Does the world make sense? Or do we make sense of the world?

May 18, 2010 5:54 PM

Web 3.0 [vimeo - flash or HTML5] A story about the Semantic Web from Kate Ray.
posted by unliteral (38 comments total)

Web 3.0 - like Web 2.0, but less well defined and useful as a term and with a 1 added on.
posted by Artw at 6:01 PM on May 18, 2010 [2 favorites]

(Sorry to leap in with snark there, but "Web 3.0" is an instant snark button for me. "Semantic Web" produces less instant snark, but still induces the feeling that there is a high probability that the speaker does not know what they are talking about)
posted by Artw at 6:04 PM on May 18, 2010 [1 favorite]

We're not calling it that.
posted by jeffamaphone at 6:05 PM on May 18, 2010 [4 favorites]

Is this something I'd need to be a Social Media Guru to like?
posted by Jimbob at 6:07 PM on May 18, 2010

I like to think the internet will instead become more accessible for animals in the future. Think about it: There are trillions of mammals in the world, but only a small percentage of them pay for internet access.
posted by mccarty.tim at 6:12 PM on May 18, 2010 [7 favorites]

I've often thought about the fundamental disconnect between those people who believe all of the data out there should be accessible, versus those people who believe that enough of the data out there should be accessible, inasmuch as the former appear to want to have access to everything to make all the judgement calls themselves, and the latter appear to just want to get things done, and are willing to tolerate potentially imperfect data and filtering to get the results they desire.

In that sense, I always appreciate that there are folks out there striving for the perfect, ideal organizational method, because someone has to be shooting for the stars -- and inadvertently giving a leg up to the other folks, who find some small portion of the results to be useful enough to make useful and interesting things happen, whether it's the right or official thing to do or not.

I also find that, on the whole, the older I get the more I'm in the latter camp rather than the former. Having kids also pushed me further in that direction. So I'll just wait it out, content in the knowledge that no matter how long I do this web thing as a career (as others have been orators and scribes and publishers and town criers and made audio and video recordings and created telephone and cable and satellite networks and, you know, built cities and roads and culture) it will inevitably hit a point at which it becomes a ubiquitous part of our infrastructure and culture upon which the next great thing will take our fancy and help us to soar.

Arguably, we're already at that point, and just don't know it yet. Check back in a few years.
posted by davejay at 6:18 PM on May 18, 2010 [3 favorites]

oh, and not a bad little documentary
posted by davejay at 6:19 PM on May 18, 2010

Yeah.

Tim Berners-Lee

What a know-nothing.
posted by unliteral at 6:26 PM on May 18, 2010

When they were talking about ontology, I'm thinking philosophy. I didn't realize is has a separate meaning for information science
posted by bhnyc at 6:29 PM on May 18, 2010 [3 favorites]

I thought the semantic web was a result of the increasing crowdsourcing of Web 2.0, not a new model but rather the outcome of the current model.

(Also I'm in library school and the next professor who says "Library 2.0" is getting papers with the margins covered in hastily-scrawled marginalia expounding on the topic of category errors.)
posted by Pope Guilty at 6:49 PM on May 18, 2010 [1 favorite]

Dude, I've been rockin' to Web 4.0 since before it was cool.
posted by chasing at 6:57 PM on May 18, 2010 [1 favorite]

Does the world make sense? Or do we make sense of the world?

Yes!

Fun doc, thanks.
posted by carsonb at 7:00 PM on May 18, 2010

I'm running Web 3.6.3 on Linux 10.04.
posted by DU at 7:01 PM on May 18, 2010 [5 favorites]

I like how the Semantic Web folks think that the main criticism is ontology impedance or something.

No, no. Go look at some people's HTML. Do they validate it? Do they close their tags? Do they even pick the right tags? People are lazy. The Semantic Web depends on people using ontologies, not which ontology they use. Damn, look at the spelling. Most people just do not possess the urge to mark up their text to any significant degree. Now go look at some more HTML and get more depressed. Tweets. Wall posts. MySpace layouts.

Semantic Web assumes everyone will bother. That's why XHTML 2.0, if not dead, is on life support. The W3C must escape its ivory tower "I am sure everyone will want to catalog this, every jot and tittle" mentality or be relegated to just one of those curiosities of standards bodies.
posted by adipocere at 7:06 PM on May 18, 2010 [7 favorites]

Artw: “(Sorry to leap in with snark there, but "Web 3.0" is an instant snark button for me. "Semantic Web" produces less instant snark, but still induces the feeling that there is a high probability that the speaker does not know what they are talking about)”

It's still an interesting video, and worth watching. If only for the part seven minutes in, when people start talking about how "semantic web" is an overused marking blurb that's lost most of its meaning.
posted by koeselitz at 7:13 PM on May 18, 2010

I don't understand what the deal with this is. They've been harping on relationships since 1998 or whatever, but did any of them think of inventing Facebook or Twitter? No. Those just naturally happened, without much thought about semantic frameworks.

