each class should focus on a single area of responsibility. therefore the login check doesn't belong in class A or B, instead it belongs in the code consuming the functionality of said classes.
<?php $login = new login($pwd, $acc, $customerAcc); $bee = new b(); if ($login->validate()) $bee->ExtractRawData(); ?> you could alternatively look into either extending the login class or using a decorator pattern (is that the right name?), or even registry pattern, e.g.: <? class A { function extract() { if (MyRegistry::get('login')->validate()) $this->processExtract(); } private function processExtract() {] } class MyRegistry { private $objects = array(); function get($cls) { if (!isset(self::$objects[$cls]) || !self::$objects[$cls] instanceof $cls) throw new Exception("$cls not registered!"); } function set($obj) { if (!is_object($obj)) throw new Exception("gimme an object!"); self::$objects[ get_class($obj) ] = $obj; } } MyRegistry::set(new login($pwd, $acc, $customerAcc)); $a = new A; $a->extract(); ?> Andrew Williams schreef: > Hi, > > I am working on a back end OOP project with about about 18 classes of Object > but some of the Objects shear one or two functions in common, like external > login system . What is the best way to avoid a repeatable function among the > classes without using global object instance. > > class a{ > > function ExtractRawData() > { > global loginObj; > *if($loginObj;->GetLoginSession(){* > ///process > } > } > > } > > class b{ > > function JohnXchange() > { > global loginObj; > *if($loginObj;->GetLoginSession(){* > ///process > } > } > } > > class login( > > function loginSession($pwd, $acc, $customerAcc) > { > $this ->validate = connect to externalServerObject($pwd, $acc, > $customerAcc); > } > *GetLoginSession()* > { > return $this ->validate; > } > } > > > > Andrew Williams > http//www.willandy.co.uk > > > > > -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php