Re: Open Source Tools + Workflow for Publishing Linked Data

Our group has been looking at Callimachus (http://callimachusproject.org/)
and Fedora 4 (https://wiki.duraspace.org/display/FF/Fedora+Repository+Home)
for interface and workflow implementation ideas for publishing and
maintaining linked data within a community of practice. Has anyone else
heard of any other open source tools that would support #3 below:

3) a service based on the dereferenced HTTP Accept request  that would
dynamically generate/produce alternate linked data serializations (e.g.,
JSON-LD, RDF/XML, Turtle) on the server of the source RDFa/HTML





-------------------------------------------------------
Advanced Distributed Learning Initiative
+1.850.266.7100(office)
+1.850.471.1300 (mobile)
jhaag75 (skype)
http://motifproject.org (MoTIF Project)
http://ml.adlnet.gov (Web)
http://twitter.com/mobilejson (Twitter)
http://linkedin.com/in/jasonhaag (LinkedIn)

On Fri, Jul 24, 2015 at 8:12 AM, Sarven Capadisli <info@csarven.ca> wrote:

> On 2015-07-24 14:47, Haag, Jason wrote:
>
>> Hello Semantic Web Community,
>>
>> I'm from the learning technology space and we have been investigating
>> the use of semantic web technology as part of our workflow for
>> publishing controlled vocabulary terms. These terms help provide the
>> specific meaning of verbs and activities supporting various learning
>> experiences. We've mostly been trying to leverage SKOS and PROV
>> ontologies for this effort.
>>
>> I'm interested in leveraging open source tools that might help our
>> Communities of Practice (CoPs) more easily publish these terms as linked
>> data. I envision a publishing tool or repository interface that would
>> bring the process together rather nicely, and also help compliment our
>> governance and maintenance concerns as well. We can't expect our
>> disparate CoPs to each have the resources or knowledge to configure
>> servers on their own to support content negotiation for the level of
>> granularity we are interested in for publishing our linked data.
>>
>> I envision a workflow that would support the following:
>>
>> 1) allow CoPs to utilize HTML/RDFa templates and simply populate those
>> with persistent URIs and the suggested metadata from SKOS and PROV.
>> 2) publish the RDFa to a web server or repository tool
>> 3) a service would dynamically generate alternate linked data
>> serializations (e.g., JSON-LD) of the RDFa/HTML based on the
>> dereferenced HTTP request
>> 4) any application could then consume linked data in any format in real
>> time based on the single source HTML/RDFa provided at each IRI
>>
>> RDFa seems to be the most user friendly for those that are not RDF
>> savvy. Also, rather than putting the responsibility on CoPs to embed
>> JSON-LD in HTML or configure / establish various rewrite rules it seems
>> a publishing server or service might handle this more efficiently. Does
>> this seem like a practical workflow for publishing linked data? Are
>> there any flaws with this proposed workflow process?
>>
>> Finally, is anyone from this community aware of any open source
>> applications that would support this type of workflow? Thank you in
>> advance for your responses and support.
>>
>> Warm Regards,
>>
>> J Haag
>>
>> -------------------------------------------------------
>> Advanced Distributed Learning Initiative
>> +1.850.266.7100(office)
>> +1.850.471.1300 (mobile)
>> jhaag75 (skype)
>> http://motifproject.org (MoTIF Project)
>> http://ml.adlnet.gov (Web)
>> http://twitter.com/mobilejson (Twitter)
>> http://linkedin.com/in/jasonhaag (LinkedIn)
>>
>
> Hi Jason,
>
> HTML+RDFa is great in a sense that you have a single document which is
> useful for humans as well as machines. While publishing (read purposes)
> only in RDFa is okay, you might want to consider the remaining CRUD
> operations, if you have to deal with it from the outside.
>
> You are welcome and encouraged to take what you like from here:
>
> https://github.com/csarven/linked-research
>
> To summarize: the RDFa templates are written as HTML5 Polyglot documents
> (fancy way of saying that it can act as HTML5 or XHTML5 given respective
> content-type in the response). See the examples, click around the menu
> (e.g., LNCS, ACM). Dereference the URLs for RDF. Use a Line Mode Browser
> e.g., links, and see in fact that all of the content is there. It is
> "religiously" progressively enhanced. There are ways to embed Turtle and
> JSON-LD into these documents, and you can export etc. More work on the
> authoring end is on its way.
>
> Feel free to ping me off the list for any details and future goals.
>
> -Sarven
> http://csarven.ca/#i
>
>