Obviously if these relationships are meaningful to people, services to mark them up will get invented, and in a large sense already have. I assume the point of this project is to make people think about how it should work in the future. But I don't see any user-friendly proposals in this video, nor have I in any writing about the Semantic Web.
posted by shii at 7:14 PM on May 18, 2010

All this talk about semantics and needing filters for your information. It's so meta.
posted by jeremias at 7:16 PM on May 18, 2010 [1 favorite]

I'm a big fan of semantic mark up. So are search engines, or at least Google is. We try out best but do make technical compromises here and there (namely how CSS handles the order of things). Sure, some developers don't give a shit, but I really don't give a shit about them, they do me the favour of shovelling their clients to me to fix the mess they made.
posted by juiceCake at 7:34 PM on May 18, 2010

Semantic Web assumes everyone will bother.

Yeah.

Big surprise that a basically technocratic proposal has some unworkable assumptions.
posted by kenko at 7:42 PM on May 18, 2010

I'm sure the semantic web will be a huge success once we make some AI smart enough to do the tagging for us.
posted by Joakim Ziegler at 8:00 PM on May 18, 2010

When I was young, back in 19... long, long, long ago, I did not have to walk five miles to school in the snow (it was Texas and I had a bicycle), but the world was brought to me in about twenty white bound books called the World Book Encyclopedias. I never watched TV without three or more open in front of me. Later some brown faux-leather books came into the house. They had all the information in the world presented twice - Micro and Macro. The whole known universe was my oyster.

Jump forward a few years and I'm at college. There is a huge wood paneled room there, Drawers and drawers and drawers of 3X5 cards, tables and tables piled high with this thing called the Guide to Periodical Literature. Wow!Eventually I found the room behind that room. It was a room full of weird machines with huge screens and little cranks or littler electric motors to read tiny little pictures of print and photographs' microfilm and microfiche. Wower. (I have a niece who is in a master of library science program. These tales have a bigger effect than the ghost stories I told her fifteen years ago.)

A professor (a good one, maybe once a household name in historian's and classicist's homes; I don't remember his name - it was long, long ago.) handed me a stack of pages (xeroxed, mimeographed?) - they were snippets and reconstructions of the Telegony; bits of a lost masterpiece and lost worlds. Another professor a few years later gave some sheets with reproductions of reports of assizes during the time Henry I so I could see the number of times men named Smith were found guilty of using "force and arms' to harm a horse's hooves. (Why? To show that negligence has always existed at law, of course, though threw a legal fiction. Duh.)

Some years later than that I was buying a new computer (not my first. I'd had one that was a word processor that could play "Zork" and chess.), and it had this bit called a modem and there were these free disk by the door from this stange company, AOL...

It's today, and I know nothing about computers (the on button, who to bookmark in firefox,... but really nothing.). I know even less about "Information Technology" and my google-fu sux. I do know I didn't completely understand this documented, I admit that. And I also know that information, every little tiny bit of it, is the most extremely precious commodity that has ever been known, more than gold or art. More than useful energy except where that energy gives us access to information. Information is mankind's greatest need and asset. What you don't value, someone will; what you can't use, may be used to build a machine or to win an argument that improves lives.

I also know that next to burning every scrap containing a certain scrap of data, the best way to destroy information is to bury it among zillions of similar bits. This is true even if each bit is as valuable as the next.

I don't need access to everything that is out there on the web perhaps, but I didn't need to know about suits about twelfth century furriers until I did. No one needs(!) Homer's third act perhaps, but wouldn't it be nice if we all had it?

And gatekeepers suck. Those cute boys, attractive girls and homely women that kept me away from the microfilm for almost two years sucked. People who demand that their platforms use only their software suck. The people that pile tripe and pilpul onto thousands and thousands of web sites - don't suck. That is all important! But I want my nieces and nephews to be able to get to what they want, to be able to learn everything - everything, even of the mistakes, and I think that is what we are talking about here, no?

So: Go Web 3.0. Go Semantic Web. Go, new searches, and to hell with the information technologist that think it is meaningless. It means a whole lot, even if it only means a dream.
posted by Some1 at 8:12 PM on May 18, 2010 [5 favorites]

Just as soon as they finish standardizing XHTML and RDF, all those semantic web entrepreneurs are gonna build useful stuff and get rich!
posted by Nelson at 8:27 PM on May 18, 2010

I'm a big fan of semantic mark up. So are search engines, or at least Google is.

You might be able to count the semantics of tags that matter to Google on one hand, and only one of them in itself probably has a weight comparable to that of the data mining done on raw text of a given document and documents that link to it in ranking the relevancy of a given document to a given search. And even that one tag is used to mean something beyond what the people who invented HTML intended it to mean.

