Well, ask and you shall receive. Kind reader K sent us his homebrew GSS generator written in PHP. It’s a fascinating piece of code using high-level algorithms and math to prove that the post length and number of uppercase characters can be used to measure the total insanity of any eBay post. Case in point:
Auction 1 – PLEASE HELP!WHAT’S GROWING ON MY HEAD!! MYSTERY AUCTION GSS Score: 66,731
Auction 2 – EXTRA LARGE HOBO BIG GREEN PURSE BAG RING NEW GSS Score: 1,273
As we see, even if the poster tends to use capital letters, the GSS generator is not fooled.
The GSS Generator is available after the jump for further study. I suspect it might be slightly insecure and therefore would like a few pairs of eyes to ensure that we can indeed host this valuable tool on a server someday, thereby allowing millions of men and women to separate the eBay wheat from the eBay insane people.
Related
eBay Deal of the Day: Vampire Killing Kit
http://www.gizmodo.com/gadgets/gadgets/index.php#ebay-deal-of-the-day-vampire-killing-kit-114284
RUN AT YOUR OWN RISK – I’m not going to go through these to make sure they don’t “rm -rf” your boot dir.
“;
$end_string = ““;
//Find Begin and end auction crap
$beg_pos = strpos($buffer,$begin_string);
$end_pos = strpos($buffer,$end_string,$beg_pos);
//grab just stuff between begin and end auction comments
$desc_text = substr($buffer,($beg_pos + strlen($begin_string)),($end_pos – $beg_pos));
//Strip out the tags then strip out space
$strip_text = strip_tags($desc_text);
$strip_text = str_replace(” “,””,$strip_text);
//Strip out line breaks and returns (proabbly don’t need to do this)
$strip_text = preg_replace(“/(rn|n|r)/”, “”, $strip_text);
//find lenght of stripped text
$len_strip_text = strlen($strip_text);
//Counter vars
$num_lower = 0;
$num_upper = 0;
$num_other = 0;
//loop through stripped text count upper, lower and other chars
for($x = 0; $x < $len_strip_text; $x++) { $test_char = substr($strip_text,$x,1); $ascii_num = ord($test_char); if( ($ascii_num >= 97) && ($ascii_num <= 122) ){ $num_lower++; } else { if( ($ascii_num >= 65) && ($ascii_num <= 90) ) { $num_upper++; } else { $num_other++; } } } //Calculate Gizmodo Strangeness Score $gss = ($len_strip_text - $num_other) * ($num_upper / $num_lower); echo "URL = $url_to_calc“;
?>
| Summary Data: | |
| Total Chars | |
| Number uppercase | |
| Number lowercase | |
| Number other chars | |
| Gizmodo Strangeness Score* | |
*(Number Chars – Number Other Chars) x (Number Upper / Number Lower)
Stripped Text (blank space removed):
“;
echo $strip_text;
} else {
echo “ERROR fetching file”;
}
}
?>
Here is an interesting command-line version in Perl by Thomas:
#!/usr/bin/perl
use LWP::Simple;
{
if (@ARGV[0] eq “”)
{
print “Usage: weirdness.pl
exit;
}
$url = @ARGV[0];
$buffer = get “$url”;
$buffer =~ s/<(?:[^>‘”]*|([‘”]).*?1)*>//gs;
$length = length($buffer);
@uppers = ($buffer =~ /([A-Z])/g);
@lowers = ($buffer =~ /([a-z])/g);
$num_upper = @uppers;
$num_lower = @lowers;
print “Non-html chars: $lengthn”;
print “upper chars: $num_uppern”;
print “lower chars: $num_lowernn”;
#$score = $num_upper/$num_lower;
$ratio = $num_upper/$num_lower;
$score = $ratio * $length;
printf(“upper-to-lower ratio: %.2fn”, $ratio);
printf(“Weirdness score: %.2fn”, $score);
