> *blinks* when I view the page it still has a > tagcloud of common tags in the > top-right... >
The data is not complete and the historical data is missing. I shall explain more later. > > Joshua once wrote - admittedly in a bit of a > different context (viz. in > > explaining that the "/post url allows only a title > and url ... [in order] to > > prevent XSS hacks," i.e., lest "publishers" use > the feature to "suggest" > > tags): > > > > Tags are a memory aid, decided upon by the person > doing the remembering. > > > If some other person with intentions other than > recall choses the tags, then > > > they are indeed improperly chosen. > > > > > > > If tags are "memory aids" - and memory aids only - > why bother displaying > > *all* of the tags from each user?... The "Common > tags" ought to be good > > enough to get the mnemonic juices flowing. Tags as just 'memory aids' is a very superficial point of view. It might have been true when Joshua wrote it, but now no one believes it anymore. Delicious was a succesfull experiment, and when it started no one realised the importance that the tags would have reached. It would have been like imagining the taste of a fruit by seeing a seed of the plant for the first time. Joshua might have said that they were just memory aids, but later he said that people pay their acces to the site by tagging things. Why would he be interested in having those tags if they were only 'memory aids'. On the 25 of May 2005 I showed how mathemathically your 'memory aids' could be used to compute a distance between web pages. http://blog.pietrosperoni.it/2005/05/25/tag-clouds-metric/ Four month before 'Ben Hyde' had proved something that many people had been suggesting from a long time. That is that tags followed a power law. He actually calculated the steepness of the power law. http://enthusiasm.cozy.org/archives/2005/01/tagging-powerlaw/ Of course he did this by screen scraping, because this kind of info was not at the time given through the API (and still isn't). Much research have been done in the meantime to study folksonomy, and in general the behaviour of your 'visual aids'. This research was generally done either directly screen scraping, or using websites that were screen scraping. The whole geek community got its fruit, as new web sites would appear using tags. Joshua, didn't all this helped you at least a bit in understanding the depth, and the importance of tags, and thag clouds? Well this research was only possible BECAUSE delicious have been offering the whole historical data of the tag cloud. By screen scraping you could not only see the tag cloud of now, but what was the tag cloud in every moment. And people who were screen scraping, yes, against the guidelines, but always telling to Joshua, and following the guidelines were giving back to the community. We would use delicious, delicious would gather the data, the data would then be offered in html format and researchers would use the data to study the tags, tag clouds, and folksonomy and would feed back in the community their discoveries. Three days later from the previous post of tag clouds metric, Terrell Russell (who now shared his frustration with me via email) made a web site: Cloudalicious. Cloudalicious could show how the tag cloud would change as the days would pass. By studying this I could MEASURE HOW SOCIETY WAS CHANGING. The dream of many sociologist, and we had it available on a golden plate. I presented my work the morning after in another entry that also became very popular: http://blog.pietrosperoni.it/2005/05/28/tagclouds-and-cultural-changes/ In this I showed how you can see the changes in the society as they are happening, and indeed measure them. I took as example the tag ajax. All this was only possible because cloudalicious was offering the data. And cloudalicious was offering the data because delicious was making it available. Delicious stops it, and a very promising line of academic research dies. The line got his first academic paper (that I know of) from Golder and Huberman, on the 18th of August. They even studied the tag clouds of every url that came out from delicious popular in a period of 4 days. http://arxiv.org/abs/cs.DL/0508082 They made some very respectable work. Reproduced some previous results and proposed reasons why tag clouds would follow a power law. This was of course only possible because delicious were permitting people to have the whole tag cloud for each single URL, and indeed the whole historical data of the tag cloud. You might ask yourself why are people so keen on using delicious respect to, let's say Flickr. It is not only that delicious is bigger than many other system, it is also that the information was more readily available and more cleanly kept. In delicious every user had a tag set of each url, and each url ended up having a tag cloud made up of all the tag sets. I will not enter here on why this is important, please refer to my first blog entry above to see the details. For now believe me, it is fundamental. The fact that people used delicious in their example all the time meant a huge free advertisment for delicious. An advertisment that would generally reach a different community from the geek community from which delicious was born. And an advertisment that delicious does not seem to value much anymore. Many many people used this information. Saying that only few people were using it totally forgets that many people were using cloudalicious, and that needed that information. And many people were reading the results from the people making the research thus indirecly having a benefit from all the data being offered back to the community. It is like saying that since only 3% of personal computer run linux than linux is unimportant. Totally forgetting that 99% of the people uses google, and google runs on linux. This line of academic research is not dead... yet. 3 weeks ago I was giving a presentation at the university on bottom-up onthologies. I contacted Clay Shirky for the occasion. In that occasion he told me: "I've been giving a talk pointing to your 'little social quake' graphic and trying to convince everyone that this is the most important view of tags currently going." My advisor looking at the results asked if I shouldn't aim to make some serious research for a Nature publication or similar. If you are in academic you know that for many discipline Nature is the top to which many people aspire. Above: the nobel prize. He actually asked if I shouldn't write a research proposal for the EU once the PhD was over. People writing academic papers; studies being done; people inspired by those studies starting new web sites; research; ideas; the possibility to calculate the distance between url; the possibility to investigate the long tail of the web through that; the possibility to study social changes through studying tag clouds. Delicious getting free advertisment from all those studies. Advertisment reaching way beyond the geek community it started (many professors at my talk never heard of delicious before). And you are being serious in saying that this data is unnecessary to the majority of people. Well, until the majority of people don't enjoy the fruit of the research being done with it. I asked Joshua 6 times now to make an API that delivers that information. He NEVER answered. Not yes, not no. Just ignored the message. The last was the one from yesterday evening. It got ignored, again. > PS - Last thought: There's probably a Greasemonkey > solution to this > > problem, eh?... Or not?... Could you not > reconstruct the old page using > > GM?... Sure by calling the page of each user who bookmarked a site. For the big site (which are the ones statistically relevant) this means calling a thousand pages for each url... plus the fact that for many user you will need to go back in time through some pages. It would be less abusive for delicious if we were to start spidering the whole site and just make a copy of the whole database. As a last element I want to point out that the idea of taking of the info because they wer cluttering the web pahe, and not really being useful is ungrounded. There are many information that we pick up from our environment subconsciously. The underline emotional tone of an email. (Could you feel that I was pissed off yesterday? Can you feel I am more calm, but more lcear today?). The expression of a face. This is analog information. It uses a higher bandwidth to digital information, but if you are able to use it it gives you way more information than the digital one. It is like a mouse respect to a keyboard. An ipod wheel respect to using 4 buttons to move through directories. The info that delicious provided was giving a lot of subconscious info to the reader. Look at the tags of a user. How often do you read all the tags? Never, right. Would you say that they give you no information about that user. Can't you immediatly see what are his interests, his way of thinking, his favorite topics? And so is the info about the URL, That info should not have been taken away, it should have been made complete, instead. And I totally subscribe that the name of the people tagging something gives no info. You can as well just put a number. Ok, it took me 90 minutes to write this email. If you read it all, I thank you. If you think I am write please add your voice. If you think I am not please tell me why too. And let's hope that it is really a decision that was coming out of Joshua, and he was not just obeying some directives from the Yahoo headquarters. Pietro __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ discuss mailing list [email protected] http://lists.del.icio.us/cgi-bin/mailman/listinfo/discuss
