No clear entry point

[dbooth-boston]

"Compare "How do I write React" Google results with "How do I write RDF"
Google results. React's first hit[3] is served by its authority (reactjs.org). It links
to a description that is compelling, welcoming, and relatively easily
scanned. . . . RDF's first hit is hosted by w3schools.com[4] and feels scanty"
https://lists.w3.org/Archives/Public/semantic-web/2018Nov/0229.html

IDEA: Canonical documentation

"There is a canonical documentation and jumping in point for communities like React, and the documentation is very good, particularly for beginners. Does rdf have any of these things? Where is the starting point for RDF -- I don't even know -- W3C specs -- that is not the same ease of documentation for a self taught developer as it would be for an academic."
https://lists.w3.org/Archives/Public/semantic-web/2018Nov/0136.html

[dbooth-boston]

"I propose that we make a W3C community-maintained website for this
purpose, and use a github repo -- to be created -- to drive it, such as:
https://github.com/w3c/rdf "
https://lists.w3.org/Archives/Public/semantic-web/2018Dec/0127.html

[dbooth-boston]

Some comments from Henry Story:

"For questions and answers that can have an answer I would like for the moment to
https://opendata.stackexchange.com/ https://opendata.stackexchange.com/
with links to the tags that are relevant to our community namely rdf, sparql, linked-data,
semantic-web"

"We need to suggest a few other tags perhaps that would be of use. It would be good
to have some tag for publication."

"For questions that are open, this mailing list would be good, or a forum that is good
at leaving questions open, in the sense explained by Floridi in "What is a Philosophical
Question?"

"There should be a section on academic papers specializing in this, and conferences,
with perhaps a feed that people can subscribe to that collates all of them, so people
know both which conferences to go to, or where the specialists are they could approach
to answer their questions."

[awwright]

This is something I've been interested in developing, but I would like to note one difference between React and RDF, in that React does one thing: It is a library that generates a document from data in an event loop; whereas RDF is a data model targeted at many different use cases.

My idea of what this looks like is more along the lines of a recipe book: "How do I embed navigation information with RDF?" "How do I query a list of items with RDF?" and so on.

[ontoconsult]

Based on my own experiences there is definitely a need for a central source of information related to RDF, with subsources for the different directly related standards like RDFa and use cases.
A simple problem I always had, and still have at times, is that every standard seems to have its own rules for the composition of IDs. And what on earth could be the difference between rdfs:Class and owl:Class. These are problems that most of you cannot imagine because you have grown up in this RDF world.
So when you experts write something for laymen, I will give my comments as layman.
To illustrate how, for a layman, complex this all is I attach a graphlike diagram in which I tried to make visual how all these concepts are interrelated.

[dbooth-boston]

@ontoconsult, wow!

[dbooth-boston]

Melvin Carvalho wrote:

I just came across this :
https://github.com/semantalytics/awesome-semantic-web
Maybe a good start?

[dbooth-boston]

Nice! For a long time Mike Bergman has also maintained a database of
RDF-related tools, though a number of the tools on his list are now
outdated:
http://www.mkbergman.com/sweet-tools/

I am still wondering if there is enough interest to create a
community-maintained repo under W3C for this purpose. So far I have
heard agreement about the need for such a resource, but no concrete
offers to help create and maintain it. I am willing to help, but I
don't want to ask W3C to create such a repo unless I know that at least
3 others are willing to contribute.

If anyone is willing to help, please say so by leaving a comment here.

[mkbergman]

I would be willing to check activities for my existing listing and to possibly update them. However, I do not think I want to continue in a maintainer mode.

[iherman]

I am still wondering if there is enough interest to create a community-maintained repo under W3C for this purpose.

Just a word of warning... many years ago (in my previous life:-) I have set up a community wiki site at:

https://www.w3.org/2001/sw/wiki/Main_Page

I had two goals when doing that:

• giving a clear and unified reference (a URI...) with a short description to each SW technology of W3C (e.g., SPARQL, SKOS, etc.)
• provide a community driven place where various tools can be listed, categorized, etc.

It was/is a Semantic Wiki, which had the advantage of providing all kinds of tools to cross-reference things, categorize entries, and even give the possibility to generate linked data from the content.

