NAME

WebService::CaptchasDotNet - routines for captchas.net free captcha service

SYNOPSIS

# create the object
my $o = WebService::CaptchasDotNet->new(secret   => 'secret',
                                        username => 'demo');

# generate a random string for this image url
# note you _must_ use $o->random() and cannot supply
# your own random string!
my $random = $o->random;

# generate an image url
my $url = $o->url($random);

# verify that the typed captcha and image are a match
my $ok = $o->verify($user_input, $random);

DESCRIPTION

WebService::CaptchasDotNet contains several useful routines for using the free captcha service at http://captchas.net.

to use these routines you will need to visit http://captchas.net and register with them. they will provide you with a username and a shared secret key that you will need for these routines to work.

CONSTRUCTOR

new()

instantiate a new WebService::CaptchasDotNet object.

my $o = WebService::CaptchasDotNet->new(secret   => 'secret',
                                        username => 'demo',
                                        expire   => 1800);

the constructor arguments are as follows

secret

the secret key assigned to you when you registered with the captchas.net service. this argument is required.

username

the username assigned to you when you registered with the captchas.net service. this argument is required

expire

the time, in seconds, after which a cached random strings should be invalidated. see the random() documentation below for more details. this argument is optional and defaults to 3600 seconds.

if the required arguments are not given you will get still get a valid object back but verify() will always fail.

to minimize overhead in persistent environments like mod_perl you can construct a single object at the package level and hold on to it for the rest of your processing.

METHODS

verify()

this is the heart of the interface, the verification routine. basically it takes the captcha phrase that the user entered and checks whether it matches the image presented by captchas.net.

my $ok = $o->verify($user_input, $random);

here '$user_input' is what the user keyed in, and '$random' is the random string you attached to the captchas.net URL. for example, if the URL you presented on the webpage was

http://image.captchas.net?client=demo&random=RandomZufall

then the call would look like

my $ok = $o->verify($user_input, 'RandomZufall');

so basically you need to keep track of the random string yourself between calls in some stateful manner. personally, I use hidden form fields, but YMMV.

keep in mind that the verify() and random() methods are tightly linked - you must pass verify() a random string generated with the random() method and cannot just use any random random string. see the random() documentation below for the details.

verify() returns true if the user correctly identified the captcha image string and false otherwise.

random()

random() is a utility method that will generate random strings for you.

my $random = $o->random;

the random() and verify() methods are linked such that only random strings generated with random() will cause verify() to return true. here's why...

suppose a would be hacker was presented with your captcha image and recorded the image url, complete with the random string. the hacker could then use the same random string to programatically fake subsequent requests, mucking with your system. therefore the random string needs to be verified by you once and only once to make certain there is a human on the other end. for the really paranoid the random string should be received within a set time limit after it was been generated. the union of random() and verify() takes care of both of these needs.

when random() is called it stores the random string in a cache on the filesystem. verify() then checks for the existence of the file, makes sure it isn't stale, and removes it if the user input was good. only if the file exists and is recent will verify() succeed, regardless of whether the user input passes the captchas.net algorithm. if the user input was bad (such as a genuinely mis-typed response) the file will remain on the filesystem so the user can try again without completely refreshing the page. at least until the file is deemed stale.

the random() cache lives in $TMPDIR/CaptchasDotNet/ by default, where $TMPDIR is defined via File::Spec->tmpdir().

the caveat to this random() implementation is that it is filesystem based so if you are in a clustered environment with no shared mount points there is the strong possibility the box that serves the random string will not be the one to verify it later, causing legitimate matches to fail. in this case you might want to subclass WebService::CaptchasDotNet, override _init(), and choose a different path for your cache files.

url()

generate a suitable captchas.net URL for embedding within a webpage.

my $url = $o->url($random);

the returned URL will have both the passed random string and the username provided with the class constructor embedded within it. for example

my $o = WebService::CaptchasDotNet->new(secret   => 'secret',
                                        username => 'demo');

my $random = 'RandomZufall';

# http://image.captchas.net?client=demo&random=RandomZufall
my $url = $o->url($random);

it is important to note that the returned URL is encoded for proper display on a webpage, meaning the ampersand itself is encoded. this makes sure your generated pages remain valid xhtml :)

expire()

set the time (in seconds) after which a random string should expire from the cache:

$o->expire(1800);

the expire time defaults to 3600 seconds, which would give a user 60 minutes to validate themselves.

DEBUGGING

if you are interested in verbose error messages when something doesn't go according to plan you can enable debugging as follows:

use WebService::CaptchasDotNet;
WebService::CaptchasDotNet::DEBUG = 1;

SEE ALSO

http://captchas.net/

AUTHOR

Geoffrey Young <geoff@modperlcookbook.org>

COPYRIGHT

Copyright (c) 2005, Geoffrey Young All rights reserved.

This module is free software. It may be used, redistributed and/or modified under the same terms as Perl itself.