Wednesday, 11 April 2007

Google News: news source - bug, or deliberate?






Google News is no longer in beta but it is, oddly, forcing "news source" searches on users.

You can do a "source" search by clicking "Advanced news search" and typing the source you want to limit the search to in the "News Source" box, outlined in red below - e.g. CNN or New York Times:


Or you can type "source:Reuters" (or whatever) in the main search box along with your search term:



But what if you don't want to do a search confined to, say, news agency Reuters as the news source? What if, instead, you want to search for news on Google News about Reuters?

Well, I've just noticed that Google News won't let you. It automatically fills in "source:Reuters" (or "+Reuters") if you just type in Reuters as a search term, e.g. if you wanted to check the news reports about the news agency Reuters setting up shop in the Second Life virtual world in late 2006.

If you type in the Google News search box: Reuters "Second Life":

and click Search News, you get this search results page:

In other words, Google News seems to assume that "Reuters" has to be a "news source", rather than the possible subject of any news, and makes it into one whether you like it or not - changing the search to source:Reuters "Second Life" (or +Reuters "Second Life").

Interestingly, it doesn't do this when you search the Google News Archives, where you get this on searching: Reuters "second life"


As you can see, the search results from the archives are from a mix of sources, not exclusively from Reuters itself.

Google News behaves in exactly the same way if you try to use e.g. "CNN" as a search term rather than a news source, e.g. search for CNN shooting and this is the results page, it's been turned into "source:CNN shooting":


I've not tried other searches yet with the names of other news corporations or press agencies, but I wonder why Google decided to do that. What's going on? Is it a bug? Is it deliberate? Did they come to some sort of deal with various news agencies?

It doesn't do it if you search e.g. "Financial Times" or "Times" or "Guardian" or indeed "New York Times" or "Washington Post" (odd that those wouldn't be deemed worthy news sources! Especially as the New York Times is listed as an example news source on the Advanced news search page). But it does change "BBC" to "source:bbc_news". Even more strangely, if you try "Bloomberg", it doesn't change it to source:Bloomberg in the search box, but it does say on the results page (just above the results listings) "Search news source Bloomberg" - yet the results clearly include reports from other sources e.g. the International Herald Tribune:


OK, I've done me share of typing names of random news sources to see what Google News does to them; if anyone else wants to sit there trying different names, feel free - but please let do me know what Google News does or doesn't "convert" into a news source! I can't find anything about this peculiarity in the Google News help or just on generally searching.

So, are we destined never to be able to read news reports on Google News about a news corporation (e.g. any takeover of CNN) - unless they're written by that corporation itself, or enough time has passed that it'll be old archived news?? That's the cynical view, of course...

Hopefully there's nothing Big Brotherly behind this. Sort of a variation on quis custodiet - who then will report on the reporters? How will we be able to find news about the news sources? etc etc. (Yeah, I know, before anyone suggests it: not on Google News!).

But still I think it's interesting quirk, if a little frustrating for the user who wants to search for news about a news source.

4 comments:

Tom Martin said...

Perhaps searching all new source makes the scope a little wide, limiting the results to one source could make it easier on the datacenters. I've not seen this before so it could just be a tempoary measure.

Improbulus said...

Thanks for your thoughts Tom. Course if you typed in "Second Life" without the "Reuters" it would no doubt search all the datacenters for "Second Life", so why don't they let you search them all for the word "Reuters" too?

I guess I don't like their thoughts about what I want overriding mine, especially when they get it wrong; if I want to search only Reuters as a source, I'd use source:Reuters or the Advanced news search page. I'd like the ability to use Reuters as a search term instead of data source if I so wish; as a user, I don't like the choice being taken away from me.

I do hope it's only temporary too!

TGOM said...

You should get the results you're looking for with the + sign.

When searching a source for a keyword use "second life" reuters.

When searching for a keyword that is also a source use "second life" +reuters.

Google News usually produces a link to the latter search right beneath the search box on the results page.

I tried the "second life" +reuters search and it gave results about Reuters and many other companies "setting up shop," which I was clueless about and find freaky, but nonetheless, Google is returning the right results, no?

Improbulus said...

TGOM, thanks for the comment. Yes the + works to stop

I should have checked the + link on the results page... I confess I'd assumed it worked the same way as in normal Google searches, so I didn't think it would add anything.

I didn't realise that it forces what Google News would otherwise convert into a source to remain as a keyword (which does make sense given the normal use of the + to force a search for "common words" like "the"). Thanks for clearing that up!

I do think that the link that appears under the search box in the results page e.g. "Search news pages that contain the term +reuters "second life" should contain an explanation.

I don't think it's obvious, otherwise. Though it's probably just me!

I'd still be interested to know exactly which words Google News convert into search sources, and why those & not other ones...