W3C home > Mailing lists > Public > semantic-web@w3.org > November 2018

SemWeb LAMP stack [was Re: Toward easier RDF: a proposal]

From: Chris Mungall <cjmungall@lbl.gov>
Date: Wed, 21 Nov 2018 18:28:21 -0800
To: "David Booth" <david@dbooth.org>
Cc: semantic-web <semantic-web@w3.org>, "Dan Brickley" <danbri@google.com>, "Sean B. Palmer" <sean@miscoranda.com>, "Olaf Hartig" <olaf.hartig@liu.se>, "Axel Polleres" <axel@polleres.net>
Message-ID: <DB059410-AEDA-49EE-B3B2-6F0882CCE9EB@lbl.gov>

On 21 Nov 2018, at 14:40, David Booth wrote:

> 1. Tools are scattered.  How to find them?  Which to use?
> Every team wastes time going through a similar research and
> selection process.
>
> One idea: create a bundled release of RDF tools, analogous
> to a standard LAMP stack, or Red Hat or Ubuntu; so that if
> someone wants to use RDF all they have to do is install that
> bundle and they're ready to go.

I very much like this idea. I would particularly like to see a standard 
web stack that could be used with a triplestore of choice. Many 
developers love Neo4J becomes it's bundled with nice visual exploration 
mechanisms.

In contrast, with a typical SPARQL endpoint, you get something much more 
bare bones and it varies depending on which triplestore you use. It’s 
not super-enticing to non-geeks, and even for us SPARQL geeks, it’s 
not always clear how to get started.

Of course, there is a lot of great code out there that can be used to 
build a generic front end, but it’s all very disjointed, and many of 
us end up rolling our own code, e.g. for providing users with templated 
queries, or deriving a simple schema that can be used for exploration.

A killer app would be a ‘semantic web generic client in a box’. 
Simply start a server pointing at a SPARQL endpoint(s). Optionally 
configure it with some example queries, metadata about what property to 
derive labels (or introspect this from the endpoint itself), and get:

  - A form interface for SPARQL queries with a menu of example queries 
that give you an idea of how to get started, plus some useful standard 
prefixes
  - Simple linked data browsing (using conneg for html vs turtle)
  - Dumb generic graphical browsing, a la neo4j
  - Basic schema derivation; or just stats on most used classes and 
properties
  - A generic semantic hypermedia API
  - An autogenerated domain-specific swagger API (e.g. using garlik)

I believe a lot of these pieces exist already, it's just that we as a 
community need to do a better job of putting these things together in an 
easy to use package. This shouldn't be so hard now it's easy to 
containerize things and orchestrate them.
Received on Thursday, 22 November 2018 02:28:49 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:42:03 UTC