|
Search this group's discussions
|
Explain IA
|
Please use this topic for debating definitions, discussing entries, and explaining IA.
Originally posted at 12:35PM, 5 January 2010 PDT
(permalink)
Peter Morville edited this topic 29 months ago.
|
|
I'd like to use the first two sentences of the IA definition currently posted at the Wikipedia (http://en.wikipedia.org/wiki/Information_architecture) as a starting point - I believe they cover a lot of scope and are concise enough to be understandable.
My humble proposals on advancing this definition:
1) The word "art" as such implies "artistic" approach to the process of creating something, which, I'm afraid, is not exactly the same as "creating ... systemic, structural, and orderly principles to make something work". I'd suggest that "IA is the art and science of" be used at the beginning of the definition.
2) Before "expressing a model or concept", we need to comprehend what we deal with. Therefore, the next part, I think, should be "analyzing the information", followed by the definition of the scope of such information ("all information in the world" would sound a bit too challenging to me), e.g. “used in a given system or domain of knowledge”.
3) The outcome of the art of analysis should be the science of “expressing models or concepts to be used in the information access and delivery systems”. I think changing from singular to plural is important, as the reality of huge info systems requires multiple models, which is now even possible to express and deal with through the use of Semantic Technologies (RDF/OWL or Topic Maps). Another important point here: when we deal with information where structure (entity classes and relationships across them) change very often, we have to create “a model of a model” (i.e., a conceptual model at a higher level of abstraction than a typical info model representing entity classes and relationships) for a sustainable design of the “access and delivery systems” (otherwise, such systems would become suicidal in terms of maintenance and attempts to cope with ever-changing info-models).
Thus, the enhanced definition can be something like: “Information Architecture is the art and science of analyzing information used in a given system or domain of knowledge and creating conceptual models at various levels of abstraction to be used in the Information Access and Delivery Systems (IADS).”
I’m not sure if the “overseeing the IADS design and implementation” should also be part of the definition (I, personally, would include it, but it is not typical for existing definitions).
Thanks for the opportunity to express my thoughts about this – and to you, reader, for reading them!
Posted 29 months ago.
(permalink)
|
|
|
To keep up, you might also want to follow #explainia via Twitter.
It's nice to see the entries starting to flow in more quickly!
Originally posted 29 months ago.
(permalink)
Peter Morville edited this topic 29 months ago.
|
|
I had a reaction to your including "art" as part of IA, so I thought I would look at the definition of Architecture from Wikipedia.
"Architecture is both the process and product of planning, designing and constructing space that reflects functional, social, and aesthetic considerations....Architectural works are often perceived as cultural and political symbols and as works of art. Historical civilizations are often identified with their surviving architectural achievements."
Seems pretty well aligned with what you said. I think this has to do with the fact that both types of architecture have to do with creating a human experience. The human interacts with the building architecture and the information architecture. Light, color, space, texture all play a part in the sensory experience in both environments. I think you could add something to your definition that could expand on this aspect of it. Maybe include something about "how the information is used" so you are including the dynamic aspect as well as the static.
The science aspect of the result of the architectural effort has more to do with what the human can do in the building or in the user interface. Is it intuitive where the bathroom is in the building? Is it intuitive how to place an order from an on-line vendor? Is it clear what functionality is available in the building...in the on-line environment. Do I have what I need to make meals for a family? Do I have what I need to make a strategic decision?
Reports are a good example of an overlap between the presentation and the structure. A report is easier to obtain if the structure is well designed, the presentation of a report is essential to the usability of the report. I find it interesting that the art aspect is not applied more readily to report design.
Some of the confusion comes from when we try to define roles. I suppose one person could do the whole job of designing the data structure and the presentation, but I think that would be a bit much. There are some very different skills involved. In general, there is more science in the data structure design and more art in the UI design although the "material" , i.e., the information, is common to both.
Interestingly, I am currently using a tool that generates a UI from the data model. This means that I am able to manipulate the UI with my design of the data structures. As someone with a strong data management background, it is coming together for me. The challenge now is to generate the process aspects of the UI as well.
When you mention "conceptual models at various levels of abstraction" are you including presentation models or just data models. What other types of models do you use? Do you have a presentation model that you use?
Posted 28 months ago.
