Re^3: regex transforms of object attributes

Replies are listed 'Best First'.
Re^4: regex transforms of object attributes by Aragorn (Curate) on Jun 19, 2004 at 09:09 UTC
Who says that the `$self->content` variable stays a variable in the future? Dealing with the attribute doesn't cuase the attribute to become uncapsulated, but it does violate information hiding. An object has a published interface (the methods) which are more or less guaranteed to work. The upside is that when the author of the module/class feels there is a better way to implement things, she can do that without changes to the external interface. Of course you can use the attribute directly. No one is stopping you. It'll probably work just fine. Just don't get upset when it suddenly breaks your programs when you install a new version. Arjen	[reply] [d/l]
Re^5: regex transforms of object attributes by Roy Johnson (Monsignor) on Jun 19, 2004 at 11:24 UTC
But honestly: if it's settable and gettable, what are the chances that it's not going to be a simple variable? And a minute after posting that, I encountered this node, that answered my question one way: it's possible that a value needs to be encoded to be stored, and decoded to be retrieved. So there I have it. On the other hand, if you use a tied variable, you can do your encoding and decoding transparently and not break encapsulation... We're not really tightening our belts, it just feels that way because we're getting fatter.	[reply]
Re^6: regex transforms of object attributes by herveus (Prior) on Jun 21, 2004 at 22:21 UTC
Howdy! ...when it gets reimplemented as a value stored in a database somewhere, or stored remotely, such that you access it via SOAP or HTTP or carrier pigeon...or as a value in a hash or array somewhere. But you don't have a need to know... A procedural interface, whether a get and set pair, or a polymorphic accesser/mutator, provides a way to obtain and change the value while keeping your actions decoupled from the specific storage mechanism. yours, Michael	[reply]
Re^7: regex transforms of object attributes by Roy Johnson (Monsignor) on Jun 21, 2004 at 22:40 UTC
Re^4: regex transforms of object attributes by belden (Friar) on Jun 19, 2004 at 16:21 UTC
This will not break due to changes in $obj's interface `my $content = $obj->content(); $content =~ s/ome/thin/g; $obj->content( $content );` [download] This might `$obj->{content} =~ s/ome/thin/g;` [download] The former uses the interface published by the module author. The latter relies on the internal workings of the object itself, and therefore violates the encapsulation of the object's data. The risk that comes with the latter form (your code is now tied to the current version of Whatever::Module, and that requirement is neither enforced nor documented) just doesn't seem worth the gain of avoiding creating $content.	[reply] [d/l] [select]