A kind of farewell

Hello all,

Unfortunately, it seems that I will not be able to contribute to the group,
at least in the following 6 to 12 months.

About two months ago we decided to stop using RDF+XML at z-Bible, and use
YAML instead for knowledge representation. (As a small startup, it's easy
for us to make such abrupt changes.)

Here are my reasons why we dropped RDF+XML. (We talked with Christian that
these could be included in the knowledge base of the group, and/or
incorporated in the publication.)

- RDF was invented around 1996, and still hasn't gained too much
popularity. I was wondering why. Now I think the main reason is the
over-simplified data model (everything is triples). It sounds extremely
elegant, but it turns out that even the simplest things become overly
complicated because of this. For example, if I want to store an ordered
list of objects (e.g. paragraphs of a text) and retreive them in order with
SPARQL, I have to do crazy hacks. It was Manu Sporny's blog post
<http://manu.sporny.org/2014/json-ld-origins-2/> that pointed me to this
conclusion.
- We use JavaScript, while I see most of you are using Java. In the Java
world, it is quite common to use XML for configuration. In the JavaScript
world, JSON is the de facto standard because it directly translates to
plain old JavaScript data structures. (We chose YAML as a human-friendly
syntax for JSON, but basically we use the JSON data model.)
- While we were using XML, we had to write importer code for each of our
XML-based file types (I was planning to solve this with an XML>RDF
conversion.) Now with YAML, we just import the data and use it, no extra
conversion code is needed. Though both technologies have advantages, for us
as a startup, the ability to make fast changes is the most important, so
YAML works better for us.
- And there is no decent SPARQL implementation for JavaScript. The only one
is quite buggy, and does not give good syntax error messages. So it seemed
that our editor colleagues will not be able to cope with it. It was also
way too slow for our purposes.

For the future, I have some thoughts:

JSON-LD was created fairly recently and it has already gained a relatively
wide popularity (it is the recommended LD format by Google
<https://developers.google.com/search/docs/guides/intro-structured-data>,
for example.) For its designers, it took several years of hard work, but I
think it was worth it, beacuse the world got a more user-friendly format,
and it helps the RDF/Semantic Web ecosystem move forward. Similarly, there
is space for developing more usable, more user-friendly formats for other
parts of the RDF world. And I think you guys now have the opportunity to
design something nice. So I wish you the best in this.

We plan to develop a format for knowledge to document (e.g. RDF to XML)
conversion. However, now we might end up with one of the (hunderds of)
JavaScript template engines out there and not use RDF at all. But if later
we happen to come back to a Semantic Web direction, I'll most probably come
back to this group.

But I'm thankful to you all, I have got some quite important insights in
this group. And it was a pleasure to work with you.

I wish you blessed work and nice results!

Bernát