Hi Folks,
Usually I lurk on this list, but I couldn't let this one go. I worry that the incessant bashing of Google and Wikipedia is becoming a professional suicide pact. Very few things damage professional credibility more than looking at a solution that works for its purposes, then claiming that it doesn't.
Steve Paling
On 07/29/13, Karen Coyle wrote:
> Google rides on the backs of content providers, and then turns around and sells them to advertisers. It's hard to imagine NOT seeing that as evil.
I have a hard time NOT imagining that as GOOD. You're the editor of a small literary magazine. Good luck getting a print magazine on the physical shelves of MegaBooks, Inc. Good luck getting good distribution through literary readings or a table at a farmers' market. Good luck getting noticed through the catalogs of the tiny number of libraries that will acquire your magazine. But if you are part of a group of magazines that link back and forth on the Web? If you do away with the cost and hassle of print, and put all your content online? Give it away, because you're not making money on the magazine in most cases anyway? Google's algorithms will pick that up. Your potential audience is larger by orders of magnitude. In what way is Google riding on your back? Of course they make money on ads. They can't operate on that scale for free.
> Yet none of this "organizes information" in the sense that library classification intends to. The Google motto of "Organize the World's Information" is insulting because THEY aren't organizing anything, they are running algorithms over text.
How is running an algorithm not organizing anything? The algorithms were written (almost exclusively) by people, and those people made decisions about what principles should guide retrieval, ranking, etc. How is rendering massive, emergent patterns in a usable form not organizing information? So what if Google's algorithms don't interpret the content of images, and rely on text instead? Do catalogers interpret the meaning of individual illustrations in books? Google may rely on the text surrounding the image, but that's more guidance than our bibliographic records typically provide. It's an odd phenomenon in this field that people will refer in a disparaging way to running an algorithm. Learn to program. Build your own indexes. Do full-text analysis. Manage massive server farms. Then come back and try to tell us that an algorithm doesn't organize information. Besides, LCCS, LCSH, DDC, RDA, etc., are just print algorithms rendered in digital form.
> There are no interrelations between texts in Google's universe,
Actually, links between sites are a key part of Google's algorithms. How is that not interrelation? Major Web sites tend not to contain random links. How often does an article on, say, global warming, randomly link to a site on Katy Perry and Russel Brand just for the hell of it? Random ads on the page don't count, because a good algorithm can ignore those links with decent reliability. How much evidence (NOT anecdotes) is there that FOR USERS our subject analysis works any better?
Received on Mon Jul 29 2013 - 14:06:20 EDT