NUMERICAL variation | CORRELATED variation

I'm publishing this post to share my exploration of a very intresting Revit's topic: it's about the way the model can vary. Me and Stefano talked about this today, he suggested me to post this in order to start (hopefully) a discussion.In my work, I've noticed that esists two ways to make the model vary: a numerical one and a geometrical one.

1_NUMERICAL VARIATIONS

The first is what we can call the numerical one: it's about constraining all the entities explicitally with paramethers, but the most important thing is the fact that some paramethers are function of other paramethers ( they exist in a form such as y=f(x) ).

This image is from my very first model, the component. As you can see, my paramethers' table 

1. is full of paramethers: 25 in total
2. contains both black p. and grey ones: the difference is that  Y = f ( X )
3. most of the dependent paramethers are related to the independent ones by trigonometrical formula, which is obvious, since it exists a geometrical relation among them and, above all, among the purpose which has led me defining them.

I had modeled a frame for the IAZZO project. It was for that my intention to sort out from this model a very simple schedule: containing the phisical dimensions of my beams only , and their total. In my intention , I could have given this schedule to a wood supplier with all the informations necessaries to make an order.

But unfortunately, I wasn't able to sort out this shedule, because Revit recognises these paramether not like a phisical feature of a 3d modeled object, but as a NUMBER-GENERATOR of a geometry. It does not matter if my (real) beams are 6 in the 3 frames: since Revit uses the GENERATOR paramether only once a instance, he calculates the paramethers 3 times only. But I have to order more than 3 beams of each kind! 

It's not the model to be unuseful, but the way in which it is thought. I need to calculate the features of beams, i don't need to create a geometry made of numbers and trigonometrical formula,because for me the geometry is the generator, not the numbers; because our subject is architecture, concerning real stuff.

2_CORRELATED VARIATIONS

For this I've discovered another way to make the model vary, it's about constrining them to a reference geometry which varies itself depending on the values of what are the independent paramethers in my previsious model only. All my other models are built up in this way.

Concernign the iazzo, it is necessary to create the beam family, whose shared paramethers are the ones I want to calculate in a shedule. I've at first tried to nest this family into a frame one: but this attempt is unsuccesful, because i create a reference geometry (made of planes and lines) and i assinged its paramethers to the beam features, to create the model: but that's the same mistake!! I've only GENERATOR-NUMBERS.

The way to sort it out is to create a reference mass directly into the project. This object is poverful, beacuse it can vary parametrically. You can put ref. planes and lines in the mass, and constraint the elements with features to be shared using the ARROWS, rather than using a paramether.

The most important thing in this model is that I HAVE FOR PARAMETHERS THE independent ONES ONLY, WHILE THE OTHERS DO NOT EXIST NUMERICALLY: THEY EXISTS IN THE MODEL BECAUSE OF THE REFERENCE GEOMETRY CONTRUCTION. ALL THE VARIATIONS ARE INDEED CORRELATED. In the model indeed a lot of implicit relations exist , such as a rectangle is always a rectangle. To take advantage of this, it is necessary do design a little bit carefully (respecting -oral or written?- rules such as a i-reference line's centre of rotation is fixed only if it is the intersection of two reference planes, or two reference lines ii-an offset is preserved in a rotation only if it is constrained with a quote iii- snaps are sacral iv-do not always lock).

According to me,  using the correlated variation in a geometry is,  much a more powerful way of modeling, because for us -future- architects the numbers are son of the model, not viceversa (with the exeption of those number which genetares the geomtry...which exists in the mass).   

Only using this method, i've been able to schedule the beams for my frame. A schedule which is alive and variable as the model is.