Tag me as
January 17, 2005 7:35 AM   Subscribe

Rebirth of the Semantic Web. On the heels of the Technorati taggregator, the Oddiophile bookmarklet, the tag search (new today!) and much ensuing buzz, Jeff Jarvis brings up people tagging. This concept drove Friendster and FOAF, both of which petered out. But with Technorati's elegant synthesis of photo, link, and post tagging, the web may once again tap into networked individualism.
posted by NickDouglas (23 comments total)
LiveJournal is probably the most succesful actual use of this. Go figure.
posted by smackfu at 8:16 AM on January 17, 2005

I don't see a way to get an RSS feed on buzzmachine. Isn't that weird? Surely it's there and I just don't see it. There's no little orange button, though.
posted by bingo at 8:27 AM on January 17, 2005

Firefox often misses feeds. Try using Bloglines, which provides a bookmarklet for rooting out RSS feeds.

On preview: Found it.
posted by NickDouglas at 8:53 AM on January 17, 2005

It seems a bit churlish to talk about tagging without mentioning Flickr and del.icio.us, which have really popularized the concept.

I love Flickr and del.icio.us, and tagging is great. But - and maybe I won't be the first to say this but I haven't seen it mentioned anywhere else in the current love-in - this won't scale. Not the same way that Google currently does. Current tag aggregators rely on the tags being "good", and that no-one has tagged the goatse image with "child birthday cake". They also rely on being able to kick off a user who does that, because - prove me wrong here - the sources of tags are all registered users.

So, as far as I can tell, anyone who thinks this will scale to the entire networked world (current population over a billion and counting) is smoking crack. But as many of the people talking this up are the same old tired crew of techno-utopians I am not that surprised... I am just counting down the seconds until I see a blog post that says "Dude... what if you could tag podcasts? That would like, be so coool..."
posted by pascal at 9:05 AM on January 17, 2005

NickDouglas: Thanks for the feed. Unfortunately, unless I'm doing something wrong, it looks like I'm not getting post headers in my 'live bookmarks,' just dates and times, and the times are all 00:00:00 -0500. Strange (?).

Isn't it weird that firefox 'misses' RSS feeds, given that it's the browser that is arguably most RSS-compatible?
posted by bingo at 9:28 AM on January 17, 2005

bingo: Try this feed. Firefox detects that on the front page of buzz machine, but not on the archive pages.
posted by monju_bosatsu at 9:51 AM on January 17, 2005

Isn't tagging people using a link with the rel attribute just XFN (which is what Mefi contacts uses)?
posted by revgeorge at 10:52 AM on January 17, 2005

I hate to keep sounding like a luddite, but I don't know how to use that feed that monju_bosatsu is proffering. I want to feed into Firefox live bookmarks. That's a downloadable file....
posted by bingo at 10:54 AM on January 17, 2005

BTW, my first paragraph was maybe a little off, as Nick was talking about tagging people, not tagging in general. But I stand by the rest, even if it's a little OT.
posted by pascal at 11:13 AM on January 17, 2005

Also, I assumed Flickr and Delicious were sufficiently implicated in the Taggregator page, as they make two thirds of it.
posted by NickDouglas at 11:55 AM on January 17, 2005

I want to feed into Firefox live bookmarks. That's a downloadable file

Firefox insists on downloading RDF files, I guess because they're not served as text/xml or whatever, so you'll have to create a bookmark manually. Personally, I think the bloglines suggestion beats trying to cram it into Firefox.
posted by yerfatma at 12:11 PM on January 17, 2005

Create a bookmark manually...from the RDF file?
posted by bingo at 12:12 PM on January 17, 2005

So, as far as I can tell, anyone who thinks this will scale to the entire networked world (current population over a billion and counting) is smoking crack.

Guess I'm on crack. Tagging (or folksonomies in general) works best the more people do it. Yes, a goatse site or blog or photo could get spagged (spam tagged), and that might carry weight if it was the only tag for a given site.

But as more and more people tag resources it becomes easier to find the popular tags and weed out the outliers. Such a process needn't care *why* such a given tag is so rarely applied to a resource, only that it is.

Filtering systems that took into account who did the taggng, how may other resoruces this person has tagged, how well these tags accord with other people's tagging, and so on, will make it tedious for spammers to bother.
posted by Ayn Marx at 2:53 PM on January 17, 2005

There are already problems with the tagging systems, which our own Anil blogged about today in response to Rebecca Blood's posting. Key quote from her:

[F]or a few hours this morning the most recent tagged photo under MLK was a picture of a protester's sign that read "Setting aside our differences to focus on our common goals: peace, love, harmony, killing Jews, and tolerance." Nice.

