At around 1.10 in this excellent 14 minute introduction to the Semantic Web, Clay Shirky makes a provocative point:

"If I was going to start a news business tomorrow, I would start a business that was not designed to produce one new bit of news."

Web 3.0 from Kate Ray on Vimeo.

That is heresy for publishers. It sounds like aggregation and for goodness sake we have enough of that already! In fact Clay goes on to mention aggregation. But before we dismiss aggregation, imagine aggregation that is actually useful as a filter. Imagine in other words a startup that delivers on the mission:

"Aggregation that does not suck"

I don't like the term aggregation. Filter is more positive. We just need better filters.

No, sadly I have not yet discovered that magic aggregation/filter service. But this post will explore:

• why publishers hate aggregators,

• why Clay Shirky is right,

• what the new publishing stack looks like,

• what role the Semantic Web can play in this.

Publisher's Instinctively Dislike Aggregation

Primal is a semantic web startup that debuted recently (our coverage here). They enable individuals to aggregate content.

Here is one reaction to Primal, from Richard MacManus of Read Write Web. Richard started as a blogger and now runs a successful online publishing business. (I used to work there). Read Write Web produces excellent content, holding their writers to high standards. They prove that blogging standards can be just as high if not higher than traditional journalism.

Richard rightly skewers Primal for their hype - "semantic synthesis platform" is a bit much! But Richard really notes that:

"this is reducing Web publishing to its most basic form, devoid of new content."

Later he returns to this theme:

"However, as noted above, it had no original content on it - which means it doesn't add much value to the Web as a whole."

This idea that producing original content is the core purpose of a publisher is deeply ingrained. That is why Clay Shirky's statement is provocative.

Publishers Have Never Produced News, They Report It

Of course we don't. The guy who kicks a ball into the net in the World Cup produces/creates the news. The person who starts a revolution in Iran produces/creates the news. The entrepreneur who raises funding and launches a new service produces/creates the news.

All we do is report that news.

But the problem is that the world has about 4.6 billion journalists. That is the number of people with cell phones as of the latest report from International Telecommunications Union as reported here on CBS.

That is 4.6 billion people who can say "wow, xx just scored". Or "somebody just threw a molotov cocktail into the presidential palace" or "we just scored $5m from SandHillMucketyMucks".

Blogging opened one door. Anybody can become a publisher. A few actually did that - folks like Richard MacManus, Mike Arrington and Om Malik. Blogging changed the game. Traditional publishers hated it and then became bloggers as well. The bloggers became serious and adopted the best bits of traditional media. Convergence happens!

But Twitter changed the game yet again. And so does Facebook with status updates. Given 4.6 billion people with cell phones, Twitters "Big Hairy Audacious Goal" of 1 billion users does not seem crazy (they have 190 million today).

While Twitter and Facebook fight it out as the place you go to to send an update, to the rest of us that really does not matter. We (publishers, users, call us what you will, we are all publishers today) can consume updates from both Twitter and Facebook and every other source via real time aggregators such as Superfeedr (see here for their great analysis of the real time web for publishers).

In that world, something like Primal becomes relevant. I say "something like" as Primal seems pretty raw to me at the moment, great UI but the data or the semantic matching was weak on my few quick tests. So I would agree with Richard that Primal is in the "wait and see" category.

We Are All Writers, Soon We Will All Be Editors

The current term is "curation". It sounds very like what we used to call "editing". At it's most basic this is something like Digg or Facebook Like. At a more sophisticated level, this is scientific peer review. These two worlds may converge, so that we get tools that curate/edit complex/sophisticated information (we look at that in the context of scientific peer review here).

What an editor does at a basic level is say "this is good, this is not" and if they are doing their job well "here is why it is good/not good and here is how you can improve".

Today, the norm is still for editors and journalists to work for the same organization. Some journalists get promoted to editor.

But in a world where everybody is a journalist, we still need editors (or curators if you prefer that term). When thousands of tweets report exactly the same news, basic machine-level filtering is needed. The automated aggregation can also help at a course level. Beyond that you have to start engaging humans. And those humans will have different weight as far as the final editor is concerned. You don't just want "crowdsourced" editorial, you want "expert sourced" editorial.

When a story is big you might want to be become the "go to" source for that story (in other words to become the top link in a Google search). So you might invest in paying for a) some research work b) your own expert to weigh in with an opinion.

The new publishing stack looks something like this:

And The Biggest Vote Goes To....

In that whole stack, one user gets a super vote. That is the person who is "consuming" the content, the "people formerly known as audience". It is not that they are smarter or better informed than anybody else - they probably are not. It is simply that only the end user knows the context within which this information is valuable.

Linked Open Data: Stop Recreating The Basics

What services such as OpenCalais and Freebase and DBPedia (and maybe Primal) do is save the time to aggregate the basic contextual information. So when you spot the Tweet that says "xyz just scored $5m from abc for dfe technology", you need to leverage all of that stack to create a post that puts the news in context. You will need the basic facts on "xyz" and "abc" and "dfe".

That should come via Linked Data sources.

Then you want to leverage the higher layers of your stack - crowdsourced votes, expert votes and so on.

In none of those things have you produced any news. You have simply leveraged the tools and open data to put that news in context.

-----------------------------------------------------------------------
• Don't forget to propose your startup for our Semantic Web Impact Awards. The deadline is Sept. 15.