Flickr, How About a Spell Checker and Thesaurus For Tags?
Here’s the thing about tags and search. The value of a human element that filters search results through tags is immense. There are two potential problems with tagging and the broader role that it will play in search going forward though.
The first problem involves malicious or misleading tags. In the future these may take form for two primary reasons, evil and greed. Think of the future attack on tags as the equivalent of today’s viruses and spam. At present tags are fairly new. They are used by most primarily as a personal way to manage and organize something, bookmarks (del.icio.us), photos (flickr), etc. But as tagging turns more and more into something of a larger and larger group undertaking the potential for evil gets greater. Whether for attention or just plain mischief, as tags become a catalyst for activity at some point when search becomes more integrated with tags we are bound to begin to see this.
Similarly, as commercial application begins to unfold in the world of tagging and search, economic incentives will arise for individuals to disrupt the natural evolution that might take place by introducing manipulations into the world of tagging.
The good news about the viruses and spam of future tagging is that there are really smart people, smarter than me for sure, who will figure out ways to deal with these issues in the same way that we’ve somewhat figured out how best to deal with viruses and spam today.
The second issue with tagging that comes up a lot is the concern that by turning tagging and organizational control of data over to individuals, uniformity goes out the window. Further, errors take place (misspellings, made up words, etc.) and you have a whole hodgepodge of a mess. To this, I say it doesn’t matter. Statistically errors are insignificant. If 2% of the time people misspell “San Francisco,” the important thing is that 98% of the time they don’t. Heck, as long as people get it right half the time and an effective search and rank algorithm exists to sort the correctly spelled tags you are more than good. Really all we need is something that can pull up the top 5% of a search term anyways. For most people this is enough.
So even though it doesn’t matter that error tags exist for the group, doesn’t mean that as an individual I don’t want my tags to be more correct.
Which brings me to my point (sorry for getting there in such a long winded round about way). At present Flickr has the ability to wholesale replace tags. If I accidentally spell “Oakland” “Okland” I can tell Flickr to change all of the “Okland” tags to the correct “Oakland.” Of course I have to know about it is the thing. I’ve caught myself making a few misspellings over time in my tags and corrected them — but as my digital photo library gets larger and larger on Flickr I know that there are now and will be even more in the future.
So how about this. I think it would be a super easy app to write. Give me a spell checker as an advanced tool that will go through all my Flickr tags and spell check them for me. Sure. There will be many purposely misspelled words and identifiers that have personal meaning (sflickr for instance) but help me fix my other misspellings where I want to. Build an add to dictionary function by the way so that I’m only reminded that sflickr is not a word once. By the way while at it let me run my descriptions and even the comments that I’ve written through the spell check app as well. Sure many don’t care about misspellings but for the truly anal it will be great fun.
Further, let me improve my tagging. With clustering results let me also run my photostream through — well not exactly a thesaurus but you get the idea. If I’m tagging a photo “sky” and others that use the same tag also tag sun and clouds or cielo, etc. Right there next to my photo throw out a list of recommended additional tags. Sure I could do this manually, but give me an advanced tool to make it easy for me to make my tags more rich and complete. In the end the group benefits a tiny bit but I benefit a lot. flickr, tagging, folksonomy, del.icio.us