I like a good utopian ideal as much as the next Filterite but somehow doubt that these mechanisms will not avoid falling prey to the same evils as most other automated web propositions (email spam, comment spam, IM spam, blah blah blah spam).
posted by billsaysthis at 5:39 PM on January 17, 2005

Ayn Marx: Yes, you are on crack. It's not like no-one ever thought of using the content of the META tag to index content before. That's what it was invented for. There is a reason people stopped using it.
Billsaysthis: thanks for the link... sadly, I have to agree with you.
posted by pascal at 6:12 PM on January 17, 2005

Good point. But my gmail lets only two or three spams into my box each week, WordPress blocks my comment spams 100% so far, I stopped getting IM spam when I stopped chatting in sex rooms, and I can put myself on a "do not call" list. We can keep the kipple down. In any case, you've named three technologies that got spammed but are still going strong. Do you think anyone will stop e-mailing soon? Neither will they stop tagging.
posted by NickDouglas at 9:44 PM on January 17, 2005

Nick, the reason you get very little spam is because gmail does not rely on any assertions that the sender makes about the content, it just relies on applying rules to the content itself. Given that tags are by definition assertions made about the content of some item, I think your optimism is a little unjustified.
posted by pascal at 12:15 AM on January 18, 2005

Ayn Marx: Yes, you are on crack. It's not like no-one ever thought of using the content of the META tag to index content before. That's what it was invented for. There is a reason people stopped using it.

Good point, but wasn't the problem with the META tag that webmasters were applying false tags to their own sites, in order to attract extra traffic?

But if you get a whole lot of people tagging content, it reduces the possibility of this happening. Sure, someone might host goatse on their site, and tag it as 'cute bunny', but all (or nearly all) other people will tag it correctly. So the consensus opinion should hopefully outweigh the minority spam, especially if filtering systesm such as Ayn Marx proposes are put in place.
posted by Infinite Jest at 12:44 AM on January 18, 2005

Pascal, the discussion here should be about the current use of tags. No one recommended having senders tag their e-mails. The uses so far discussed are in:
  • images
  • links
  • blog posts
  • people
The first two have been widely used for a few years now; the third will rise quickly because the early adapters have the incentive of extra site traffic. The last may or may not work, and I never promised it would. My "optimism" is merely in pointing out that it's as feasable as the other technologies we've implemented since '93.posted by NickDouglas at 5:40 AM on January 18, 2005

The notion of "service autodiscovery" is starting to take root in libraries. Slightly more here. Services like CiteULike are interested in this and it's going to be critical for internet localization. More here. [last two links are self-referential]
posted by Human Stain at 9:15 AM on January 18, 2005

Nick: I didn't suggest tagging e-mails. You brought up e-mail because you suggested that the success of gmail filtering your spam was an indication that filtering could be used to avoid the problems of abuse of tags. I don't believe this is true for the reason I have already stated.
As I mentioned in my first comment, tags have been used successfully for images and links in cases where the contributors of the tags are registered users of the system - basically the tags are trusted because the users are trusted.
As Infinite Jest mentions, one way tags could be trusted in the absence of a registered user mechanism is if the tags are not applied by the people who own the content, but by the consumers of the content. This is actually an important difference between the way del.icio.us works and the way Flickr and the Technorati systems work. But ultimately, how different do the results end up being from Google's PageRank? PageRank operates on the same principle, just using linking instead of tagging. To vote, you just have to link - and you don't have to say why because Google can make a pretty good guess.
BTW, the "incentive of extra site traffic" was exactly what killed the meta tag in the first place...
posted by pascal at 9:28 AM on January 18, 2005

Sorry, yes, we got bogged down about the e-mail tags. I was first responding to your example of e-mail as a technology overridden with spam. I only sought to counter by negating this assertion, demonstrating that e-mail works despite spam. Similarly, tagging can work despite spam.

Note the population who cares about tags. This is not a random sampling of the internet population; a particularly savvy, young, and heavy net user reads blogs. This is why comment spam has not reached the prevalence of e-mail spam; it is easily blocked because its victims are adept with the technical tools to counter it. Del.icio.us uses minimal registration, and a person-tagging system with such minimal registration could be easily implemented.

Indeed, page rank is a tagging system, both consumer-run and, among smart creators, creator-run (you've seen the sites teaching how to raise your own pagerank). And it's seen abuse, as every system has. But the system has worked better than any other search system, giving Google the majority of all searches conducted online.
posted by NickDouglas at 10:50 AM on January 18, 2005

OK, so here's my last word: the reason Google works is because the ranking is based on the content of a page, and the content of the pages that link to it. The content, not any direct assertions that people make about the content. The way PageRank has been improved to deal with abuse is by becoming smarter in understanding the content in order to differentiate good content from abuse. Without trust, metadata is useless, you have to go with the data.

Remember, my original point was not that I don't like current tagging systems - quite the opposite, as the amount of time I spend on Flickr will attest - but that I don't believe that tagging, or the semantic web, or whatever, will scale in the way that Google has.
posted by pascal at 11:29 AM on January 18, 2005

« Older Does this FPP make you want to be a hooker?   |   Why is Red Peters not famous Newer »

This thread has been archived and is closed to new comments