Less combined objects and better ways to connect them

[Christiaan]

This is more of a high level discussion than a specific wish. The more I model buildings in VW and the more I understand VW, the more I want elements of the model to be separate objects, instead of having such elements combined into one tool. But with better, more intelligent ways to connect them all.

 

The general principle should be that if it's installed in real life as a separate element then it should be a separate element in VW too.

 

Combining objects into one tool simply because they're attached to each can create limitations that don't exist in the real world; and those limitations degrade each element's ability to change direction and interact with other objects. Then you end up with complicated schemes like automatic wall/slab/roof connections that just aren't reliable.

 

The stair + railing is the prime example:

https://forum.vectorworks.net/index.php?/topic/90711-stairs-and-railings-should-be-separate-tools/

 

The same argument can be made for the three main elements of a building envelope (internal finishes, external finishes and structure of floor, walls and roof). But the wall tool is the one most degraded by the current paradigm. For instance we end up making hundreds of different wall styles to break the wall down into sections horizontally where the components change slightly (e.g. slightly different internal lining, a different cavity, passing a column, etc), when in fact in the real world the wall is installed as three main elements: finishes to one side, structural, finishes to the other side.

https://forum.vectorworks.net/index.php?/topic/79110-workflow-to-model-structure-internal-elements-and-external-elements-separately/

 

Windows, sills, lintels cavity closers (and cavity barriers) is another one. Why are the sill and lintel objects combined with the window object? Perhaps the window tool should have it's sub-sill settings, but the main sill should be a separate object. Again, falling back on the principle of how they're actually installed.

https://forum.vectorworks.net/index.php?/topic/49479-window-sill-improvements/

https://forum.vectorworks.net/index.php?/topic/49481-brick-header-and-decorative-window-lintels/

https://forum.vectorworks.net/index.php?/topic/98035-support-for-modelling-cavity-closers/

https://forum.vectorworks.net/index.php?/topic/98080-new-tool-for-modelling-cavity-barriers/

 

 

Edited July 4, 2022 by Christiaan

[line-weight]

1 hour ago, Christiaan said:

This is more of a high level discussion than a specific wish. The more I model buildings in VW and the more I understand VW, the more I want elements of the model to be separate objects, instead of having such elements combined into one tool. But with better, more intelligent ways to connect them all.

 

The general principle should be that if it's installed in real life as a separate element then it should be a separate element in VW too.

 

Combining objects into one tool simply because they're attached to each can create limitations that don't exist in the real world; and those limitations degrade each element's ability to change direction and interact with other objects. Then you end up with complicated schemes like automatic wall/slab/roof connections that just aren't reliable.

 

The stair + railing is the prime example:

https://forum.vectorworks.net/index.php?/topic/90711-stairs-and-railings-should-be-separate-tools/

 

The same argument can be made for the three main elements of a building envelope (floor, walls, roof). But the wall tool is the one most degraded by the current paradigm. For instance we end up making hundreds of different wall styles to break the wall down into sections horizontally where the components change slightly (e.g. slightly different internal lining, a different cavity, passing a column, etc), when in fact in the real world the wall is installed as three main elements: finishes to one side, structural, finishes to the other side.

https://forum.vectorworks.net/index.php?/topic/79110-workflow-to-model-structure-internal-elements-and-external-elements-separately/

 

Windows, sills, lintels cavity closers (and cavity barriers) is another one. Why are the sill and lintel objects combined with the window object? Perhaps the window tool should have it's sub-sill settings, but the main sill should be a separate object. Again, falling back on the principle of how they're actually installed.

https://forum.vectorworks.net/index.php?/topic/49479-window-sill-improvements/

https://forum.vectorworks.net/index.php?/topic/49481-brick-header-and-decorative-window-lintels/

https://forum.vectorworks.net/index.php?/topic/98035-support-for-modelling-cavity-closers/

https://forum.vectorworks.net/index.php?/topic/98080-new-tool-for-modelling-cavity-barriers/

 

 

Agree with all of this entirely.

[E|FA]

I agree.

