Any chances of getting rid of the redirects in search results? It's so annoying that I can't simply copy the result link URL and paste without getting some crazy Google tracking URL instead. It's such an obvious fail that I can't see how it's possible to not have it fixed yesterday, it made me switch to a different search engine.
I don't think we'll get rid of this any time soon. This is rolled out only for a subset of the users [1], and I guess it provides Google with valuable analytics data.
[1] As I can read in several articles in the internet, but can't find an authoritative source.
A very brief test indicates that search strings containing "Cocoa Webview" no longer return a bunch of useless crap* about Android webviews with the word "Cocoa" nowhere to be seen on the page. If that's the case: it's about time.
*"Useless crap" is context-sensitive, of course -- no disrespect toward Android intended, but I go to the trouble of including "Cocoa" in my search Google shouldn't be giving me pages that don't even contain that word. It might be different if it were smart enough to just give you Objective C or iOS-focused pages that included "Webview", but not Cocoa, since there's at least some probability that those pages would be of interest. However, it wasn't that smart.
The one that's been driving me crazy lately: 4 as a synonym for "for." I understand some percent of the population "rites lik dis 4 serch" -- but I certainly don't! When 4 = for, and "for" is on every page on the internet, it almost had me pulling my hair out.
Here's an example (but not my exact query). "iPhone 4" will match pages with "iPhone" and "for."
It should be easy enough to write an extension or even a bookmarklet for this. You just have to add &tbs=li:1 to the end of the URL to enable "verbatim" mode.
> Improvements to freshness. [launch codename "Abacus", project codename "Freshness"] We launched an improvement to freshness late last year that was very helpful, but it cost significant machine resources. At the time we decided to roll out the change only for news-related traffic. This month we rolled it out for all queries.
> More precise detection of old pages. [launch codename "oldn23", project codename “Freshness"] This change improves detection of stale pages in our index by relying on more relevant signals. As a result, fewer stale pages are shown to users.
Hopefully this means fewer instances of searching for code help and seeing prominent results from 2005.
>Improvements to handling of symbols for indexing. ...we’ve now started to index the following heavily used symbols: “%”, “$”, “\”, “.”, “@”, “#”, and “+”. We’ll continue to index more symbols as usage warrants.
That, and the new "Verbatim' option, are really good news. There are also some five (10% of the entire list) timely tweaks to synonyms.
All in all, it seems Google listens and tries to improve its service, which has become rather paternalistic lately ("you wrote X, but actually mean Y", no I don't!). Well done.
I found relevant results for $_, $argv, @memoize, useful completion but no results for $@ and $#, nothing at all for $!. Python programmers rejoice, Perl programmers still mourn code search.
With all the bashing around them stagnating in search, it's nice to see that they are continuously improving even if it's more incremental than revolutionnary.
I particularily like the indexing of symbols, it will make searching for C# results easier.
Click through to the video of their search meeting. It seems really strange to me that something like that would be made public. I mean, it is pretty content free, but why spend so much effort to show the world deliberations on how to change search?
I find it amazing that there are 30-something people in the room & 9+ more by video.
Some are almost certainly checking their email or doing other things – despite the small text insert that tries to suggest they're all looking at advanced analytics about the proposed changes.
Very few speak or seem to be in a physical (or status) position where they could speak, though maybe as the topic changes, some of the giant peanut gallery gets engaged.
Altogether, it fits a lot of meeting antipatterns. When I see perplexing things like this about Google, I have to wonder: is this a counterintuitive part of the secret sauce, or an indulgence whose cost is covered up by their gusher of quasi-monopolistic profits?
That's a weekly meeting where proposed changes are discussed. Most of the people not speaking are there because they have a launch to present. Each one gets a lightning round of deliberation somewhat like the one in the video, a decision is made, and they move on to the next.
Because it gives a sense of transparency to the process, when in reality google is far from transparent. If they made a big deal about how no one could see the meetings behind the curtain, the secret mystique might erode some of google's goodwill in the eyes of the public - especially if the us govt got involved.
Most people, if they don't have a strong interest in tech & the Internet, only see the surface PR efforts, without having the background to notice that they're in fact heavily orchestrated PR.
The strategy certainly seems to be working thus far - but I feel they might be pushing it a bit recently. More people seem to be noticing and thinking about it.
Take a look at when google chooses to release news, especially "fun" news like robotic cars. They are great at the art of PR - which is not really a surprise if you own a tinfoil hat, because you never know - maybe they can get leading indicators through query & behavior mining across their dataset? :)
The new additions of UEFA, KHL, and tennis scores would seem to indicate that each snippet type is manually built and added by Google.
I would have expected something a little bit broader; an algorithm that attempts to automatically parse score results for all major sports based on the pages they pull results from, or something to that effect. The former is way easier, and probably a whole lot more accurate, that the later, but it removes a little bit of the sex from the whole process. It also highlights how much work must be involved in building the snippet generators for a site like google.
Tools and systems that take a corpus of root text and create large number of variations of that content, to avoid duplicate content penalties when trying to stuff sites to catch Google's attention.
> Implements and setups that take a corpus of tuber text and create meaty number of changes of that tranquil, to avoid duplicate content sanctions when trying to stuff localities to catch Google's contemplation.
Like serkey.com? God I hate that site; it came up with most of the front page for pretty much every Oracle related query lately. That it until these changes, apparently.
