Kyla’s Reflection
I found examining organization from the perspective of the digital records rather than the physical items on shelves to be very enlightening. Initially, I thought it would be very simple to include a search function. Wordpress has a search widget, so all we would need would be a simple front page with a search box, and all the pages hidden, ready to be searched for. However, we quickly discovered that a browsing function would also be crucial. Having added this as a less important information seeking option, I thought it would all work out.
The search widget we had available from WordPress searches all fields equally, including the description. This seemed to me like it would be excellent, because we wouldn’t be limited to the resources that mention the search term as a subject, but anything that may be related: a much more flexible, up-to-date, Google-like way to organize information. I was quickly shown, however, that this assumption was incorrect, and that there was a lot more to take into account. For instance, we started by adding our MARC records as pages, but this clogged up the search results. We had to go back and attach them as PDFs instead, to keep that data hidden from searches.
Additionally, even with so few resources, some search terms (such as plants) brought up almost every single one, with no sense of relevance or importance: they may as well not have been organized at all. We realized that the browsing and hand-sorted categorization were much more effective tools to find information that the simplistic search, which is better suited to finding specific resources from title or author name. I learned that this task is much more complicated than I thought. In order to use the search function as the primary information seeking tool, we would need to be able to edit the search widget. We would need to be able to control which fields are searched, or to develop a complex relevance algorithm, like Google has to . Even with our small number of resources, I learned a lot about the challenges that come from having too much data.
In the end, we developed a system that we think will work best. The MARC records are hidden from searches, the reflections and explanation would be removed before use, so as not to clog search results. Also, the hand-sorted browsing options outlined above will be the primary information seeking tools. By going through the process, I was forced to question my assumptions, which changed the way I think about these ideas.
No comments:
Post a Comment