[blanger]

Please remember that architecture is not the only application. For set design purposes I often wish for more integration.  For instance the ability to include base mould as a wall component would be a huge plus for Set Design. I'd say we work primarily with wall finishes and the flat itself is the simplest element. @Andy Broomell do you have any opinions on this?

[Matt Overton]

4 hours ago, blanger said:

Please remember that architecture is not the only application. For set design purposes I often wish for more integration.  For instance the ability to include base mould as a wall component would be a huge plus for Set Design. I'd say we work primarily with wall finishes and the flat itself is the simplest element. @Andy Broomell do you have any opinions on this?

If I understand correctly the wish is for 2 parts

A) a better system to combine a group of elements together and have that group react to each other if further edits are required. 

So once you've created the the flat and trimmed it with a baseboard you could combine them to a symbol-like "thing" or a super-group. 

The Object info would expose some controls from the combined object (like styled objects do) to allow selective changes to the object.

ie a Flat with Baseboard could be resized and it would all grow in the way you've told it to grow. 

 

B) on the back of new system complex plug-in objects break up in to smaller easy to maintain object, allow 2 complex objects to use the same sub-object so for example a Stair-object and theoretical Balcony-object would both internally use the same balustrade object to the point of exploding either would lead to a group of the sub-objects. 

 

 

[blanger]

I think I'm following now.  I had this situation which was pretty annoying and could be solved by a revised approach.  I had two walls that required different thickness both with the same finish. I needed 2 wall types, one for each thickness and then if I revised the finish I had to update it in both walls.

 

[Tom W.]

8 hours ago, Matt Overton said:

If I understand correctly the wish is for 2 parts

A) a better system to combine a group of elements together and have that group react to each other if further edits are required. 

So once you've created the the flat and trimmed it with a baseboard you could combine them to a symbol-like "thing" or a super-group. 

The Object info would expose some controls from the combined object (like styled objects do) to allow selective changes to the object.

ie a Flat with Baseboard could be resized and it would all grow in the way you've told it to grow. 

 

B) on the back of new system complex plug-in objects break up in to smaller easy to maintain object, allow 2 complex objects to use the same sub-object so for example a Stair-object and theoretical Balcony-object would both internally use the same balustrade object to the point of exploding either would lead to a group of the sub-objects. 

 

 

 

Excellently put

[Christiaan]

4 hours ago, blanger said:

I think I'm following now.  I had this situation which was pretty annoying and could be solved by a revised approach.  I had two walls that required different thickness both with the same finish. I needed 2 wall types, one for each thickness and then if I revised the finish I had to update it in both walls.

 

Now imagine you have hundreds of wall types with the same problem! 

[zoomer]

On 7/4/2022 at 11:00 AM, Christiaan said:

Why are the sill and lintel objects combined with the window object?

 

 

I think because e.g. this way the Window size will determine the Sill length

automatically, which I think makes sense.

 

But as I agree completely with your proposal in general.

So the "Sill Tool" would need to be "inserted" into a Window

(like a Window into a Wall) and read the dimensions from the Window (?)

 

I am a big fan of Structural separation from Finishes.

Of course For Walls and Slabs but even also for Stairs.

(As for now the Stair isn't even able to solve the geometry of the Structural Part)

So Stair construction + Finishes.

[Christiaan]

9 hours ago, Matt Overton said:

B) on the back of new system complex plug-in objects break up in to smaller easy to maintain object, allow 2 complex objects to use the same sub-object so for example a Stair-object and theoretical Balcony-object would both internally use the same balustrade object to the point of exploding either would lead to a group of the sub-objects. 

Perhaps. But I'm thinking more that you would have a stair tool with no railing settings and a separate railing tool (that could be used anywhere, including stairs, balconies, landings, etc.). And the railing tool would understand what a stair object is and be able to snap and connect to a stair intelligently, with few adjustments needed, and move with the stair if it is edited.

Edited July 6, 2022 by Christiaan

[Christiaan]

4 minutes ago, zoomer said:

I think because e.g. this way the Window size will determine the Sill length automatically, which I think makes sense.

 

So the "Sill Tool" would need to be "inserted" into a Window (like a Window into a Wall) and read the dimensions from the Window (?)

I hadn't thought about that, but, yes, that's a good example of more intelligent connections while being a separate object.

