Re: KIT releases monumental dataset of more than 15 *trillion* triples

On 1 April 2018 at 23:16, Denny Vrandečić <vrandecic@gmail.com> wrote:

> Thank you. We were considering using the blockchain for sustainability, as
> it would then be a distributed app, but I was told "that's not how it
> works, Denny!"


A block chain is not suitable for pseudo deterministic data sets

First of all, it could be easily 51% attacked, which could cause potential
unstable behaviour

More importantly, the entropy of the data set is much smaller than that
entropy you would send across the wire

Please use a torrent, or better still, ipfs


>
>
>
> On Sun, Apr 1, 2018, 13:12 Axel Polleres <axel@polleres.net> wrote:
>
>> Dear Denny,
>>
>> given the size of the dataset and it's potential impact in terms of
>> resource needs when the full dataset is being processed,
>> I'd strongly suggest to switch to CO2-license (see also my blogpost no
>> Sustainable Computing)!
>>
>> just my two cents, HTH, and congratulations for this huge advancement of
>> LON!
>>
>> Axel
>>
>> --
>> Dr. Axel Polleres
>> url: http://www.polleres.net/  twitter: @AxelPolleres
>>
>> On 01.04.2018, at 10:31, Denny Vrandečić <vrandecic@gmail.com> wrote:
>>
>> KIT is proud today to release an extension to an existing dataset, which
>> will increase the size of the dataset by a factor of more than 1000
>> <http://km.aifb.kit.edu/projects/numbers/web/n1000>. The widely cited Linked
>> Open Numbers <http://km.aifb.kit.edu/projects/numbers/> dataset (more
>> than 30 <http://km.aifb.kit.edu/projects/numbers/web/n30> citations) has
>> been updated. Every single triple was regenerated, and even though the size
>> has been dramatically expanded, we remain confident in the quality of every
>> single triple.
>>
>> http://km.aifb.kit.edu/projects/numbers/
>>
>> It has been - on the data today - eight
>> <http://km.aifb.kit.edu/projects/numbers/web/n8> years since the
>> original publication of the Linked Open Numbers dataset. Today, we are
>> proud to announce to increase the size and thus utility of the dataset by
>> three <http://km.aifb.kit.edu/projects/numbers/web/n3> orders of
>> magnitude.
>>
>> The page has received a thorough remake, not only refreshing it optically
>> and updating it to display better on mobile devices, but also introducing a
>> number of new features:
>>
>> * the previous limit to the first billion
>> <http://km.aifb.kit.edu/projects/numbers/web/n1000000000> natural
>> numbers has been lifted, since the page has in the meantime moved to a 64
>> <http://km.aifb.kit.edu/projects/numbers/web/n64> bit architecture. We
>> expanded the supported numbers to the first trillion natural numbers,
>> therefore creating 999 billion
>> <http://km.aifb.kit.edu/projects/numbers/web/n999000000000> new entities.
>>
>> * all links to Wikipedia and DBpedia have been refreshed. In the eight
>> years since the original release, Wikipedia and DBpedia have in an effort
>> to catch up with Linked Open Numbers created new entities for numerous
>> numbers. We have updated the links to all of those.
>>
>> * also links to Wikidata entities representing these numbers have been
>> created and added, extending the linkage between Linked Open Numbers and
>> the LOD cloud by thousands and thousand of new entities.
>>
>> * the whole dataset is now published under the terms of the CC-0 license,
>> countering long years of discussion that resulted in fear, uncertainty, and
>> doubt. Now the Linked Open Numbers dataset is standing on a solid
>> grounding, joining other major datasets in choosing the perfect license for
>> data.
>>
>> * we expanded the ontology and the dataset to also provide the digit sum
>> of the numbers, allowing new applications on top of that.
>>
>> * we refreshed the links to Linked Data browsers. The original six
>> <http://km.aifb.kit.edu/projects/numbers/web/n6> browsers are all not
>> available anymore to allow to browse over the Linked Open Numbers dataset.
>> Therefore these links were all removed, and replaced with two
>> <http://km.aifb.kit.edu/projects/numbers/web/n2> current browsers.
>>
>> * we also support the URI4
>> <http://km.aifb.kit.edu/projects/numbers/web/n4>URI project and
>> providing data about the Linked Open Numbers URIs in the URI4URI
>> <http://uri4uri.net/> scheme.
>>
>> * the page has been updated to support Unicode's UTF8
>> <http://km.aifb.kit.edu/projects/numbers/web/n8>, thus showing the
>> number names in their new full glory.
>>
>> Eight <http://km.aifb.kit.edu/projects/numbers/web/n8> years - 2922
>> <http://km.aifb.kit.edu/projects/numbers/web/n2922> days - after the
>> original publication Linked Open Numbers still gets tens of thousand
>> <http://km.aifb.kit.edu/projects/numbers/web/n40000> hits per month. We
>> are happy to have updated the resource and expanded its lifetime
>> considerably.
>>
>> The community is invited and challenged to provide a SPARQL endpoint to
>> the dataset. We think that the size of the dataset would provide for an
>> interesting challenge.
>>
>> An open source release of the code base is being planned.
>>
>> The update was created in collaboration by Denny Vrandecic, Steffen
>> Thoma, Andreas Thalhammer, Andreas Harth, and York Sure-Vetter.
>>
>>
>>

Received on Monday, 2 April 2018 13:42:21 UTC