How can I use Linked Open Numbers projects

Question

There is Linked Open Numbers project. I don't understand how I can use it. It can be used with sparql query, I think. But I can't do it. I have found http://km.aifb.kit.edu/projects/numbers

Answer 1

Do you know that Linked Open Numbers was an April fools' day joke?

Note the publishing date on the project website. The paper accompanying the project is firmly tongue-in-cheek, and was published in RAFT , a joke journal.

At the time, many Semantic Web research groups were republishing existing data sets in Linked Data format, sometimes without much consideration for the actual usefulness of doing so. Linked Open Numbers pokes fun at that fashion. Linked Open Numbers is a very earnest effort at doing something that is entirely useless. Except, maybe it isn't useless after all? Maybe we just haven't opened our minds to the possibilities? Who knows! It's just like the entire Semantic Web effort!

With that out of the way, you asked how you can use the site, so I will explain.

First, there is no SPARQL endpoint, and no downloadable RDF dump either. There are an infinite number of numbers, so the storage required would be infinite!

The only way to use the site is to look up the information it has for any particular number. For example, for the number 42, we have:

• the linked data URI: http://km.aifb.kit.edu/projects/numbers/n42
• the web page: http://km.aifb.kit.edu/projects/numbers/web/n42
• the same information in RDF/XML format: http://km.aifb.kit.edu/projects/numbers/data/n42

So you can write code that downloads the RDF/XML data about a particular number, and accesses the various properties of the number, such as its spelling in different languages, its Roman numeral, and its prime factors. Most RDF libraries will make this job very easy.

And if you publish a dataset in RDF format and/or as linked data, and users of the dataset could potentially benefit from this information about numbers in the dataset, then you can include RDF links from numbers in your data to Linked Open Number's linked data URIs.

Is any of this useful? I don't know!

Answer 2

Agree with almost everything Richard said in his answer, besides a few tiny point:

• Linked Open Numbers, unlike the natural numbers, is actually not infinite - the dataset used to stop at a billion (minus 1), and since an update it goes to a trillion (minus 1). So that makes the amount of data you need to store also less than infinite - merely unwieldy large.
• I agree with Christoph Lange's comment that one could have added more information to Linked Open Number which would have made it even more useful, such as factorizations for large numbers. As it is now, we only factorize the numbers below a million. This is because the results are not actually stored, but calculated on the fly. Storing all the factorizations for the larger numbers could indeed be a useful service, but would require quite some refactoring in the back-end.
• Surprisingly, Linked Open Numbers is (or at least, was) actually used by a number of external projects - for getting the name of the numbers in different natural languages. There were some projects who, instead of implementing a name-generating code themselves, just used a look-up to Linked Open Number.
• Another use case for Linked Open Numbers is actually for testing purposes and as a showcase - for testing a Linked Open Data browser, crawler, and other tools. We put a lot of effort in ensuring Linked Open Numbers implements the standards correctly and comprehensively, so you can always find a dataset that kind of makes a bit more sense than just a synthetic dataset, but still is large enough to test your tools. I think that makes Linked Open Numbers genuinely useful already, even without implementing Christoph's suggestion (which I would entirely support).