Edited July 6, 2022 by Christiaan

[Tom W.]

2 minutes ago, Christiaan said:

And the railing tool would understand what a stair object is and be able to snap and connect to a stair intelligently, with few adjustments needed.

Good luck with that... 🙂

[Christiaan]

6 minutes ago, Tom W. said:

Good luck with that... 🙂

The technology for this already exists in VW. Associative dimensions, wall connections, roof-wall connections, to name a few.

 

And ArchiCAD already works like this I believe.

Edited July 6, 2022 by Christiaan

[Tom W.]

I agree + am not questioning the suggestion for one second. I am just a bit underwhelmed by the current capability of the Railing/Fence Tool in 3D. Very buggy in my experience. It is a real pain creating runs of fencing that follow the terrain of a site model. Should be very easy

[Christiaan]

Oh yes, the current railing tool is a bit of a dog. But I have great confidence in the engineering team VW now has onboard. They just need to be let loose on it.

[twk]

Agree with all of your comments totally @Christiaan.

There is quite a common trend in NZ where you have different claddings above, beside and below an opening. We have to split the wall into parts to cater for the change in cladding even though the substrate is the same (timber framed etc)

 

 

[line-weight]

I think it's inevitable that VW (and any BIM package) is going to have to get on board with the concept that walls (and roofs and floors) need to have a properly defined structural element, an internal finish element, and an external finish element, (and I'd say possibly also a thermal insulation element) each of which can be dealt with independently to some extent.

 

Things have gone about as far as they can, treating walls as unvarying assemblies from one side to the other- the problems are revealed as soon as you get into things like wall closures, and workarounds will only go so far. Even quite simple things like junctions between walls in plan don't really work properly at the moment, certainly not at any level of detail. Vectorworks needs to understand the structural core as the thing that is all connected together, and then inner and outer layers are stuck onto that as a separate operation.

[Jeff Prince]

This is an interesting discussion.  It seems you are all wishing for virtual construction that would be similar to what manufactures use to define things like cars, jet engines, and other complex assemblies.  Reminds me of Frank Gehry’s collaboration with Dassault Systems back in the day.

I believe “virtual construction” is a likely the future for a segment of AEC on complex building, prefab, and now 3D printed buildings.  That’s kind of how I describe BIM processes to my students with our current technologies within Vectorworks.  I’m not a programmer, but I imagine the implementation of such processes would require a completely new “Vectorworks” program rather than the continual band aid approach to expanding current tools and the breakage that is ever present.

 

FWIW, I did my first degree in manufacturing engineering and it was that skill set that landed me my first job in the AEC industry, training teams on new processes and the use of the Softdesk suite to define complex medical, athletic, and infrastructure facilities.  This was replaced by Autodesk’s discipline specific “desktop” packages and eventually Revit.  In the AutoDesk example, Revit will eventually be replaced with the emerging tech in that company.  AEC as an industry trails manufacturing by 20 years or more in terms of modeling and documentation technology, IMHO.

 

Going down that path… just be very careful what you wish for.  I think it will require the current generation of AEC leaders to retire before the next massive push forward occurs.  Then it will take another 10 years or so before AEC processes catch up due to software and staff skill development.  Think of the hand drafting to CAD to BIM transitions that have been taking place over the past decades.  These are entirely different skill sets and processes that will require a new generation of technology and people to use it.  It’s already started, but it takes time before widespread adoption.  Then there is the whole construction side of things that has to be pulled into the discussion along with manufactures of systems that are installed into buildings.  It will be interesting to see how the larger picture comes into focus on these topics.

 

We humans are creatures of habit and innovation in AEC is relatively slow compared to other industries.  I’m just glad I’m a landscape architect who gets to draw by hand and design in the field on occasion still.  I think we are a long ways off from 3D printed trees and shrubs 🙂 

[Christiaan]

Jeff, no I think that's overstating the problem really. I don't see the output necessarily being any different than it is now. This is about refining the process of modelling in VW. Essentially not grouping objects into one tool unnecessarily. And if they're not in the same tool, what objects should be able to associate with.

