"In real world applications... Millions of images are added every day (e.g.300 million images are uploaded to Facebook per day, with a total of 100 billion images), rendering these methods impractical even to index the daily uploads." 1CNET 08/2012, http://tinyurl.com/9jfs7ut
I’d already decided a block quote from one of my yet to read books and drafted something to work on – later.
I’ve had my twitter research and lessons I learnt from that session going round my head. I’d use the term ‘crash-tagging’ in my post parodying social media marketing. I decided to google that term and found my blog post was third result on the first page above a politics of hashtagging blog and a whole load of gamer stuff after that.
What I’d meant for crash-tagging (in twitter type use) is apparently known as hashtag spamming – but to me spamming implies direct communication aimed at individual user accounts. I prefer the term crash-tagging, meaning to gate-crash a #conversation with irrelevant or inappropriate content.
The internet’s an amazing way of getting lost and found on unplanned browsing and research journeys. I came across the PDF document download linked above and although I don’t understand all the terminology and certainly not the equations, I enjoyed reading it and considering some of the issues raised in that document.
I also find incidental juxtapositions of information intriguing – such as finding at the bottom of the left hand column on-screen, if you continue reading from the top of the right-hand screen some entertaining or unfortunate strings of text can occur. Of course that’s entirely taking original text out of context and changing its meaning but can be quite fun and screenshots can save those moments to return to.
I’m not sure how safe it might or might not be to start ripping microsoft downloads to shreds and representing such material in new form… perhaps it should be done without referencing the source…
I’m not at all surprised by the figures quoted in the document I’ve referred to regarding image uploads but wonder how many are duplicate images, how many are original new content, how many are pirated…
It’s also not surprising that over 50% of images on Flickr aren’t tagged. If a user doesn’t wish their images to be indexed for the convenience of search engines or other uses then they won’t wish to use tags. However the very nature of uploading to those services allows the service-provider world-wide royalty-free license to use any of those files any way they wish (though privacy policies come into play too I expect, depending on the nature of the images).
While tagging might not be important to some service-users/consumers, it will be an integral necessary function of the image-hosting facility to be able to index those files. The idea that this could be automated fascinated me.
Briefly.Until the next distraction…
(I wonder if it’s my theme that won’t allow a conventional block quote and will only present a scrolling text …)
#crash-tagging, #microsoft, #research, #meta-tags, #images, #indexing, #ArtificialIntelligence, #spamming, #terms, #S.E.O, #F.I.T,#S.U.A., #blockquote, #writing101