypical users – the ironist and the common sense user – when I assessed the web structures and the methods for organizing web services. I did this in a response to the assumption that users behave rationally and with clear goals when searching. I do not believe that users always apply the same paradigm or mental schemata as the information architects behind when searching the web service. This is why I bring in the two extreme positions of the ironist user and the common sense user.
I claim that the library structure of the computer supported the common sense user while ignoring the ironist, but the hyperlink and the search algorithms now support the ironist. I claim the usability business for many years has be focusing on the common sense user and offering him ‘global’ solutions, leaving the ironist highly frustrated. I claim that behind some information architects’ and usability experts’ wish to make a definitive structure and interface of a webservice, web portal or application, the assumption lays that a final vocabulary is possible.
I – as an ironist – think that the project of making the perfect interface, the perfect organizing of information is impossible.
I believe that the positions of the two fictitious users will continue to co-exist. The task for an interaction designer or information architect must be to satisfy both positions. Ignoring the ironist web-user will push her away to competing webservices and applications, ignoring the common sense user will burden him with the meaningless job of creating his own categories not profiting from cognitions and insights made by other people.
Notes
[1] www.boxesandarrows.com/archives/
card_sorting_a_definitive_guide.php
[2] e.g.: Bing Pan, Helene A. Hembrooke, Geri K. Gay, Laura A. Granka, Matthew K. Feusner, Jill K. Newman (2004) The determinants of web page viewing behavior: an eye-tracking study
Proceedings of the Eye tracking research & applications symposium on Eye tracking research & applications, Pages: 147 - 154 March 2004, San Antonio, Texas, USA. ACM Press.
[3] Turing, Alan (1950) Computing Machinery and Intelligence Mind: A Quarterly Review of Psychology and Philosophy 59 (236): 433-460. October 1950.
[4] http://www.useit.com/
[5] http://www.boxesandarrows.com/view/
creating_a_controlled_vocabulary.php
[6] http://www.w3.org/2001/sw/
[7] http://en.wikipedia.org/wiki/Semantic_web
[8] http://www.scientificamerican.com/
article.cfm?articleID=00048144-10D2-1C70-
84A9809EC588EF21&catID=2
[9] http://www.sigchi.dk/
[10] http://www.eleganthack.com/me.html
[11] http://scholar.google.dk/
[12] www.oreillynet.com/pub/a/oreilly/tim/
news/2005/09/30/what-is-web-20.html
Literature
Cooper, Alan (1999) The Inmates Are Running the Asylum: Why High-Tech Products Drive Us Crazy and How to Restore the Sanity. USA: SAMS
Rorty, Richard (1989) Contingency, Irony and Solidarity. UK: Cambridge University Press. pp. 73-74 http://books.google.com/books?
Jonas Löwgren and Erik Stolterman (2004) Thoughtful Interaction Design. A Design Perspective on Information Technology. Cambridge, Mass. USA: The MIT Press. http://mitpress.mit.edu/catalog/item/default.asp?ttype=2&tid=10334