Here my opinion on the situation. There is no license or law to guide or steer us, but past experiences can help decide which opinion leads to the best solution.
On 2015-12-25 08:28 PM, Alexandre Torres Porres wrote:
2015-12-23 18:36 GMT-02:00 katja <katjavetter@gmail.com mailto:katjavetter@gmail.com>:
Summarizing, the discussion in this thread has so far rendered three practical and simple solutions to improve MaxMSP compatibility in Cyclone without breaking Pd patches (with average~ as an example):
- MaxMSP compatibility through an extra inlet / outlet
I still think that introducing an extra outlet is the least complicated and least intrusive.
A signal outlet, at the right of a message outlet, is not very common for Pd. And it leads to an object doing two things. Because of POLA*, added complexity and work, it would not be my first choice. But not many objects are expected to have a fix like this, so just for once...
- MaxMSP compatibility available through an extra operational mode
Not sure how an "extra operational mode" would work, but seems a little complicated.
It would mean using an argument or message to switch behaviour. As average~ already has two (optional) arguments, which become mandatory just to specify a third, I do not see it as a reasonable option.
- MaxMSP compatibility available through an extra class
An extra class breaks compatibility, as you need another class name - seems like a drastic or last resource solution.
IMHO, this fits the situation best, as several objects have different names in Pd and Max/MSP. Apart from the extra object name.
- MaxMSP compatibility available with a -legacy startup flag
The -legacy startup flag would mean you can have only one of the two solutions per Pd-instance. Introducing this flag (in Vanilla?) just for this object would be a bit of overkill.
So I will try to combine the average2~ functionality into average~.
cheers and merry xmas
Greetings & happy 2016,
Fred Jan
*) https://en.wikipedia.org/wiki/Principle_of_least_astonishment