Looking at what ProSteel / Structures and others can do, Triforma's Form relationships are starting to look positively outdated.
Maybe structures has enough of a business case to drive a Triforma 2.0 overhaul since Bentley is supposedly the 600-pound gorilla in the structural market.
1. Walls should get an intelligent analytic workline like the structural beams have. They all have to preserve their intelligence as an analytical line and node framework for RAM/STAAD to work off the same model, BIM-like. This means that vertical elements walls/columns need to be intelligently glued to horizontal elements like beams and slabs.. and roofs.. . trusses etc. If not, you end up with rubbish when the model gets thrown over the wall to the structural engineers to be analysed. Coming as more and more elements become ISM-kosher? Hey, primary hot rolled steelwork is one of the true success stories of BIM. Good interoperability and semantics have allowed the same model to be passed on and re-used by the design, analysis, detailing and onto the fabricators before being Trimble TotalStation'd on site. This has resulted in big efficiencies through reduction of rework, concurrent design/analysis/detailing/fabrication scheduling. The way the structural members are modelled needs to help maintain the integrity of the analytical model at all times.
PS: You don't see the roadway designers extolling the use of fence stretch or romanticizing the virtues of 'no fly by wire' way of doing mods. They go parametric / associative all the time. See the new Civil Cells and Civil Geometry stuff. The workflow is based on changing key 'worklines' or alignment geometry that drive the bulk of the model downstream. Maybe TF would tap into the same rules based constraints set up they have. A road network or grid looks very much like a floor or structural framing plan if you squint.
2. Triforma's Booleans are bit basic. TFBOOELAN UNITE/SUBTRACT/INTERSECT ...etc should also have a DE-UNITE and DE-SUBTRACT tool. Most of the TF / Accudraw/Snap tools are locked out after a boolean. Users need to be able to de-boolean a feature... move/mod it using the usual tools...then re-boolean it so that your concrete FEA analysis package will have a proper solid to mesh up later.
In fact it what it really needs is an option for Feature Solids to subsume and manage the Forms that undergo a Boolean operation. What I see all the time are poorly constructed solids that look like trouble for any analytical app to pick up. Triforma suffers because it does not have a big feature/history based parametric tool library like ProSteel or the dimension driven topologically-savvy direct modeling that CreoDirect, Kubotek, Spaceclaim etc have.
3. Triforma's Booleans need to be updated to recognise sub features like the products of previous Boolean operations, and their input profiles as well as output edges and profiles.
4. TFJoins: Have a look at the work GS have done with Priority Based Connections....Not sure I agree the way everything is implemented, but it does look quite powerful.