 

I don't think we need to wait for anybody to retire for that to happen. I think VW may be on this path already. Or at least I know that they're discussing things such as 'should the sill actually be part of the window or part of the wall or something else?'

 

 

Edited July 6, 2022 by Christiaan

[line-weight]

It strikes me that round about now might be quite a good moment for someone to build something new from the ground up. There are probably enough people now semi-seriously modelling in 3d (and using those models to actually produce at least some of the working drawings) that there's a decent market, along with enough people coming to an understanding of what does and doesn't work (based on trying to make existing software, such as VW, do what they actually want it to do).

 

I think that's quite a big change compared to 10 years ago, say.

 

I don't know enough about other packages like Revit, to know if they have similar problems to those that VW does.

 

But it's surely difficult for any of these programmes, which have to keep happy a large userbase that still wants to use things in a variety of "legacy" workflows, to change quickly.

 

Progress in VW is painfully, painfully slow, on the evidence of the past few years and I don't see much reason to hope that this is going to change.

[Christiaan]

I would love to see an Onshape for architecture, in web browser only. But it's a chicken and egg thing. To write something from scratch you'd have to produce something fairly comprehensive out of the box in order to compete (the built environment is really complex). Otherwise people are just going to keep using the tools that earn them the fees. 

 

My money is on VW etc staying ahead of any such startup because it gets the job done. While VW (and others) have a lot of baggage to deal with in their code base, they also have a huge code base!

 

You could argue that a CAD developer like VW should write something from scratch, maybe even use some of that code base. But how do they do that without development of VW itself being negatively effected?

Edited July 6, 2022 by Christiaan

[zoomer]

1 hour ago, line-weight said:

Progress in VW is painfully, painfully slow, on the evidence of the past few years and I don't see much reason to hope that this is going to change.

 

 

For me it was from VW 2014 to VW 2021.

Because there were of course some new, important and helpful features over the years,

but some basic annoyances where never addressed.

 

That changed for me with the release of VW 2022 drastically.

(Screen Plane deprecated, Walls now Solids, .....)

 

Unfortunately these larger infringement came a bit on cost of VW's stability and reliability.

(disappearing elements, location/coordinates issues, ...)

So I hope for VW 2023 to just consolidate a bit and do not so much new features and changes.

 

Since a few years I feel like VW SP0 is no often less problematic than SP1-3, same for

macOS Public Betas vs official.

(So often I had Public Betas that worked well, so I resigned from PB, just to realize that

I get the same old issues back in official version - and went back to Public Beta Program.

Like Sleep issues, not exited external drives correctly, Mail issues, Safari issues, ....)

 

[zoomer]

3 hours ago, line-weight said:

I don't know enough about other packages like Revit, to know if they have similar problems to those that VW does.

 

I haven't used it ever but I think it feels a bit bureaucratic, complicated and not very user friendly.

At least for the target group of Vectorworks.

 

Similar like Autodesk is a whole different world.

People like engineers or programming affin users love the command line input and LISP.

But for me it is too abstract and disrupting from creativity.

 

 

 

Back to this Topic.

Bricscad's way in this regard looks far more future proof.

You do just (excellent) direct modeling with Solids, so you can create anything and decide

later how to IFC Tag them (or let its AI do for you). And later refine details like how Wall Components

interact with Slab Components at one place and let Propagate AI search and adapt similar situations.

 

Of course Bricscad's "PIO's" are currently much simpler.

Like Windows have a third of VW Settings but produce a similar result. But Like in VW, if you want something

deviating from available Settings and need to insert custom Symbols in Walls you will need to create your

own custom parametric constrained Window Part.

(which you can as complex if you want if you are able to set it up)

Or the Stair "Tool" will give you just basic Threads and cut an opening in the floor, but that is far away from

a "Stair". And opposite to VW it is not that much fun to place Windows or Doors in Walls.

And so far Bricscad also has no answer to separate Structural vs Finishes like reading the Window openings

from Walls to adapt a separate outer Insulation/Finish Layer to follow changes.

 

So as long as your design fits VW PIOs, VW is the easiest, non disturbing and fastest BIM App.

But you will run so easily into limitations which also disrupts creativity.

 

Edited July 6, 2022 by zoomer