Do People Like Hierarchy?
This is a question I routinely ask myself -- especially in my Cognitive Psychology course at UMO. Often with computers, we are asked to put things in hierarchy via folders. A file can only exist in one folder and one folder only. A few classes ago, we talked about the root of this idea which might be linked to how humans cognitively process data. I was flabbergasted, according to the theory we were studying, people do like hierarchy! Well, as it turns out, the theory (Quillian's TLC Model) is based on a software program written to understand human language in 1969 -- a computer programmer helped found this idea. Ugh!
Well, in my opinion, programmers often tie in their programming experience with other ideas assuming that it is the best fit, since it is scientifically the best idea. This is often the case with user-interfaces, and I am not immune to this disease. It would seem as though the theory was mostly based on Object Oriented Programming (OOP), where the programmer sets up a hierarchy of "classes" and then uses those to control the properties of "subclasses". This leads to very readable code that often makes lots of sense and is very clean. Does this programming model necessarily reflect the way our mind works? I don't think so.
I've often thought of our minds as more like a relational database, where we make loose connections between nodes and there is no hierarchy. We just look for things that relate to the item in question. For instance, when being asked the question "Is a canary an animal?" we don't go from the canary node in our brains, up the hierarchy to the "bird" node, and then up again to the "animal" node. We have a direct connection between canary and animal. Just like we have a connection from canary to feathers. It just makes sense to us.
The same can be said for filestructures. If humans don't really visualize or process data in hierarchy, and they only accept it because that is what they're used to, shouldn't we be coming up with a new concept of information organization? One that is perhaps based more on the psychology of memory and how people *really* think rather than the basis of a computer program?
More on this later. Time for Psych class. :)
Well, in my opinion, programmers often tie in their programming experience with other ideas assuming that it is the best fit, since it is scientifically the best idea. This is often the case with user-interfaces, and I am not immune to this disease. It would seem as though the theory was mostly based on Object Oriented Programming (OOP), where the programmer sets up a hierarchy of "classes" and then uses those to control the properties of "subclasses". This leads to very readable code that often makes lots of sense and is very clean. Does this programming model necessarily reflect the way our mind works? I don't think so.
I've often thought of our minds as more like a relational database, where we make loose connections between nodes and there is no hierarchy. We just look for things that relate to the item in question. For instance, when being asked the question "Is a canary an animal?" we don't go from the canary node in our brains, up the hierarchy to the "bird" node, and then up again to the "animal" node. We have a direct connection between canary and animal. Just like we have a connection from canary to feathers. It just makes sense to us.
The same can be said for filestructures. If humans don't really visualize or process data in hierarchy, and they only accept it because that is what they're used to, shouldn't we be coming up with a new concept of information organization? One that is perhaps based more on the psychology of memory and how people *really* think rather than the basis of a computer program?
More on this later. Time for Psych class. :)
posted by Matt James at
10:00 AM
2 Comments:
Your canary example is kinda interesting, since it relates to the scientific classification system for lifeforms, which is a hierarchy.
I'm surprised you didn't mention tagging that's been catching on lately, with notable examples being del.icio.us and flickr. I think tagging systems have a lot of problems, but it's just catching on so it probably just needs time.
I've been frustrated with the folder system for a while. Sometimes I've thought about using shortcuts as a way of putting one file in multiple folders (and *nix hard links would be nicer way of accomplishing the same thing), but it would be hard to manage everything that way I think.
I'm having more hierarchy type issues in planning out how I'm going to tag my music files. I want to have multiple artist tags for songs with more than one artist, but most software doesn't support this very well. For example, I have a file tagged with both "Enya" and "Howard Shore" that in WinAmp's library doesn't appear under either artist, but under a separte artist called "Howard Shore; Enya".
If you want to find some good articles about the concept of tagging, here's a good place to find stuff:
http://del.icio.us/tag/tagging
This one especially relates to what you've written:
http://www.blumpy.org/tagwebs/
By
Doug, at 3:28 PM
Sorry I've gone so long without replying. I skimmed your comment when you first put it up, but I've just now found the time to respond.
I think that tagging has its possibilities as well. Not only does it break from the single folder, multiple file paradigm, but it also applies more metadata to the file in question.
My only problem with tagging is the way that you tag things. From my experience, the interfaces for applying tags is often lacking. For instance, if you want to apply a tag to multiple files, you need to have a way of doing it that makes sense -- like dragging a tag onto a group of selected files. It would then be good to see some kind of state-change in the files to show that they had been tagged correctly -- a method of feedback. It's been my experience that software developers might use tagging in their applications, but not follow through with the idea of tagging in the interface, if that makes any sense.
I'm about to make another post about this same subject, so you can see where I was headed in my last one.
By
Matt James, at 6:12 PM
Post a Comment
<< Home