(permalink)
|
|
|
What a thoughtful posting! If you are, by any chance, planning to attend SemTech 2010 (unfortunately, I was not able to squeeze IA Summit into my conference budget for this year) I'd be honored to have an extended conversation since the topics mentioned and the questions asked are at the focus of my current work. Meanwhile, I'll try to give some brief answers/explanations.
I believe that “information used in a given system or domain of knowledge” should include not only the traditional “business information” (i.e. business data, documents, e-mails, etc.), but also the information describing such systems or domains of knowledge. Just consider “information” as “verbalization of tangible and intangible reality” (on a high abstract level), then the description of a system or a domain of knowledge becomes part of IA. Informational models, in my opinion, cover all aspects of an information system, including data, processes, presentational layer, business rules governing the system, etc., etc.
Semantic technologies (Topic Maps or RDF/OWL) enable creation of one holistic model, that may include various aspects of an information system, in an agile, iterative, continuous process. However, subdividing such a big model into aspect-specific models makes better sense from practical usability point of view.
A report, I believe, deserves its own (abstract) model that would enable data mapping to presentational layer and can facilitate implementation of data delivery services and UI. For example, a report can have a summary that would be mapped to the main info entities, which, in turn, have “child” entities (a.k.a. “drill-down” parts); it can contain quantitative and qualitative (narrative) parts with the latter also being “children” of main reportable entities; quantitative parts may be represented by graphs or tables, or alternation thereof, etc.
One of the main problems in IA (as in system design in general), I believe, is “structure vs. continuous change”. When we try to organize (informational or other type of) reality we tend to impose structure onto it. The reality, however, is continuously changing. A solid “top-down” structure has poor chances of long and happy life. However, organization is impossible without structure...Irresolvable paradox?
Not really. Why not apply the same principles as we would use to conceptualize the reality to information modeling? Any reality has various aspects: main concepts or “notions”, relationships among them, processes, conditionality or rules, forms of manifestation or interaction with other realities etc.
Each aspect can be further stratified into the layers that represent:
- the core, or something that is the quintessence and, therefore, is not likely to change often (e.g. a model for the statement: ”This company manufactures products from parts/resources obtained from suppliers and sells its products to users“);
- main entities (e.g. types representing products, suppliers, users, phases of manufacturing process, etc.) and relationships among them.
- defined attributes (factual data) related to the business entities;
- relevant information defined only by a generic type (e.g. “document”, “folksonomy tag”) relevant to an instance of a business entity type.
The more stratified approach we use for information modeling – the better chances for a “long and happy life” for our aggregated info model.
Semantic technologies represent a high level of abstraction, i.e. “model for modeling”, and thereby enable flexible modeling at the higher (core & business entities) levels and aggregation of all aspect- or strata-specific models. To a certain extent, a similar effect can be achieved in a relational database (via highly denormalized tables) or XML Schema based modeling, however, processing of such models with existing technologies would be highly expensive, while semantic technologies are optimized for handling such highly “denormalized” constructs.
I guess, I should stop at this point not to abuse the posting space.
Would be glad to answer more questions and – especially – to read your thoughts on the above.
Originally posted 28 months ago.
(permalink)
Igor_I edited this topic 28 months ago.
|
|
My name is Bob Griffin, and I am a full-time faculty member who teaches Information Architecture for the Web Design & Interactive Media Department at The New England Institute of Art. I joined forces with Dorothee Shamonsky and Kay Aubrey (who were the Interface Design and Usability adjunct faculty, respectively) to define the interaction design curricula at the college. The reason that we felt it necessary to get together and define the interaction design curricula was due to our student population complaining about too much overlap in the three separate areas of study.
It’s my feeling that Information Architecture, Interface Design and Usability are separate areas of study, and yet they are also three peas in a pod. You have to have one to help define the other. Trying to talk about one without some semblance of understanding of the other two, is like trying to describe the Three Stooges by only talking about Larry.
Please take a look at what we came up with below. There will always be agreement to disagree — but in the long run — after we did this analysis, our students stopped complaining.
Check our documentation and diagram on: bg411.aisites.com/ and see the link under WDIM390 for EXPLAINING IA
Thanks!
Posted 28 months ago.
(permalink)
|
Would you like to comment?
Sign up for a free account, or sign in (if you're already a member).
|
|
Explain IA /
Discuss
Feed – Subscribe to Explain IA discussion threads
