Posting to Autocat
On 21/06/2013 18:06, J. McRee Elrod wrote:
While my messages are much more frequent that James’, my impression is that my individual messages are much shorter. Like this one.
According to my calculations (which could be incorrect but I don’t believe so), this is what I have for numbers of lines in messages for the top five posters of 2013:
|Aaron Kuperman||Frank Newton||J. McRee Elrod||James Weinheimer||john g marr|
|Total Number of Messages||146||68||284||108||123|
To be honest, I have posted relatively little lately, but the results surprised me as well. Probably if it were limited to the first three or so months, the results would look different. Still,I hope none of this changes what anyone is doing (it won’t change what I do!), and I hope that Mac and the others post as much as ever.
This illustrates what I think are some of the subtle dangers when working with metadata. While you can manipulate metadata up and down, in and out and over the moon, and while I agree it is “data”, I do not believe that in most cases, the result of this manipulation creates “information” and even less is it anything that approaches “knowledge”. The danger is that after all of this manipulation, somebody might actually think they do have some “information” that can serve as bases for decisions. No one should draw any conclusions from any such manipulations. The reason I say this is that the validity of much of statistics is based on “randomness”. For instance, you can do a statistical analysis of blackjack or poker and everything can work out rather well–so long as everything remains random. But the moment you get an active human mind involved, such as a card shark, your statistical analyses get thrown out the window–and if you are fool enough to persist in following your analysis against a card shark, you will soon find yourself sitting in nothing but your underwear. Here is a great scene from the wonderful movie “The Sting” http://www.youtube.com/watch?v=n3bnMv3ULes The hands in the movie clip are not Paul Newman’s but of the great John Scarne. I would not want to play in a card game against Scarne, no matter how good I thought my statistical analyses were!
In the same way, what I write is not a random process (although I realize that some probably think it is!). I don’t think any of my colleagues above write things randomly but what they write, they do so in a thoughtful way. I want to read what they wish to share with us. Shakespeare certainly did not write at random. Therefore, any statistical analyses of his writings must remain highly dubious. And consequently, putting too much faith in statistical manipulations of many other types of metadata and its manipulation, hold similar dangers.
Does any of this concern cataloging? I think it does, since catalogers are supposedly supplying much of this “metadata” and the future of RDA/FRBR/BibFrame/Linked Data is for the public to “manipulate” the data/metadata that we create. I do not believe that catalogers can ethically ignore these sorts of concerns. And for differences in what IT people believe is metadata vs. what catalogers believe is metadata, I discuss in my podcast at http://blog.jweinheimer.net/2013/01/cataloging-matters-no-17-catalog-records-as-data.html
In sum, what I gave out in my previous post and the numbers here are “data” but I do not believe it is information, and certainly not “knowledge” or anything that is really useful. It may be interesting for a moment, but that is about all.