Since I am not in charge anymore, I do not follow the evolution. It is still operational (I see some modifications from earlier this year on, say, Jena), but I do not think it was an overall success. It would be interesting to analyze why such sites (this was not the only one, that is for sure) did not really succeed in general; this may help us to avoid pitfalls for any new endeavor.

Some of my own recollections on this wiki's problems:

• Using (mainly setting up, actually) Semantic Wiki was fairly complicated (I seem to remember that I spent weeks on this thing), brittle, and very much non-user friendly. No offense intended to the creators of Semantic Wiki, it was a great tool for the time, but nowadays we'd need a better and much more user-friendly tool. "Just" setting up, say, a github list, for example, will not cut it in my view.
• It needs a maintainer. My goal was not to need (and be...) a maintainer, but this did not work out. I had to act as a maintainer for a long time: people needed help, I had to check the content, etc. Some of these are related to the user-friendliness (or lack thereof) of the site, lots of questions on the categorization, etc. I did it when it was my job, but when I moved on to other pastures, nobody took it over and it got into the process of bit-rot. I guess any site of the sort has this danger, and I am not sure how we can avoid it.
  One of the maintenance problems was, if I remember well, the inevitable issue of spam, which meant that the system people imposed, eventually, a stricter access control to the wiki.
• The quality of the tool descriptions greatly vary. It is actually a hard job to give a 1-2 paragraph description of one's tool; it may become understandable for the non-initiated, too concise or too verbose, etc. In this case this was mostly up to the submitter, but the overall quality of the site suffers as a result.
• It is a dry list of tools. We need demos, real-life application examples and references, sexy pictures, etc. Without trying to claim that everything is perfect in the Semantic Web land, we are also our worst enemies in our inability to sell what we have. This shows on this site, too...

I hope this helps...

[awwright]

I think this would be a good place to share what I'm working on: a framework for websites and especially wikis. It is like Semantic Mediawiki in some of its capabilities; it collects RDF information from the RDFa pages of the website (and possibly other sources), and can build tables of RDFa documents by filling in placeholders/variables where RDFa properties would be expected.

For example, this table of HTTP headers is built using a query: https://fullstack.wiki/http/http-headers

Here is the source code for that table: https://github.com/awwright/fullstackwiki/blob/ebab383f8c3308277982eea8e09e85bab5e6ce13/web/http/http-headers.html

Note how the generated document is an RDFa document that contains all the RDF statements used to build the document: the rendered document is partially isomorphic with the data store!

Anyone can check out the source code and build the website themselves: https://github.com/awwright/fullstackwiki. Several other repositories host libraries that support this, you can check the dependency list and my repository list for some of them.

I think this website would be a good fit for this task. I intend for it to be device-, developer-, and platform- neutral, and focus on best practices for accessible Web applications. I'd like to add content to this site focusing on Semantic Web technology.

To do this, we need to figure out exactly who the audience is, then come up with a structure of how it should look.

First, I ask about audience because I don't want to cannibalize the audience of the specifications themselves: Implementors should refer to the specifications, and not secondhand descriptions. However, my view is that specifications define should stick to normative requirements, and then put non-normative guidelines and logical conclusions in a website or wiki. We take this approach with JSON Schema; we publish Internet-Drafts for normative behavior, and describe design motivations and provide examples on the official website.

Second is the outline. We need to describe to developers why they should adopt RDF at all, and what problems it solves. Often RDF solves purely hypothetical problems instead of actual technical ones faced by a company.

So, I'd guess the outline list the problems that RDF aims to solve, then a technology or solution used to solve it.

I take this approach in a document I wrote describing how OAuth works: https://fullstack.wiki/oauth/index — note how it describes a single, narrow problem (e.g. how can a machine authenticate with a username/password), then presents the solution (use Authorization: Basic), then introduces another problem and its solution (passwords are expensive/insecure, use session tokens), and so on, until the entirety of OAuth (plus its foundations) are described.

But I'm having trouble coming up with a coherent, similar list of "problems that RDF solves". Things like 5-Star Linked Data describe very nicely how RDF is good for the Web, but why does it help e.g. me or my organization?

I think we need to describe some of the tooling available, and develop more & better tooling (see the other issues on this repository), in order to be able to justify how publishing RDF can solve important problems for the publisher.

I think I have a vague idea of what an outline might look like, but I'd like to see if other people have any good ideas before I spoil your minds with my own. Any ideas?