W3C home > Mailing lists > Public > public-lod@w3.org > October 2017

What Happened to the Semantic Web?

From: Sean B. Palmer <sean@miscoranda.com>
Date: Wed, 11 Oct 2017 12:06:30 +0100
Message-ID: <CAH3-oEcKZ135Up-KK1KF+=K2HAi3UTeOwXF0TN_27cFSxeVgrw@mail.gmail.com>
To: SW-forum Web <semantic-web@w3.org>
Cc: www-archive <www-archive@w3.org>, public-lod@w3.org
I was curious as to what was going on with the Semantic Web these
days: where did it go, who's using it, who's talking about it? As
usual I decided to chat about it in the Semantic Web Interest Group
channel #swig on Freenode, but since there is no logger there I
scraped the log out of my IRC client and am posting it here.

In case you don't want to wade through the waffle, the summary is that
there were roughly four phases of Semantic Web development starting
with the eponymous golden age: Semantic Web (2001-2005), Linked Data
(2006-2010), JSON-LD (2010-2014), and now the Data Activity (2015-).
The biggest tangible results are Schema.org in conjunction with
JSON-LD for SEO, Dbpedia, a few tacky database products, and very
loosely the API economy.

The only work that I could find still taking place on the Semantic Web
is under the W3C's Data Activity, which is very quiet and of limited
scope. On the other hand, the conceptual legacy of the Semantic Web is
still quite strong, and I make some notes on that in the log. As I
say, the Semantic Web was originally the conception of graphs instead
of trees, with global symbols, published in an open and decentralised

The main problem seems not to have been the proliferation of the
Semantic junk such as RDF/XML and SPARQL, as is sometimes argued
(references below), but rather that the Web side did not provide an
open and decentralised system in which to host the Semantic side.
Ongoing efforts are being made to rectify that, but there are no
promising solutions in that domain and so many of the Semantic Web
ideas will remain dormant.

