Hugh Glaser wrote:
Sorry, but you probably knew this was coming...
The esw page starts by saying:
"The Open Data Movement aims at making data freely available to everyone."
This is the context for the LOD activity.
And I guess you knew this was coming:
You already have DBpedia on the Web with an open SPARQL endpoint paid
for by OpenLink.
We are also offering DBpedia on EC2 so that anyone can use DBpedia
freely but handle the cost of infrastructure in the E2 Cloud.
Finally, as your namesake (at OpenLink) indicated, you can also build
DBpedia yourself (ditto the rest of LOD) using Virtuoso Open Source
Edition or any other Quad or Triple store capable of handling the data.
Personally I am quite happy with an LD world and a LOD world, but I think
that some people have a deep commitment to the Open.
At the very least, we should be quite clear about whether we are talking LOD
or LD.
Linking Open Data Sets on the Web, is about publishing RDF archives with
the following characteristics:
1. De-referencable URIs
2. Linkage to other Linked Data Entities & Data Spaces
3. Making data available in the form above from a myriad of sources
"LOD" has nothing to do with the cost of computing & intellectual
resources required to make the above happen.
"Open" has nothing to do with the cost of computing resources &
intellectual resources required to make the above happen.
So the recent postings suggesting the questions and answers for an FAQ (well
done guys!) need to be (more?) careful about the Open. This is especially
true because the data providers will have very genuine concerns about the
work Open when we talk LOD.
"Open" has always been about "Free Speech" and not "Free Beer".
I wonder if it would be better, for these purposes to drop the Open most of
the time, and just talk LD?
I don't think there is a problem with the "O" in LOD.
But to get back to the Amazon world.
It looks to me like it is being proposed as a LOD facility.
It's being proposed as an option for:
1. Situating the RDF Data Sets that adhere to the LOD publishing best
practices in the EC2 cloud
2. Quad Store instances (e.g. Virtuoso) that use this to provide
personal and service specific instance alternatives of the LOD Data
Space (sparql endpoint + HTML viewing pages as we've done with DBpedia)
I think it is a wonderful thing to do, but it is hard for me to consider it
Open in any real sense.
See my earlier comments.
Today you have:
1. http://dbpedia.org/About (hosted and paid for by OpenLink amongst things)
2. EC2 AMI option to make your own for personal or service specific
purposes (e.g. OpenCalais folks already have their DBpedia AMI going as
part of their march toward a Linked Data enhanced version of Open Calais)
Soon you will have:
1. An EC2 AMI variant of Neurocommons (as in:
http://sparql.neurocommons.org/) for example I've place it here:
http://kingsley.idehen.name (temporarily)
2. An EC2 AMI variant of Bio2Rdf.org (*this one comes after neurocommons*)
If I was to grab lots of linked data and then start to offer it to the
community, but say to you that you needed to rent some server space from me
in order to use it, I think there would be a lot of complaints that I should
not be using the word Open at all.
Again, there is nothing mutually exclusive about the places from where
LOD data sets are deployed via Data Servers.
The Amazon situation, as far as I can see, is no different.
I disagree with your perception, completely.
And I would say that the idea they are being "generous" in letting people
host data is simply part of a hard-nosed business model, where they will
draw people into their (paid-for) world, which is fine, but hardly Open.
The location is just an option. It has no effect on LOD.
So I am very happy for this activity to be badged as Linked Data, and I hope
it is very successful.
But from my point of view, please don't call anything part of the LOD world
unless it offers me free access over http.
I am happy to debate you onwards re. this matter.
Your premise is inaccurate.
If EC2 and S3 were the only options, then your point would be valid. But
as I've stated (repeatedly) we are talking about options that cover
deployment and LOD usage patterns. No more, no less.
I am happy to be told I am alone in these concerns, or have misunderstood
the terms of use of the Amazon cloud, and if so will then modify my
perceptions accordingly.
We'll I try my hardest to clarify the EC2 option as it relates to LOD :-)
Best
Hugh
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com