Semantic markup seems like such a quixotic term for the very small set of meanings afforded by HTML and the even smaller set that's utilized in the context of search.
posted by weston at 8:50 PM on May 18, 2010

Semantic web?

<hamburger>This sounds like a wonderful new idea, and is sure to gain traction.</hamburger>
posted by UbuRoivas at 8:54 PM on May 18, 2010 [2 favorites]

If google really cared that much about H1 then SEO assholes would have every last bit of their pages done in H1...
posted by Artw at 8:56 PM on May 18, 2010

For the first five minuets I thought they were talking about what this was supposed to do.
posted by Hoenikker at 9:52 PM on May 18, 2010

If google really cared that much about H1 then SEO assholes would have every last bit of their pages done in H1...

I'm sure they look at H1 frequency, and its relationship to the text surrounding it. Similarly, my webpage that just repeats "wikipedia" seven thousand times would not get a higher page rank than en.wikipdia.org.
posted by Hoenikker at 9:56 PM on May 18, 2010 [1 favorite]

The Problem: there's so much information out there that it's hard to find the news stories/camera reviews/tweets/emails that you're interested in.

The Solution: some kind of magic that in some way represents relationships of some kind between some kinds of things. This may or may not require "ontologies", whatever they are.

The Payoff: stuff will be cool, in ways that Tim Berners-Lee can't imagine.

For a 15 minute documentary with presumably some of the top minds in this field, I'd expect a bit more content than that. We don't even learn what kind of thing these people are envisioning. Is the "Semantic Web" a singular monolithic entity, like an all-encompassing graph of all the categories and relationships of the world (to which things like news articles and camera reviews refer)? Or is it more like a design methodology, just a trend toward adding tags and labels to things wherever it might help? Are Flickr tags part of the Semantic Web (and if so, why weren't they offered as example #1 in the documentary)? Can someone give at least one sketch of how a Semantically-enabled web might make some task easier, without requiring AI-complete computation behind the scenes?
posted by logopetria at 11:43 PM on May 18, 2010

People will start tagging their content right after they start flossing.
posted by mecran01 at 11:49 PM on May 18, 2010 [5 favorites]

gatekeepers suck.

It's true, but it's worth noting that people bearing formal richly semantic document schemas and ontologies are potential gatekeepers as well -- and even if they turn out to be both wise and good the abstractions themselves become gatekeepers of a sort.
posted by weston at 12:54 AM on May 19, 2010

Semantic Web assumes everyone will bother. That's why XHTML 2.0, if not dead, is on life support. The W3C must escape its ivory tower "I am sure everyone will want to catalog this, every jot and tittle" mentality or be relegated to just one of those curiosities of standards bodies.

But people do use semantic HTML. Not everyone, and maybe not you, but many do. I just used <blockquote> and <em>, which are semantic elements. Metafilter, which is fairly old, and not especially taken with Web 2.0 in my estimation, uses <h1>s for headers! A website can use semantic HTML without jumping on the every-latest-microformat, every-last-tag bandwagon.

Not only that, but many people publish on the web using a blog, and nearly all blog templates (at least the ones I've looked at for wordpress and tumblr) use at least some sort of semantic html. Many of them also use microformats. HTML 4.0 limits somewhat the tags available for describing the semantics of a blog, but with HTML5 taking off rapidly, and a method in place for supporting styling it in older browsers (yes, even IE6), I think we're poised to see a lot of things happening in terms of semantics.

There's a sense that things like this are only done in the ivory towers of academia, but the reality is that I'm a web developer, and HTML5 with semantic elements makes things easier on me. Instead <div class="article">, <div class="wrapper">, <div class="sidebar">, I will soon be able to simply style the <article>, <section>, and <aside> elements.

A great deal of information is also published by social sites like Facebook. Which tag contains your name on a facebook page? The <h1>.
posted by !Jim at 1:14 AM on May 19, 2010 [1 favorite]

I'm confused: is this the one where we switch to lisp?
posted by wobh at 3:28 AM on May 19, 2010 [1 favorite]

But people do use semantic HTML.

Don't confuse semantic HTML, which is easy to do, with the Semantic Web, which is a completely unrealistic morass of jewel-encrusted boxes designed to obscure and hide data. Semantic HTML is focused on the data, has meaning, and is useful to some degree. In contrast, Semantic Web is focused on the containers rather than the data, and is 100% useless (ok, fine, maybe 99.9% useless). As others have pointed out, this entire video is devoted to talking about the problem without actually talking about any solutions or applications. This deadly misfocus is endemic to the entire field. The best advice I ever got was to ignore this crap, but I to deal with some data in RDF (OWL actually) and I had to dive in anyway. It's a perfect storm of architecture astronauts ruining a plausible sounding idea.
posted by Llama-Lime at 5:49 AM on May 19, 2010 [1 favorite]

Wobh, I think that'th been deprecated. It turnth out lithp hath thome acthethibility ithsueth.
posted by nicepersonality at 6:24 AM on May 19, 2010 [3 favorites]

Crap, Pope Guilty, people are STILL saying Library 2.0? We are so fucked.