<sbp> looking at how active Semantic Web stuff is now, out of curiosity
<sbp> it seems that http://planetrdf.com/ is just a CFP spam site now
<sbp> there is some activity on https://www.w3.org/wiki/Special:RecentChanges
<sbp> none of the recent stuff seems to be SW related though
<selckin> some people refuse to give up
<sbp> let's try
<sbp> selckin: who?
<sbp> only a few edits to this over the past year
<sbp> mostly on https://www.w3.org/2001/sw/wiki/ShEx it seems
<sbp> appears to be some sort of not very good schema thing, though
ericP did make one of the implementations
<sbp> the http://answers.semanticweb.com/ site that the IG page links
to doesn't even respond
<sbp> and https://www.w3.org/2007/11/Talks/search/query%3fdate=All+past+and+future+talks&%3bactivity=Semantic+Web&%3bsortInverse=yes
is a "Discontinued service"
<sbp> the CG page at https://www.w3.org/2001/sw/CG/wiki/ asks for log in
<sbp> the RDFa WG is listed as active but according to
https://www.w3.org/2010/02/rdfa/ it was closed in 2015
<sbp> same with the RDF WG but
https://www.w3.org/2011/rdf-wg/wiki/Main_Page says that was closed in
<sbp> the LDP WG closed in 2014 too, https://www.w3.org/2012/ldp/wiki/Main_Page
<sbp> ah, an explanation: "A few days ago W3C started a new activity,
called Data Activity, that also subsumes the (by now old) Semantic Web
Activity." — https://www.w3.org/blog/SW/
<sbp> strange that the SW page itself doesn't explain that
<sbp> heh, dsr is the lead. well how about that
<sbp> they have a "Dataset Exchange Working Group",
<sbp> trying to work out what this is
<sbp> they were working on DCAT, but that was RECed in 2014:
<sbp> oh I see. it goes dc > dct > dcat
<sbp> then there's https://w3c.github.io/poe/vocab/ which is a rights
policy thing
<sbp> I dunno if the W3C should be working on anything rights related
these days. they seem to abrogate their responsibilities to society as
hard as they can
<sbp> there's also https://w3c.github.io/dxwg/ucr/ which says that
DCAT has some shortcomings they want to address. that's been three
years in the making then...
<sbp> (uses ReSpec. "ReSpec is a document production toolchain, with a
notable focus on W3C specifications.")
<sbp> there's also the POE WG, which is actually doing the rights
thing rather than DX WG: https://www.w3.org/2016/poe/wiki/Main_Page
(see above)
<sbp> SD WG are working on spatial data:
<sbp> ah, and the Data Shapes WG seems to be covering the ShEx thing I
found above: https://www.w3.org/2014/data-shapes/wiki/Main_Page
<sbp> seems to have been active, with Sandro no less, up to 2017:
<sbp> shame really; I thought Sandro would continue to go on to do
more interesting things
<sbp> it's like when you see tech stars from the 1980s and they're on
twitter moaning about Trump
<sbp> and then others are writing lisp machine hardware verified in
Coq or something
<sbp> heh, https://www.w3.org/2013/dwbp/wiki/Main_Page is said to be
active but was closed this year
<sbp> aand the Data Activity blog hasn't been updated since 2015:
<sbp> .t https://dl.acm.org/citation.cfm?id=3078751
<yoleaux> sbp: Sorry, I don't know what timezone that is. If in doubt,
see https://en.wikipedia.org/wiki/List_of_tz_database_time_zones for a
list of options.
<sbp> what
<sbp> .title https://dl.acm.org/citation.cfm?id=3078751
<yoleaux> 403 - Forbidden Access to The Digital Library
<sbp> "hat Happened To The Semantic Web?"
<sbp> no PDF of the paper that I can find, but slides:
<sbp> says that 2001-2004 was the golden age
<sbp> then 2006 introduced Linked Data
<sbp> ah, there we go: "Current status (2017)"
<sbp> reduced standardisation (not kidding), narrower research focus
(is there ANY research?), and maturing technology
<sbp> the technology is Neo4j, Virtuoso, and Allegrograph. plus Oracle
and Microsoft stuff that I won't even bother to mention
<sbp> says that a big problem was that ontologies were centralised
(sigh), and data held privately (well, we should have seen that one
coming). also, no use of Linked Data because of trust issues! what
happened to the crypto that I helped timbl add to CWM then, eh? eh?
<sbp> mentions FOAF eventually
<sbp> quietly mentions the fact that the web has changed, data silos, all that
<sbp> "Some useful data", cites schema.org, dbpedia, and wikidata
<sbp> not exactly a competitor to WolframAlpha though is it?
<sbp> funny, really. again one centralised site is the leader in a
space that should have been covered by Semantic Web technologies.
actually that would have been a really nice thing to push for.
probably still would be, though I tend to want my local computer to be
able to do those sorts of calculations
<sbp> like the other day I wanted to compute the prime factors of a
number. didn't really have anything lying around that does it in the
stdlib, so eventually I downloaded sage. but that's 4 GB of stuff!
you'd think we'd be able to get batteries included standard library
balances right by now
<sbp> .title https://www.quora.com/As-of-2015-is-the-semantic-web-dead
<yoleaux> As of 2015, is the semantic web dead? - Quora
<sbp> heh, hadn't seen http://5stardata.info/en/ before
<sbp> Alan Morrison answers: "The thing about the semantic web idea is
that the development cycle has turned out to be the opposite of what
Tim Berners-Lee anticipated in the early 1990s. With the semantic web
concept, TBL hoped for a giant structured open web. What we got first
was a bunch of closed structured webs and a mixed web that still isn’t
very well structured."
<sbp> it's not like open things can't work. Wikipedia is a good
example of that, despite the deletionism
<sbp> Shidan Gouran: "Definitely RDF/XML is dead as a data exchange
format, and soon, XML will become obsolete as well, which is a good
<sbp> I haven't seen anything XML based in years, except I guess for
