How would you easily explain the concept of the Semantic Web to someone with basic technical knowledge? Re-factoring something I said in a blog a while ago, I’ve now edited it down to this:
- Web 1.0 was about connecting-up documents.
- Web 2.0 is about connecting-up people.
- The Semantic Web is about connecting-up data.
So, it’s a bit basic, but I think that gets the premise across. I’d be interested to hear any suggestions you have for other easy explanations. Similarly, the Semantic Web really needs a re-brand, doesn’t it? Maybe if it sounded cooler or less techy, people wouldn’t be so scared of it. What would you call it?
iWeb? Web 3000? Ultra-web? Ultra-Intra-Web? The Dark Web? (After all, it’s a bit like dark matter, and is intended to be the invisible glue that makes up the majority of the mass of the web). Maybe we should bring the word ‘cyber’ back into fashion, I kinda miss it… Cyberdata! No, on second thoughts… Anyone have any bright ideas!?
cyberglue...?
The problem of the Semantic Web is not that the name is uncool, but that it has been hijacked by theoretical minded people who tried to build the ultimate logical language.
OWL and RDF are similar to SGML, what is needed is a pragmatic lightweight approach that is similar to HTML. So that anybody can annotate their knowledge.
We are building Semantic Web engines and we find that explaining OWL to computer science university graduates takes some time. So explaining it to a mass user base is virtually impossible.
TBL should the whole thing get back to his unique vision of simplicity and less is more that made the web so powerful. Other wise the Semantic Web is RIP.
Every definition I've ever seen for the Semantic Web has been about structuring documents to allow software to comprehend the data contextually instead of just searching for words. It's hard to get people excited about that because it seems like it's meant solely for the machinery - Not enough "What's in it for me?". How can experts so concerned about linguistic context not understand how linguistics affect their own audience?
My take? If you want to excite more people, then use the definition itself to directly involve those people:
Web 1.0 is about letting a webmaster organize the data.
Web 2.0 is about letting the audience organize the data.
The Semantic Web is about letting each individual user organize the data as he/she sees fit.
Who cares that it's actually software that shuffles the data around for them? It's the USER that sets the goals. Make all discussions about the Semantic Web more about enabling people and less about enabling machinery.
If you just must deal with the mechanics, you might also present a list illustrating the expanding scope of the source for "the data":
Web 1.0 is the presentation of all the data previously collected by the webmaster.
Web 2.0 is the presentation of all the data previously collected by the audience.
The Semantic Web is the presentation of all the data previously collected by the entire web itself.
In the previous definition, the empowerment and authority gets closer and closer to the hands of the User. In this second definition, the magnitude of work gets larger and larger, putting less demand on the User. The important part is in phrasing your description in terms of the benefit to the User, not the machinery.
Web 4connect
i.e., connecting-up document, connecting-up people, connecting-up data and connecting-up applications.
Okay, so maybe it's not as catchy as "Web 2.0", but how about --
"The Internet"
? :) :D
Infoweb?
It's worth noting the difference between data and information, by my understanding, Semantic Web deals with information and not data, which is the reason for the logical language and metadata requirements.
To an extent I agree with MonkeyT, we need to concentrate on the end user's experience. So long as it works and they can do what they need to, the technology behind it is irrelevant to most users.
Most people I deal with on a daily basis don't understand dynamic websites and care even less about how the sites they use work. They just want to be able to get on with whatever they want to do and not worry about the technological how.
Since we're talking about data about visible data leading to informed data, how about Meta-web?
And I agree with MonkeyT and Loki: people are not interested in how it works. It's like with cars: they want to turn the key and have it running. We must not forget that, in the end, we're doing it for the user.
"OWL and RDF are similar to SGML, what is needed is a pragmatic lightweight approach that is similar to HTML."
See microformats.org.
The brand name you are groping for is "Telepathy". It is precisely accurate and conveys overwhelmingly why the maximal charter of the Semantic Web is impossible and would not even be a good idea if it were.
That said, there would be large benefits in enabling efficient semantic mapping through finite, domain-specific, governed ontologies. But that would be a strong reduction of the Semantic Web charter as it is usually informally understood.
I'll willingly play devil's advocate a bit here...
> "there would be large benefits in enabling efficient semantic mapping through finite, domain-specific, governed ontologies"
Isn't that how many 'real' Semantic Web advocates suggest the SW should be built, though? The SW doesn't rely on a global ontology (in fact, unlike many XML Schema, RDF is built around the notion of being able to extend and map to other schema), and you can quite easily build simple, finite, domain-specific ontologies with RDF Schema? I completely agree, that we should just forget about the big-picture stuff, and probably forget about OWL even for the time being, and get some simple ontologies (and, more importantly, applications that use them and exchange data with them) out into the world.
> "See microformats.org."
