Hacker News Reader

Comment by frik | Hacker News Reader

As you mentioned Google Knowledge Graph, it's based on the open Freebase ontology. Google bought the company behind Freebase and the ontology will be no more in March 31, 2015: http://www.freebase.com , http://en.wikipedia.org/wiki/Freebase

A rather sad demise. There is an half-hearted(?) attempt to donate some data to WikiData of Wikipedia. But given the track record of WikiData history and how different the ontology is, it looks like Freebase will go offline forever and used as internal datasource for Google Knowledge Graph.

It would be great if Google would donate the data and the tools to generate Freebase to Archive.org or another open source community, so that they can regenerate the ontology on a monthly interval from its Wikipedia, and various other data sources. Given that Freebase is also what powers Siri, Cortana and Watson, maybe another corporation can help.

View on HN

I agree. I have used Freebase since Metaweb became a company. Really sorry to see Freebase disappear. Archives are runnable on a publicly available AWS EC2 image, but there will be no more community contributions to Freebase.

Is there a way to migrate Freebase into a community-funded project? Having that kind of knowledge set disappear is very concerning.

Is there evidence that the migration to Wikidata is half-hearted? This was just announced in December.

Not sure this is evidence of a "sad demise". Wikidata is non-commercial project managed by a foundation. Freebase/Metaweb was always a commercial project. Seems like this is a move towards openness if anything.

Full Freebase data dumps have long been available.

https://developers.google.com/freebase/data

Freebase updated the data dumps monthly. The main data source is Wikipedia, with many other data sources that update their content frequently.

Without releasing the tools and without a community who adds/fixes the data, it will be a lot less useful/useless. It's like reading a years old printed lexicon.

There is a lot evidence that the up-coming event will be a big loss - do a web search. A lot of software has been written and a lot of effort at e.g. http://schema.org has been made, all based on Freebase ontology with links to Freebase website. Freebase is basically the machine readable form of Wikipedia plus a lot of other data sources.

It's basically like closing down Wikipedia. And then a data dump of Wikipedia of March 2015 will be useless in 2020!

Do Facebook, Microsoft, Nuance, IBM, WolframResearch all have already an in-house Freebase clone. Or how do they update their databases in near future - the data used for Siri, Cortana, Siri, Watson, WolframAlpha. If not, would one of these companies be so kind and boot-strap/support an open Freebase rescue project?

Freebase doesn't have an ontology in the Semantic Web sense. It's always been a specialized tool with non standard protocols like MQL. Schema.org is somewhat related but not tied whatsoever to Freebase.

Wikidata is fully editable and has an ecosystem for updating data so concerns about stale data aren't valid. Since Wikipedia will be pulling from it (is already in some cases) you could argue that their will be more sunlight on the data.

There's been a lot of code written against Freebase APIs (me included) but it isn't sad it's going away. It's the risk you take when relying on 3rd party services. I think your confusing short term developer convenience with a real loss of open data.

Freebase has 2,751,750,754 entries, Wikidata has only 13,734,841.

Lydia Pintscher (Wikidata manager) admits that the Wikipedia/Wikidata notability guideline is a real problem and the process of re-using Freebase data is slow: https://groups.google.com/forum/#!topic/freebase-discuss/s_B...

How cares about an API, it's all about the data dump that is available for download and used to be updated monthly.

I am not sure what your agenda is, if you are somehow related to Wikidata or Google and why this is only your eight post in 1526days on HN. But one it cannot be denied that there is evident that the demise of Freebase will hurt us all (except Google) in the long term and a lot of data won't survive/be included in Wikidata.

(I do like Wikipedia, but I noticed their notability guidelines and some admins gone wrong do hurt for example the German Wikipedia a lot, which is actually shrinking as more pages get deleted than added. Some good projects from German Wikipedia like the Toolserver which hosts the map-data and the geo-database of all cities and landmarks are great. But Wikidata with its long development history should have been done by Wikimedia itself not Wikipedia Germany. It started from SemanticWiki research project and it took way too long and is still not that good) http://www.heise.de/newsticker/meldung/Blutet-Wikipedia-aus-..., http://www.heise.de/newsticker/foren/S-Artikel-schrumpfen/fo...

Reply on news.ycombinator.com