Slashdot Mirror
Nepomuk Brings Semantic Web To the Desktop, Instead
An anonymous reader writes "Technology Review has a story looking at Nepomuk — the semantic tool that is bundled with the latest version of KDE. It seems that some Semantic Web researchers believe the tool will prove a breakthrough for semantic technology. By encouraging people to add semantic meta-data to the information stored on their machines they hope it could succeed where other semantic tools have failed."
I've tried Symantec products in the past, and they are worse than actually having a virus. They slow your PC to a crawl, get their claws into every part of your computer, and are extremely difficult to purge when you finally give up on them.
What exactly is semantic web, and why haven't I ever heard of it?
--- "When you gotta do something wrong. You gotta do it right. (Fighter)"
I've tried out Nepomuk and, while I have to say that it's promising, it's got miles to go before it's even near ready. The main problem is application support. Sure, you can rate and tag and describe your files in the Dolphin file browser. So what? You can do the same in Vista. This doesn't mean anything if applications don't hook into this and make use of it. Of the apps I've used, Gwenview (a photo viewer) has Nepomuk partially implemented but it's buggy and you need to compile it yourself with it explicitly enabled (this will apparently change in KDE 4.2). Digikam, which allows you to rate, tag, and describe photos already, says that they have no plans of integrating with Nepomuk anytime soon. Amarok 2 has work towards a Nepomuk collection, but the devs say that this will always run along side the main, MySql-based collection and it's nowhere near ready yet. My email is in the cloud so I can't even begin to talk about KDE-PIM's support or lack thereof.
The other problem at the moment is a lack of ability to query your semantic data. Can I get anything to show all photos with my wife in them that I've rated four or above? Not at the moment. Hopefully this is coming in KDE 4.2, but as it stands at the moment it makes Nepomuk a case of write-only memory.
So, maybe something to get excited about in the future, but not quite yet.
NepoMUCK? Anything ending in "MUCK" doesn't sound like a good product. The concept is very interesting but the name isn't the best I've seen.
I'm glad that they don't prefix everything with K though.
Yes, I know that Nepomuk means "Networked Environment for Personalized, Ontology-based Management of Unified Knowledge" as stated in the article.
You are not entitled to your opinion. You are entitled to your informed opinion. -- Harlan Ellison
It's not as bad as GIMP :)
And I'll tell you why.
The Nepomuk Web site wants to make me chew my own arm off. Now, I'm familar with the Semantic Web, I'm excited by the idea of semantic organisation. But this site is the epitome of grim, lifeless European research-ese. It completely fails to convey the technological approach, how it works, or why you should give a damn. I get the impression that the team was more interested in the EC funding then actually developing a disruptive technology.
Why why can't researchers spend 15 minutes thinking about how to convey the importance and excitement of what they are trying to do in terms of practical examples.
I'm afraid you'll probably have to wait until some enterprising 3rd party to grab the source and build some of the technology into a different product.
All information is semantic. This slashdot post is information encoded using English semantics. Unfortunately for the machines, the English semantics are way to complicated for them to understand. So they need a simpler set of grammar rules to be able to parse it. But why would anyone want to waste time marking it up just for the benefit of machine readability when google basically can accomplish the same thing without all that metadata markup cruft?
Football Odds
There's a good rant from Corey Doctorow about this. I think the best phrase that summaries people's high hopes for the semantic web is "nerd hubris".
Easy: Just have a bot add "untagged" tags to everything not yet tagged. Then it's tagged, because it's tagged "untagged".
Also easy: Just remove all wrong information before making your deduction. OK, so how is the computer to know what is wrong? Well, that's of course again semantic information, so just tag anything wrong as "wrong". If some "wrong" tagging happens to be wrong, you can still tag that as "wrong" as well.
Just upload them onto any p2p network. Sharing is what they are for, aren't they?
Well, if the software gets stuck, it still can ask a human.