(/graduated six years ago)
posted by clavicle at 8:59 AM on May 19, 2010

My current line of work is related (in a weird convoluted way) to the semantic web concepts.
The semantic web will come about, but many are seeing the wrong trees in the forest. One assumption that is commonly held by both supporters and detractors of the semantic web is that the content creator will be the one that classifies the content. Ha! I say, in fact Ha!Ha!. Here's a little secret that should be obvious, there's billions (if not trillions) of dollars being spent creating semantic data. The issue is that the data is isolated in its own vertical clusters. It can't be correlated with other content in other clusters. On Metafilter there's the tags, on your computer there's file folders and filenames, on your browser there's bookmarks, on facebook there's your social network, on google there's search terms, and on and on and on. We classify information All.The.Time. Now, the detractors of the semantic web are absolutely right that no one's going to bother going to any extra effort to conform to some external practice when the practices they currently have work fine for them, thank you very much. But the content creator doesn't have to be the same entity as the content relationship finder. Remember that tagging effort a bunch of Mefites did to tag old posts? That is the future of the semantic web, except just one step abstracted. Rather than creating a relationship between an identifier (tag) and content (post), what will happen is the creation of relationships between relationships. Here's an example that's probably doable with a greasemonkey script. Anytime there's a youtube post on mefi, the tags for that post are used to find other youtube videos with matching tags. So when you click on the youtube link, up comes another window with videos related to the *mefi* tags (as opposed to the youtube "related content" which may not be as an ideal match). That's the semantic web. Or more precisely, the semantic web would be the mostly invisible underpinings that would make it easier to build those relationships.

However, having said all that, the current approach the semantic web folks are taking will fail, it requires extra work with no direct benefit, at least it will fail if the goal is for content creator to adopt their framework. But, there are ways to make it work, one example, simple plugins for your web framework (wordpress, rails, etc) of choice that automatically did all the heavy lifting. Or, third party solutions that takes siloed relationship information and converts into a more extensible semantic web format. But the current focus seems to be on the plumbing, before the house is even designed. A better approach would be to create semantic web applications that are cool, easy to use, and useful, and then have a discussion on the underlying technology.

We are definitely moving towards the semantic web, mashups, tags, social networks, these are all about relationships, and as the relationship information on the web grows more complex, there will come about a technology to help organize it. What that technology will actually look like is anybody's guess at this point.
posted by forforf at 10:10 AM on May 19, 2010 [3 favorites]

with HTML5 taking off rapidly

To be honest HTML5 is up there with Web2.0 and Semantic Web as far as a marker for potential bullshit goes these days. Fancy new names for DIVS is not that impressive, VIDEO vs Flash is pretty much neutral, and if HTML5 is used to mean that the site is all rendered in the canvas tag via javascript then it's actually a step back for semantics.
posted by Artw at 10:49 AM on May 19, 2010

First MeFi post. This thread lured me in. Your remarks especially, forforf, are full of win, biggest being "the current focus seems to be on the plumbing, before the house is even designed." Truth.

There's an irony in this video's title beneath the face absurdity of web versioning: had the '90s Web known itself as 1.0, Semantic Web people would certainly have been calling dibs on 2.0. Of course, the technologies we actually take to be Web 2.0 "create valuable meta-data without needing any of the trappings of the Semantic Web." These interfaces work by funneling gestures we take for granted – pointing at things we like, typing a few words about them, moving them around – into small-s semantics that machines can use to reconstellate our affinities without having to parse a thing. And in the 2010s, it will again be simple improvements on working ideas that gradually converge into Web 3.0, not the promised unveiling of a castle in the clouds built by Information Science PhDs who finally figured out a method to encase human language and every truth it captures into that uniform data structure Llama brilliantly describes as a "morass of jewel-encrusted boxes". (I mashed up several metaphors there. W3C noncompliant.)

That quote about emergent metadata is from Clay Shirky's The Semantic Web, Syllogism, and Worldview, also referenced at the outset of the video. He posted it in 2003. Predictably, SemWeb bloggers raked him over the coals for it, saying he'd strawmanned their scholarship or failed to grasp the RDF/OWL schema. Seven years later I can't find a single assertion of Shirky's on this topic that hasn't proved out.
posted by Municipal Hare at 3:41 AM on May 21, 2010 [3 favorites]