Wrap-Up of Meet-Up at ESWC2014

Dear Enthusiasts!

We had multiple meet-ups on the RDFJS topic during ESWC2014, now already two weeks gone. Further we could convince some more to join this list. Welcome!

In general the list shall be used for everything RDF-JS related, but especially to show off what great stuff there is around! 

Now I like to wrap-up the animated discussions we had for everybody, probably with a slight subjective touch from my side:

#Topics

## Holding and accessing a graph inside JavaScript
Most people reported handling graphs “by hand” in their applications. After being quite a bit of work and handling often only a subset of RDF this makes the applications often specific to a dataset/ontology and overall not really robust. Multiple times the use of “rdfstorejs” was mentioned, which is the amazing work of Antonio. But as some of you know that this implementation is not complete and also not yet 100% bug free. Antonio himself obviously wants to change this by starting over (https://github.com/antoniogarrote/rdfstore-js/tree/v2) … But it looks like he is not able to invest as much time as he would like.

On the other hand for most applications, it is not necessary to use a full fledged triplestore. Rubens N3 parser which reads Turtle and N-Triples and also plain JSON-LD is often good enough to hold the data inside an application. The tricky part begins with accessing the graph. There is the RDF-API (http://www.w3.org/TR/rdf-api/ stalled right now) which provides a specification how to interact with a RDF Graph. But we don’t know of a implementation expect the partial one build in rdfstorejs. Another take on this would to implement other access methods like e.g. Gremlin (https://github.com/tinkerpop/gremlin/wiki) based on JSON-LD or on other triple representations. And finally we could also write a more robust SPARQL engine which can be used on the mentioned structures.

Any other inputs on how to store a graph inside a JavaScript context and access it, without writing tons of nested loops and so on? Do you think a better way to work with graphs inside a JavaScript context is necessary?
And the most important question, are there people around, able to spend some time creating such a thing?

## Autocompletion
One of the features JavaScript is used from a broader community (as this depends on the use of JavaScript if you don’t want to fall back into plugins) is to integrate autocompletion inside a GUI. We like to ask if there are some best practices to achieve this easily? Especially how would we create a autocompletion service based on RDF directly, without a specialized service for that? (Keep in mind we need response time in less than 100-200ms!) Or would it make sense to create a AutocompletionAsAService (AaaS :-) provider for LD?

## CORS on SPARQL Endpoints
To use data of a SPARQL endpoint directly in a browser application is by the “same origin security policy” in general not possible. Only if the SPARQL Endpoint uses CORS and sets the Access-Control-Allow-Origin: * response header correctly we can directly access such access points. (http://www.w3.org/wiki/CORS_Enabled)

(We are aware that solutions to circumvent this problem exists, like CORS proxys and YQL, this can not be considered as longterm solution.)

We should make sure that especially smaller endpoints go the extra half-meter to enable CORS!

## Tutorial for JavaScript newbies.
Another need identified is a introduction tutorial for newbies; not explaining JavaScript itself. The tutorial should be more directed in proposing some best practices for our RDF community. This could help to keep, the different efforts made to create tools and libraries, interoperable. And further helps new developers to shortcut a tedious exploration and decision process on the utils dimension. More concrete, can we as community propose the preferred module system(s): AMD, Node.js, Bower, … The preferred testing tool, the preferred compile-testing service and so on…

Everybody knows this is highly depend on philosophy and opinion, but possibly we can produce as community at least a proposition what to use in case of ignorance?

## IDE / Editor War!
And a last, more fun topic: What kind of IDE, Editors and so on are using? vi, emacs, sublime, … ?


Thank you for any inputs on the topics mentioned. Let’s give the developers some more reason to use the semantic web you build!

Regards,
Michael Luggen



On 28 May 2014, at 15:19, Michael Luggen <michael@oiu.ch> wrote:

> This E-Mail is mainly intended for ESWC2014 participants!
> ************************************************************************ (<- lots of stars)
> 
> Dear Javascript Enthusiasts,
> 
> This week the ESWC2014 is happening. Probably a significant subset of this mailing list is present in person. So why not meet for an exchange of thoughts about everything in between RDF and Javascript?
> 
> Lets meet Tomorrow 29. May 2014, 10:30 during the Poster Session in front of the Check-In.
> 
> And to stimulated you guys, ponder the following questions until tomorrow:
> What was the library/toolkit I was looking for in the RDF field lately? (for Client/Server?)
> Which code did I recently hack together, to make a RDF based application working?
> 
> Looking forward to meet you there!
> 
> Cheers,
> Michael Luggen
> 
>