Comment by liamzebedee | Hacker News Reader

Comment by liamzebedee | Hacker News Reader

Parent Full thread

Definitely. I agree completely. For someone relatively uninformed on the scene, the NYT certainly seem like a leader in the news industry for going forward with new digital media [1].

I really like the idea of Particles. I had a similar thought this past year about building a news MVP that simply consisted of atomic facts (no more than a sentence) with an attached probability and discussion. Kind of like those you see in the IPCC (international panel on climate change) reports, where they have their claim and degree of certainty.

We were getting there with the whole Wikidata initiative, but it's still in its primitive stages with respect to content being statically authored and updated (AFAIK). It would be very interesting to see where we go with GraphQL - I could see that becoming the dominant machine-to-machine protocol for Semantic Web 3.0. The whole idea of taking what are usually REST resources and turning them into hierarchical JSON documents which you can query very flexibly is immediately appealing.

If this Particle concept catches on like cards have in UI design, I'd expect Twitter to be the first to ride the wave.

[1] http://www.niemanlab.org/2014/05/the-leaked-new-york-times-i...

Larry Sanger, Wikipedia's co-founder, already tried to make a site using atomic facts. It was called Infobitt, and lasted about a year or so before he closed it down a few months ago. The problem with using the fact as the smallest data unit is that it invites fact checking, which slows down the news reporting. Ironically, if sanger had actually used his system to create Wikipedia pages, it would probably have had more success. Wikidata is not going to work. It suffers from the same issue, as well as having problems relying on Wikipedia's "facts".

As for Twitter, I doubt they can do it either. Their "Moments" initiative shows how little they understand curation.

Vox is doing a pretty good job of it. The NYT already has quite a large number of topic pages on a variety of subjects [1], [2]. It's not Wikipedia, but it's not a small number of articles either.

[1] http://topics.nytimes.com/top/reference/timestopics/people/i...

[2] http://topics.nytimes.com/top/reference/timestopics/subjects...

I checked one of the feeds [1]. It's just a list of existing articles, by NYT writers, that don't have meta data. That's not what the NYT article is about. To make the meta tagging work you have to also strip down the article into each individual news event. For example, imagine a list of NYT articles about Yoko Ono. Almost every article about her will mention that she was married to John Lennon, and that he died in 1980. In an article you have to repeat this information because you assume that the physical paper was thrown out each day. But in a news database these events would be separate data events, and would not be repeated. Vox is the same - they still use articles, not data in their place.

[1]http://topics.nytimes.com/top/reference/timestopics/people/g...

There is actually a job title called "Fact checker" at newspapers. That's why people tend to trust high end newspapers like NYT over some random blog that just posts whatever they hear asap to get traffic. I doubt NYT plans on opening up their particles to any randos on the web. They already have great in house employees who can do this job very well.

> The problem with using the fact as the smallest data unit is that it invites fact checking, which slows down the news reporting.

Could you elaborate on this? Specifically, what do you think is the better alternative?

Let's say the news is about a plane crash. At first it is reported that there are 10 deaths, but a few hours later it's reported that there are 100 deaths, and then 99 deaths. So imagine we are writing an article for a newspaper, or even a Wikipedia article. The reporter or contributors are trying to find out the truth, because their objective is to create "the perfect article". So when the later death toll comes in they will often rewrite the original article to make it correct, but in doing so they hide the original reporting. The reader will see an article that says: Plane crash: 99 dead.

Newspapers don't print feeds, there's no room. But if you have the archive to hand, as a database, then you don't need to make a perfect article. You just need to add updates as they happen.

Event 1: Plane crashes 10 dead Event 2: 100 dead confirmed Event 3: Actually 99 dead

This leaves the original events as a record of what was reported at the time, but the reader can also see the update. This is important to readers because while information can be added to articles, it can also be taken out, to try to conform to a narrative. Wikipedia is particularly prone to this kind of selection bias dressed up as fact-checking.

Confused: correcting mistakes is 'selection bias'? How is incorrect early reporting, news at all? The act of reporting isn't the news; its the thing that actually happened.

Often we don't know the facts until much later. Even then the facts may be a matter of opinion, or may have conflicting sources. The gun control debate is a good example where facts on both sides are used to support each side's biases. On Wikipedia we see one side using a particular set of facts and then deleting the other side's facts. The same is true in newspaper articles where certain awkward facts are conveniently left out of reporting and analysis. But what if both sides are correct? We'll never know what is the truth because there are always different ways of looking at something.

In many case, fact checking is often a way to hide bias under a veneer of authority, by proclaiming a selected set of selection bias facts as "the fact-checked truth". You only need to follow the major fact checking sites for a short time to see this in action.

The question is how can you effectively present conflicting information. I believe the article format is biased from the outset, whereas a more data-driven approach leads to less bias.

Infobitt is still technically live, but it is so riddled with problems I don't know where to start.

Reply on news.ycombinator.com