19 comments

[ 4.8 ms ] story [ 61.3 ms ] thread
Slightly OT, the given answer makes more sense than some (semi) organic result I see for a search for 'population nairn'. It returned a "people also ask of":

"What is the fastest town in Scotland"?

Answer: walkwithdinosaurs

Nonsensical question and answer, not related to the query. Clicking through to the result you can see how G turned a statement into a question, just the why doesn't seem to too clever.

The question makes sense. It’s a common joke in Scotland. “What’s the fastest town in Scotland? Naaairnnnn… “ (imagine car noises)

The “answer” on the other hand…

Edit: Just did a search and the answer comes from the top result which is a blog with that name. That blog has the joke posted on it. Google could be better here, but I think it's a little unreasonable for Google to understand jokes.

I must live too far South to have heard that one, makes more sense to me now. I guess Google's seen it often enough posed as a question that it has decided it's relevant enough for a wider breadth of Nairn queries.
> It appears that Google search uses the following syntax for conversions between different units or currencies: target_unit! <expression> (Though I can't find any documentation for this).

I've always been baffled by why Google search — the flagship product by the world's largest tech company — has so many totally undocumented features.

I understand having some easter eggs (the question "did you mean 'recursion'?" when googling "recursion" ), but even the sort of basic features noted in TFA seem totally undocumented.

I suppose they do this because it makes it easier to change features. But that is silly because the undocumented features don't get used.
> I suppose they do this because it makes it easier to change features.

Yeah, and for a scrappy startup trying to achieve feature parity despite having far fewer employees, that might make sense.

But it's not like Google is a nimble startup without the luxury of hiring dedicated technical writers.

I also think that documentation implies long-term support, and a commitment not to change the expected behavior. Any time you document something, you are telling people that this is not only the way that the software works now, but also the way that you intend it to work in the future. That established intent seems anathema to Google's culture, with the amount of churn and project cancellation that exists.
> I've always been baffled by why Google search — the flagship product by the world's largest tech company — has so many totally undocumented features.

If I can speculate on their overall behavior, there's no intent behind this. They are relatively decentralized, and everyone is throwing stuff at the wall and seeing what sticks.

To have a feature known you need concerted effort by design teams to make it accessible, marketing to make it known and so on. There's nothing of the sort with Google.

A tiny feature like this can be added by a single engineer in a few hours/days. They will require 1 or 2 other team mates to code review.

Adding public documentation of that feature will require translation into all supported languages, which will incur a cost, and sign off of those costs will require approval from a whole slew of higher up and other-team people. It will probably involve processes and systems the engineer has never used before. Translations are often batched (so that a single human does all the translations to avoid inconsistency of terminology), so it might be months till the next batch is done. The copy will need to be reviewed and rewritten by various technical writers.

Suddenly the 'spend a few hours adding a tiny little feature' becomes a week or two...

I think they have no bearing on the quality of the actual product. If anything they are Easter eggs that promote discussions like this one which is free marketing in itself.
The theory that this is a very dry joke is cogent.
First -- I think the person you quoted is just wrong, it's not a feature at all.

Google "cm! 5 inches" or "cm 5 inches", or even "5 inches to cm" or "5 inches cm" and you get exactly the same result in all cases. The exclamation point isn't a feature, it's irrelevant.

But to answer your question more broadly: basic search syntax is documented here [1]. But then everything else -- like calculator or conversion functionality or snippets -- just falls under the broader umbrella of Google trying to give you answers directly whenever possible, rather than just links. And that isn't documented, first, because it's changing/improving constantly, and second, there's just too many custom features like that for any consumer to ever keep track of, so what would even be the point?

[1] https://support.google.com/websearch/answer/2466433?hl=en

> that isn't documented, first, because it's changing/improving constantly, and second, there's just too many custom features like that for any consumer to ever keep track of, so what would even be the point?

Same point as any documentation, to make information about features and changes available to the people who DO want it.

> Google "cm! 5 inches" or "cm 5 inches", or even "5 inches to cm" or "5 inches cm" and you get exactly the same result in all cases. The exclamation point isn't a feature, it's irrelevant.

That appears to be correct, and I added a comment to the quoted answer to that effect. Thanks!

On the broader point, you are also correct – and, in particular, the calculator/unit syntax is documented at https://support.google.com/websearch/answer/3284611

I was mistaken; Google provides better documentation of their search than I thought. I appreciate the correction.

If you have to do a search to find documentation on searching... :-(
There's a time to parse mathematical expressions, a time to leave refrain from parsing altogether.
Only slightly related, but this reminds me of searching on Google “how many electrons are in the Sun”. The quick answer page parser fails beautifully.

I wonder how often Google returns something wrong or completely unexpected for its quick answer result.

About half the time in my experience