the continued use of SVG
<sbp> on the other hand
argues that the SW achieved what it set out to do, just somewhat
<sbp> it seems to base that conclusion on schema.org alone
<sbp> .title https://medium.com/@dennybritz/the-semantic-web-is-dead-long-live-the-semantic-web-eddbca0a8b6
<yoleaux> The Semantic Web is dead. Long live the Semantic Web.
<sbp> "But let’s stop arguing over details and terminology. Let’s
forget about RDF, SPARQL, OWL, triple stores and quad stores for a
moment. Instead, let’s take a step back and look at the high-level
goal of the Semantic Web."
<sbp> yeah, well, it did get lost in those details
<sbp> kinda hard to ignore that
<sbp> "If you are a developer you may be able to query APIs or write a
custom scraper to get the data you need, and then write code to
aggregate them into something meaningful. But in many cases even that
wouldn’t be feasible due to technical, legal or time constraints."
<sbp> basically argues that the contemporary web didn't learn from the
Semantic Web at all
<sbp> also asks: "Did the API Economy succeed where the Semantic Web failed?"
<sbp> I find the API economy such a pain in the arse. I tend to write
scrapers instead, because I find that those are usually more stable
than the APIs. it's ridiculous how often I have signed up for a key,
used a JSONic API, only to find that either the key issuing process
changes and I have to sign up for a new one, or the API breaks in any
of a myriad unfathomable ways
<sbp> "Technologies like JSON-LD may change this."
<sbp> I have never seen any JSON-LD
<sbp> three years ago: https://news.ycombinator.com/item?id=8510401
<sbp> "JSON-LD, which is a profound improvement on and compatible with
the original RDF, is the only web metadata standard with a viable
<sbp> 'On a more down-to-earth level, there is now a solid web
metadata standard in place in JSON-LD. The big search engines index it
and presumably use it to give better results. Any startup can add
value to published data by adding links - in a significant extension
to the "API economy".'
<sbp> *lots* of chatter about JSON-LD here
<sbp> from 2010, https://en.wikipedia.org/wiki/JSON-LD
<sbp> but it's funny because Wikipedia says: "The encoding is used by
Google Knowledge Graph[6] and others[who?]."
<sbp> one of the top results for JSON-LD is an SEO weblog saying: "In
this post, we'll shed some light on Schema.org and JSON-LD. What is it
and how can you put it to use for your website?"
<sbp> lots more SEO posts too
<sbp> something more recent:
<sbp> so I guess JSON-LD was the latest fad, covering about 2010 to 2014
<sbp> mostly in conjunction with schema.org
<sbp> Manu Sporny on the creation of JSON-LD:
<sbp> "RDF is a shitty data model. It doesn’t have native support for
lists. LISTS for fuck’s sake! The key data structure that’s used by
almost every programmer on this planet and RDF starts out by giving
developers a big fat middle finger in that area."
<sbp> http://manu.sporny.org/2014/json-ld-origins-2/
libby has left IRC (Quit: libby)
<sbp> I mean, it absolutely did have lists, and N3 showed how to do it
<sbp> it didn't have them as an enclosed object though. I wonder if it
should have done, a bit like how Haskell has Text now (opaque blob) as
well as String (inductive data structure)
<sbp> you can unpack one to get the other. Text is for
representational efficiency
<sbp> 'That said, after 7+ years of being involved with Semantic Web /
Linked Data, our company has never had a need for a quad store,
RDF/XML, N3, NTriples, TURTLE, or SPARQL. When you chair standards
groups that kick out “Semantic Web” standards, but even your company
can’t stomach the technologies involved, something is wrong.'
<sbp> now he can add JSON-LD to that list :-)
<sbp> lots of people I remember mentioned in the Postscript
<sbp> there's also a bit on
which takes the Betteridge's law tack
<sbp> so we had the Semantic Web (2001-2005), Linked Data (2006-2010),
JSON-LD (2010-2014), and now the Data Activity (2015-) rump. some
interesting phases there. I stopped working on this in 2007, so that
explains why I hadn't heard of JSON-LD
<sbp> (actually I had)
<sbp> these days I tend to think that the problem with the Semantic
Web was not the Semantic part, it was the Web part. renting a domain
from ICANN, putting it through the torture of DNS, serving things over
HTTP/2 (Google Lightning™), through centralised CAs, it's not the
easiest thing to do and it's not the cheapest thing to do
<sbp> so people share through centralised services, and we get the
problem that Peter Mika pointed out
<sbp> but the original idea, which is something like upgrading the
lisp machine from trees to graphs (incurring a bunch of
algorithmically intractable problems like graph isomorphism in the
process, whoops! nobody ever said timbl was a great coder), making
symbols global and universal, and then linking together all the graphs
in a decentralised system, was actually a pretty good one
<sbp> if only we had a decentralised system to put it into, eh?
<sbp> these days we have the blockchain for name resolution, but
nobody has figured out a decent DNS alternative through it yet (there
have been many attempts). actually the best alternative is still Tor's
.onion, which is kinda funny. and as for decentralised storage, IPFS
was the closest there, but that seems to be losing traction too
<sbp> so no Semantic Decentralised Web is possible just yet
<sbp> so yeah, the Semantic Web may be out of the golden age, but I
think like lisp it'll be one of those things that sticks around and
gets rediscovered over and over, continually informing (and warning)
the future

Sean B. Palmer, http://inamidst.com/sbp/
Received on Wednesday, 11 October 2017 11:07:02 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 11 October 2017 11:07:03 UTC