Note to the humour impaired: Imagine a smiley after each sentence!
The Tao of math: The numbers you can count are not the real numbers.
You got that exactly backwards.
The WWW was an earlier doomed attempt at semantic markup, and up until the summer of '93 or so it looked like it might work. That's when the early rants about people using the tags to control layout instead of too convey meta information (e.g. using em to get italics in a bibliography, dt/dd to make roman numeral lists, etc.) started--or at least when I first became aware of them. In fact, pretty much the entire history of HTML has been a tension between the language's designers and purist, who want users to care about what markup means, even if it does nothing, and the vast majority of users who only care about what it does regardless of the "meaning" that may be ascribed to it. Once you can get your head around both perspectives some of the goofier things in the whole tawdry history (the Table Wars, XML, CSS) make a lot more sense.
Ok, a little more sense. But only if you already knew what people are like.
--MarkusQ
There's actually a pretty good introduction to the semantic web in this month's Communications of the ACM. You're right when you say that the semantic web is, as yet, mostly unrealized. But it has huge potential.
Relational databases were in the same position in the late 60's/early 70's. We needed ways to combine and extract information automatically with a simple and expressive language. Relational database management systems, combined with SQL were the result of that, and they were a smashing success. They are now a standard business tool. The key to that success is essentially the role that the database's ontology plays in an RDBMS.
Having spent a lot of time professionally and academically working with and studying database technologies, most of the work is in understanding your data. Specifically, building a data model. A well-built data model is essentially an ontology. There are various techniques used to make sure that your can be handled automatically, mainly by normalization. This requires a tremendous amount of work on the part of the database designer, but the end result is that the end-user can query this data in fairly simple terms and get an enormous richness of data, sometimes in ways that even the database designer did not foresee. I think the success of database systems is what is driving a lot of the work in building the semantic web.
So you can see-- the big problem with the web is not just that data is not just unstructured, but that there are no standardized ontologies out there. RDF is an attempt to solve some of these problems simply, because you can embed your ontology, but it may be well off. On the other hand, if new tools make structuring data very easy or natural, people may be motivated to do the extra work because they'll personally benefit from it. For example, many people annotate or organize their photo collections naturally, so that they can share them with others. A smart photo gallery software writer may be able to come along and take advantage of that behavior to further enhance the meaning of that data.
> ... the semantic web never did, and never will take off without significant AI involvement.
I understand that the point of Nepomuk is to allow for automated tagging by the standard tools of the KDE desktop. For instance, say you receive a picture from an IM contact who KDE also knows (through the address book framework, Akonadi) lives in Europe.
Then Nepomuk would allow you to make search queries as "Bring up all the pictures that people living in Europe sent me last week". Well, that's the theoretical goal anyway; we will see if they ever get there.
There's one nifty application already: you can create a Folder View plasmoid on your desktop, and instead of making it display ~/Desktop/ as usual, you can make it display the result of a query through the Nepomuk KIO slave. See here how it works.
-- B.
This sig does in fact not have the property it claims not to have.
"Everybody and his uncle tries to make systems that will index every piece of crap on your PC and it invariably results in a useless and horrible waste of resources."
On the contrary, we should seriously be asking ourselves *why*, when all our data is sitting there on our PCs, we've let ourselves get into such a state of disorganisation at the operating system level that a class of program called 'indexer' exists as a third-party tool in the first place.
How come it's not already taken as given that the primary thing an operating system *does* is, you know, *know where all its data is*?
It's as if we're living in an age before 'directories' were invented - or before databases had 'indexes' and 'queries' - and we have to manually write down and key in raw sector numbers every time we open a file. And we're okay with that, because we think - and teach - that that's 'just how computers work'. We've accepted that there's a whole class of things our computers can't do 'because there's no application to do that'.
Something is wrong with this picture.
You are not a brain: http://books.google.com/books?id=2oV61CeDx-YC