Wednesday, May 03, 2006

Google Exact Phrase Search is BUSTED ALL TO HELL!

Have you done any exact phrase searches in Google lately?

It's serving up all sorts of off topic shit, I mean, it's really busted.

Couple of examples of topics from this blog:

Do a search on "Lava Pockets" and you get crap liks this in listing #13:

Take Five - The Boston Globe(Correction: Because of a photographer's error, the names of two members of the Click Five were transposed in some edition's of Sunday's Arts ... - 41k - Supplemental Result - Cached - Similar pages
Neither the word LAVA or POCKET are even on that page

Even b
etter is "Florida Drinks Cat Piss" with 99% of the results complete shit!

Let's examine the #2 result:
eg Forums -> Pretentious bottle--ordinary wineOutline · [ Standard ] · Linear+. Pretentious bottle - - ordinary wine. Track this topic | Email this topic | Print this topic ... - 71k - Supplemental Result - Cached - Similar pages

That's right, nothing on that page about Florida whatsoever but the irony of wine coming up doing a search for cat piss certainly didn't pass me by.

Hey Google, you even know about SQA?

That's Software Quality Assurance for those that don't know, and I can recommend some very good people if you need them down at Google.


Anonymous said...


That along with what I've seen dealing with a site that was deindexed last year tells me that they have at least a screwed up view of what link/linktext pair is:

1. Located (and/or)
2. What link text goes with what link.

A screwed up view of the link structure of the web would also account for other troubling happenings.

The exact phrase search glitch has been reported to Matt Cutts, however it wasn't mentioned by the reporter exactly what he saw.

Google and SQA, hell, I'd first of all like to see real QA.

The last place I worked for regular wages the QA department was the most clueless lot I'd run into in almost 30 years. It was almost as if the filled out ass covering paperwork would make everything correct.


IncrediBILL said...

When it comes to QA, or QE (quality engineering, as you can engineer quality but not assure it) at my wife's company, people would get dragged out and publicly flogged for shit like this.

She runs a tight ship and her team is top notch, I'd trust them to test just about anything.

Then again, her background was managing a team testing financial spreadsheet-type products and there is zero margin for error in such products.

BTW, how do we know Matt didn't cause this bug? ;)

Anonymous said...

I've found that Matt is one cool cat and I'll cut him tons O slack ;).

It could all be a missing set of () (x-org had this for a problem on a couple of their more recent releases) or these {} or a == where it shouldn't be.

Maybe even a missing . (who knows maybe they write their super secret code in COBOL). I can remember finding one of those at 2:30 AM during a conversion (not a dry or test run, the real conversion).

Anonymous said...

Oh that case you are talking about is also one of the funnies going on with supplementals that g1smd has commented on.

Unknown said...

These searches work for me now.

IncrediBILL said...

Thanks Matt, THOSE are fixed, but some others are still broke, some I didn't want to publish publicly.