Yes, microformats are fun, and shoe-horn a little more 'semantics' into a basically unsemantic language (HTML). But how much more juice can you squeeze out of them? Even 'tagging', which is probably the most common (and evolved) type of microformat, isn't particularly useful...
> "we need to concentrate on the end user's experience"
Well, I sort of agree, and sort of disagree. In some respects, yes, if we could think of an end-user killer-application for the SW, then that's a huge plus. But the SW isn't an application. It isn't a web-site. It's not something that the end-user should even know they're experiencing; it's basically a protocol. To that end, I'm not sure how much we should get bothered about 'end user experience' at this stage, but rather just get on with it and put the foundations down. I'm not sure how much the end-user was considered when TCP/IP was under consideration, but it's a similar sort of thing. A tiny number of application (not necessarily 'web') developers need to consider the data types and issues, but for most people, it should eventually be an invisible part of the infrastructure, like DNS or HTTP.
>"> "we need to concentrate on the end user's experience"
Well, I sort of agree, and sort of disagree. In some respects, yes, if we could think of an end-user killer-application for the SW, then that's a huge plus. But the SW isn't an application. It isn't a web-site. It's not something that the end-user should even know they're experiencing; it's basically a protocol. To that end, I'm not sure how much we should get bothered about 'end user experience' at this stage, but rather just get on with it and put the foundations down. I'm not sure how much the end-user was considered when TCP/IP was under consideration, but it's a similar sort of thing. A tiny number of application (not necessarily 'web') developers need to consider the data types and issues, but for most people, it should eventually be an invisible part of the infrastructure, like DNS or HTTP."
Noting that you're playing devil's advocate, the thrust of the point was, not in terms of a high impact, marketable product, but rather an underlying, unasuming, and to futher anthropomophise, modest protocol that allows people to get the most out of all the information available without them really realising it. DNS, DHCP are all good examples of this.
The problem seems to be some folks seem more interested in having grand theories and ideas published and getting the plaudits for that than just getting on with developing something useful. Again, it's the small steps and little innovations that will make SW possible in the long run, TCP/IP was the start of the begining for the internet, not the major breakthrough people are looking for for SW.
Yes, I completely agree - it's those tiny steps and innovations that are imperative... a pity they're so elusive!
I dropped my thoughts on this issue at FUP to Explaining the SW in 10s.
Cheers,
Michael
the semantic web in less than ten seconds:
Computers have their own idea about the meaning of different words. The Semantic web is an attempt to create a dictionary which all computers can use.
I make that about six seconds ;)
Dependent on the actual implementation of the technology, "Web 3.0" could make sense. If the mechanics of the so-called Semantic Web are applied to the existing workings and aesthetics of Web 2.0, which would be perhaps the best outcome, I can see no reason to call it anything else. Besides, your everyday user is still going to call it "the internet", as previously suggested.
I think it's wrong, at least, wrong in that "web 3.0" won't look like the semantic web you have been talking about. It will look much more like a Wikipedia troll war spreading across the entire Internet, googlewashing everything it can touch, which will be more and more and more. Once the people are connected, they'll form factions and this means politics as usual. The advantage of the semantic web technologies is that they let academics actually participate efficiently. Which is why they don't, as a rule, bother to edit Wikipedia directly. But when they can work inside an institution or group of peers to generate a very influential semantic mapping between concepts, then release it on the world when they think it's ready, then get lazy people to adopt it by default, they turn their opponents into mere trolls opposing it.
Disempowered by this, the trolls fight back, but the only way to do that is to agree on a dissensus position, probably a looser definition of the same concepts, with more rigorous connections to others. Then release that and convince others to use it by a combination of political and economic suasion. If it really does work better to solve integration problems between fields, it will catch on. If not, it'll have a much longer tougher uphill fight.
Consider each data item to be a fight. If it is, then the semantic web is about setting rules to have lots of fights fast, and let large factional armies set up to fight much faster than before and consolidate victories more completely than before. A lot like nation-states, which let a conqueror take the entire country over by just taking its capital city...
So a more correct way to say this is:
"Web 1.0 was about connecting-up documents.
Web 2.0 is about connecting-up people.
Web 3.0 is about connecting up beliefs so they can fight out in the open, making it easier to show people documents they need to read to make that fighting more ethical, efficient and decisive."
Those who don't think that ontology editing is a form of warfare have probably not spent any time examining Wikipedia edit history... I refer you to Cory Doctorow's various writings on this:
"Britannica tells you what dead white men agreed on. Wikipedia tells you what live Internet users are fighting about."
That is exactly correct, and that's ALL that Wikipedia does. Its great achievement is its namespace and links, not its article text. The RDF is useful precisely because it omits the text, so it's not clear that you get more out of adding semantic tags in than you would out of increasing the depth of data on each link.
Id call it Ultra Cyber Web 2.0 + 1.0