I have a bunch of patches, something like modules, conforming to a kind of API (they modify the same table, one after the other). Now I'd like to add an enable/disable mechanism to all of them without having to add this functionality to each and every patch. I thought about doing a generic module host that has this feature, telling it which module to load as the first argument, something like [modulehost mirror] (mirror is the name of one of the modules, located in its own patch). [modulehost] could add all common features (like enable/disable).

Now, how should [modulehost] look like? It should load the abstraction given as its first argument, which is a patch. Just putting [$1] won't do the work since PD thinks $1 is a float. This raises the question of inlets/outlets, but I'd be happy with something like [pd~] does, say [modulehost -in 3 -out 2 mirror], or only being able to communicate with the module using [send]/[receive].

Any ideas?

@szekelyisz Hello,
1)You could add and remove objects to/from your modules with messages, but you would have to know the orders of object creation in those modules in order to connect the new objects correctly.
2)You could give your modulehost a gateway function to the table, by adding a parameter to your module names that would be included in messages that would then be sent to your modulehost, rather than directly to the table.

3. Your modulehost can use $1 for alphabetic names (i.e symbol names for the modules), but of course there must be no spaces!... and I seem to remember that they must not start with a float? (e.g. NOT "2_module").
   Try this out....... open the master first..........try_me_3.zip
   P.S. If you are on osx or Linux you will have to use [moonlib/absolutepath] instead of [getdir] to get the current directory into the "open" message....... unless you type in the full path.
   David

@whale-av Thanks a lot for the reply, I didn't know about these messages to dynamically construct patches. Your example is very clear, now I'm having trouble getting additional arguments into the dynamically loaded module. I could create a message with the desired value inside it, but I can't connect it to anything that's already in the patch, so actually I'm unable to use it. I found other interesting messages at http://puredata.info/docs/tutorials/TipsAndTricks#patch-messages, but they don't seem of any help. My problem is I can't query the ID of any pre-existing object in the patch, so I can't make a connection to any object already inside. I don't even know the ID of the newly created object (or message or whatever).

The most elegant solution would be to create objects like [r $0-arg1], [r $0-arg2], etc in the module, and send messages to these to get the arguments through, but PD doesn't return $0 of the loaded patch...

There are ways around these sorts of things, but they're not easy. Take a look at the iemguts library, which has lots of objects that are very useful for dynamic patching. (assuming you're on extended here).

[iemguts/canvasselect] can tell you the ID of an object so that you can build connections. Note that it's often easiest for you to put all objects you wish to create / connect inside a small subpatch. That way, the object numbers will be 1, 2, etc. and much easier to keep track of.

You can't dynamically create an object with $0 in it, but you can dynamically create an object with the real number equivalent of $0. So this should create the receive object you are looking for.

[bang]
|
[f $0]
|
[obj 0 0 receive $1-arg1(
|
[send pd-mypach.pd]

To back up a little, I didn't fully understand your first post, so I'm not exactly sure what you're trying to achieve here, or whether dynamic patching is necessary. What do you mean by enabling and disabling the patch? Are you familiar with the [switch~] object, which will switch off all line level objects and save CPU? Also, are you absolutely clear on what $0 does and why it differs from $1? (I hope I don't come across as patronising here--I'm just trying to make sure that we're going down the right track).

@szekelyisz Hello again,
I have been away for a week...... no sleep......no time....... sorry.
And now I will reply straight away, but I am very tired.
You can connect to objects that are already in the patch using the [connect x x x x x ( messages. BUT you need to count down through that patch in a text view to count the objects down the list and so work out the "number" of the object that you wish to connect to.
Unfortunately it depends upon the order in which you constructed the objects in that patch. As Liam says, if you had built the whole patch dynamically it is easier (you know the order of object creation for every patch created), and small patches with "sub" patches are easier to keep track of.
I might post my (up to) 64 input (up to) 64 output mixer one day. It already works well, but it is a bit hard to understand. It builds itself dynamically from a first page where you can decide how many inputs and outputs are required. If I can find the time to "notate" the patches it would provide a good "tutorial" even if nobody actually wants to use it as it is.

When I am tired I talk too much and it is all rubbish.........
Here you go.............proof.zip
David.