Wikipedia talk:What Google liked
From Wikipedia, the free encyclopedia
I renamed this page from "Wikipedia:What Google likes" to "Wikipedia:List of articles frequently visited through Google", because it is the combined result of what users of Google like and how Google indexes. - Patrick 11:03 Apr 25, 2003 (UTC)
- I thought the old title was more memorable. Google, as an inanimate object, can't "like" things, so I think the possibility of confusion is minimal. Ahh well - aesthetics... Martin
-
- It was not clear, it could mean something like Wikipedia:Subjects for which a Wikipedia article is in the Google top 10, A-K. - Patrick 11:44 Apr 25, 2003 (UTC)
---
How do one see the statistics related to this page? It would be easier, then, to update it. Yves 17:34 May 15, 2003 (UTC)
For that matter, it would be helpful if we could somehow see failed searches within Wikipedia, where "failed" maybe means a search that was performed without clicking on any of the results (I know this may be hard, particularly for non-cookied users, but perhaps something is better than nothing). A lower-tech tactic would be to include a prominent link to the requested articles page on the search page, to make it easy for users to record what they were looking for. Zashaw 02:39, 19 Aug 2003 (UTC)
This page hasn't been updatd since March and is getting a bit stale. Do we have an update coming soon? --Raul654 19:27, 18 Oct 2003 (UTC)
[edit] From VfD
- Wikipedia:List of articles frequently visited through Google hasn't been updated since early March. I posted in the talk pages in early October asking if there was any plan to update, and no one responded. Super-stale pages like this give a false impression of where effort should be concentrated. --Raul654 11:14, 14 Dec 2003 (UTC)
- Unfortunately, I have to agree. If a time-sensitive page isn't regularly updated, we are better off without it. Rossami 16:00, 14 Dec 2003 (UTC)
- Don't delete, but restore to former title "What Google Likes." I've edited the page some, to make it so it is not so time-sensitive. It is linked from quite a few places under stats under the heading "what google likes." The present heading is inappropriate and should be updated regularly or deleted. The specific data should be updated regularly, but at present is listed as a case study. However, posting at least weekly, possibly even daily a list of what pages Google is referring would be a good quality control check to make sure those pages are in good shape. Sterlingda 07:59, 16 Dec 2003 (UTC)
- While what you are saying sounds good (I for one would love to see the page updated rapidly), as a computer engineer, I'm inclined to think it is a logistical nightmare. Specifically, running such an intensive task on such a large log (whatever webserver wiki is running) would probably take a LONG time to complete - on the order of hours for each run. And last I had heard, wiki is short on developers as it is. --Raul654 08:11, 16 Dec 2003 (UTC)
- Keep and try to update it. Its a useful tool to know which articles we should try to improve. Muriel Victoria 11:29, 17 Dec 2003 (UTC)
Given the lack of updates, if anybody cares to try, I'm moving this content from Wikipedia:Ranking items by Google hit count:
In case it's useful to anyone, a perl script to rank a long list of items by "importance", defined by number of Google hits. Rather crude, but could be helpful in cases where it's a first stab at an NPOV ordering -- developed it in order to pick a few people from an alumni list for a highlight list of particularly famous alumni.
This is not the most beautiful perl script. But for what it's worth (maybe not much!) you may copy it under GPL, additional to your rights under GFDL.
#!/usr/bin/perl
# simple, but crude script to rank terms by "importance", defined by
# number of Google hits.
#
# runs on unix; takes on stdin a list of terms - one per line
#
# outputs numbers of hits as it progresses, followed by an HTML table
# of sorted hit counts at the end.
#
# requires lynx
#======================================================================
while ($term=<STDIN>)
{
$term =~ s/^\s+//;
$term =~ s/\s+$//;
$n{$term}=ngoogle("\"$term\""); # NB adds quotes (\") around item
# remove them if you don't want that.
print "$term $n{$term}\n";
}
print "=============\n";
print "<table><tr><th>Term</th><th>Google hits</th></tr>\n";
for $term (sort {$n{$b}<=>$n{$a}} keys %n) {
print "<tr><td>$term</td><td>$n{$term}</td></tr>\n"
}
print "</table>\n";
sub ngoogle
{
my $term=shift;
my $url="http://www.google.com/search?num=1&q=".escape($term);
open GOOGLE,"lynx -source '$url'|" or die $!;
my $n;
while (<GOOGLE>) {
if (m{Results <b>.*</b> of about <b>([0-9,]+)</b>}) {
($n=$1) =~ s/,//g;
last;
}
}
close GOOGLE;
return $n;
}
sub escape {
#
# turn string into "escaped" format for GET url
#
my $s = shift;
$s =~ s/([^a-zA-Z0-9_\-.])/uc sprintf("%%%02x", ord($1))/ge;
$s =~ tr/ /+/;
return $s;
}
Why does this page even need to exist?
[edit] MfD debate
This article survived a deletion debate. The discussion can be found here. -Splashtalk 21:32, 2 December 2005 (UTC)
