This week we have looked at the different digital databases and information retrieval systems. Information retrieval systems are defined as the process of representing, managing, searching, retrieving, and presenting information.
Databases:
We have looked at Relational Databases (RDBMS) which according to the BBC – GCSE bitesize website list the following as advantages:
– The book’s details and the customer’s details need only be entered into the database once.
– Because of this, mistakes are less likely to happen and if there were a mistake in a customer’s record, for example, correcting it will correct the mistake database-wide.
– Duplication is avoided – this keeps the database’s file size down.
– Details about books and customers are easily accessible using their unique IDs.
– Queries can be performed and reports generated, eg a list of books a customer has borrowed since joining the library.
Information Retrieval:
Recall = proportion of relevant documents retrieved.
Precision = defined as the proportion of retrieved documents which are relevant.
The most widely used metrics for information retrieval are recall and precision, measuring respectively the success of a system in finding all the relevant material that there is to be found, and in returning only relevant material (Introduction to Information Science – D. Bawden & L. Robinson, p.153).
It has been interesting to read more about the different ways in which not only databases are created and function but also to consider how people search for information. The different categories of image queries in Image User’s Needs and Searching Behaviour by Stina Westman (p.6) were especially familiar to me in my job as a graphic designer. That we “differentiate between searches on the three semantic levels: general, specific and abstract (Jörgensen 1999). General searches consist of finding images of a general topical or subject category: a kind of person, group, thing, event, location or action. Specific searches consist of finding images of an individually named person, group, thing, event, location or action, that is, a specific instance of a general category. Queries may be designed to find types of objects (e.g. computer) or specific objects (e.g. Big Ben). An abstract search involves finding images that communicate certain emotional or abstract concepts (e.g. warmth). Image queries and selections may also be based on the moods and emotions which could be associated with the image or elicited in the viewer (e.g. sadness).”
Image searches or queries became almost automatically the first place I would begin when looking to communicate a visual design solution. In part I would search image libraries or even google for abstract queries such as for example: pride. Not because I don’t know what that might look like to me but mainly to see how other people (the people who had categorised the images in the image libraries and general people on the internet) would have tagged or categorized particular images.
Also I would look to image libraries to jog me when I was stuck or had drawn a blank. I didn’t always expect the perfect image to be there in the results of my search but I did hope that it would spark something in my mind. Something that related to my query or opened up another way of thinking about all the things which that particular query suggested. As Stina Westman notes (Image User’s Needs and Searching Behaviour, p.8): “Some users may not want a specific image at all; instead they want to browse a collection for inspiration or ideas.”
I think many people are now so used to searching in different kinds of image databases such as Google or Getty Images for example, that people become aware of any limitations and adjust their expectations. As Stina Westman points out again that (Image User’s Needs and Searching Behaviour, p.7) “describing images for retrieval, i.e. querying, seems to result in more semantic terms and less syntactic attributes than free description of image content (Hollink et al. 2004b; Jörgensen 1998). This may be due to users adjusting their image needs to the perceived capabilities of an image search engine.”
This goes to confirm why I had been doing what I had been all this time. Which is to see the way other people perceive things. What meaning or natural associations have they attached to an image. How would other people describe images? But also it has made me think more seriously about the chicken and egg of digital searching. Are we searching this way because this is how we know the search systems function or because this is how they have received information from us and our own instinctive ways of categorizing?
Laine-Hernandez and Westman (2006) found that constrained image annotation resulted in more terms depicting the story, setting and theme of the image than a free description task, which led test subjects to enumerate individual objects and describe their locations within the image. Imposing either a retrieval system or a format of description (preselected terms, limited number of terms) on users may change the way they describe images.”
When considering the assumptions we make about search systems as discussed by Westman above. I found the following quote (What does the internet have to do with my library? – T. Ballard) sums up the trap that most of generation Y have fallen into: “The rise of search engines such as Yahoo! and Google gave users such a wealth of data that it led to the false idea that all worthwhile human knowledge was already online.”
Lisa Baatsvik-Miller DITA14 