Geometry creation and modification

simple (though geometrically complex) .... you said it.
And it will be tricky to model in ABD. And not so easy for a carpenter as well.

Cut Solid by Curves, modify and connect form tools will do what you show. It will be easier if you allow boolean tools also.

But I agree- ACS and Accudraw RE is tricky understanding when to move the compass and not. And the real expert on those is Brien Bastings - and he doesn't touch TF or ABD. You have to ask in MicroStation forum to reach him.

Demotion yes - but exactly what do you want to achieve? In what stage of the project?
Drawings? It will do.
Quantification - much harder.
Illustrations for client?

Easy modification like Implied Relations - no, not yet.

3d-modeling is nice - but it also forces you to resolve all joints and meetings everywhere in a building - 2d is very economical.

Thanks Thomas - I think you know ABD like no other (when did you start?)

Unknown said:

simple (though geometrically complex) .... you said it

but is typical of what Architects did, until competitive need to adopt the limitations of early CAD ‘temporarily’ suspended that freedom, which has never since been reinstated by the 3D Building modelers – so far.

Unknown said:

it will be tricky to model in ABD

yes, in today's ABD - but I for one am looking to the coming year, when ABD's geometry creation and modification gets major re-think by Bentley. The signs are good - both in 64bit MS's Platform capabilities, and in hints and responses from the ABD team. I hope we users can do a lot of well considered work to make the case heard at Bentley, for the changes we need. I don't think we can just leave this to Bentley, unaided. Does anyone agree?

Unknown said:

And not so easy for a carpenter as well

'not so brainlessly routine', I’d agree - but carpenters regularly handle multi-axis 3D cuts with precision. I find they love to do something interesting for a change – no problems there. It does require me to provide precision dimensions e.g. height above datum of wallplate/purlin tops at key points - then the rafters span between these and shapes and spaces form themselves. I can do this in 2D, with much manual trigonometry. It's ridiculous that a 3D Building modeler can't even create such geometry.

Unknown said:

Cut Solid by Curves, modify and connect form tools will do what you show. It will be easier if you allow boolean tools also.

Those Forms at the apex (and most throughout) are full of MS Cut Solid by Curves and Boolean operations

to fill in for the inadequacies of ABD's modify and connect form tools. ABD tools as first preference, MS ones where ABD tools refuse or send Forms shooting off to some remote intersection etc.

Yes I can work around all these tools, but with so much failing - I lie awake figuring out the approach I'll try tomorrow. Always I retain the Forms as Forms, not using those MS tools that result in demotion to SmartSolid, because Forms' behaviour is always more suitable for further modification. But after multiple MS operations, although still reporting as a Form, ABD operations cease to work on it.

Believe me, after months of trying everything, it's true that "Cut Solid by Curves, modify and connect form tools [boolean tools also] will do what you show" - but at completely unacceptable cost in abortive work, frustration and time, and way beyond the programmers' concepts and intentions.

Anyone got any more ideas?

Thomas, if important, could you say more about how ACS and RE come into it, what I should be asking Brien Bastings about, and also Demotion?

Unknown said:

what do you want to achieve? In what stage of the project?

Drawings? It will do.

Quantification - much harder.

Illustrations for client?

First, just to model, then Drawings, Illustrations, but not Quantification.

Unknown said:

Easy modification like Implied Relations

I don't need (or trust) Implied Relations – prefer to extend, trim, re-connect as I go, maybe with ability to re-apply 64bit MS Feature cuts.

Unknown said:

3d-modeling is nice - but it also forces you to resolve all joints and meetings everywhere in a building

Yes, of the basic layers, to the ABD Connect (butt or mitre) level of detail, as in my dgns above. That's no problem if only extend, trim, connect would work on all Form edges equally, not exclusively on Form Ends. That’s all it would take (for a start – there’s more, naturally).

Unknown said:

2d is very economical

yes sometimes the way to go, but presently necessitated only by ABD’s inability to model full constructional detail - not everywhere but just in those areas where detail SV cut-planes and peel-back 3D pdfs are planned. Again no problem if only extend, trim, connect would work as suggested above.

Today’s ABD is so far from adequate, but so close to being magnificent, over the coming year. We hope!

But after multiple MS operations, although still reporting as a Form, ABD operations cease to work on it.

Yes - I see that also.

Tom

I use compound wall for everything. Part of what I had to do was determine the amount of detail that I wanted to perform.
I want my building to create the section of 1/2" scale or larger AND
The plan for 1/8" scale or larger

Here are the issues I have and how to fix them.
The first is CENTERLINE. With the CENTERLINE I can create many materials that would require 2d drafting in the Drawing Model.
For examples. Metal Panels, Shingles, T&G planks, CMU Walls, etc

First CENTERLINES must be fixed to allow for full width of centerline cells.
Metal Panels - I have a linestyle for a 4", 6" and 8" Pan. I use it as a centerline for a form. The color for the outline of that form is the background color.
Shingles - I do the same as the metal panel but to do this right I must have one for EVERY slope. So on a hip roof I have 4 patterns. This is where CENTERLINE needs orientation toward base.
T&G - fairly easy
CMU walls - very Tough as a Centerline recognizes a line in relation to the origin centerline.cell however the end of the centerline does not recognize the end centerline.cell. That needs to be addressed.

I also create a 2" blocking Centerline with an "X" - So for that detail I have to eigher make the decision to break blocking down into 2" stips of non-unifying material or wait for the CENTERLINE to be fixed and provide a single centerline.

So currently I can create most sections fully and the only 2d addition is CMU, Metal Pans, etc.

NOW

Then something like the roof. I do as you are doing but KNOW that roof (sloped compound walls) Only join with other like walls and you have to struggle with the orientation. Part is a FORM or WALL and an ASSEMBLY is a Compound WALL or Compound Form
THE FIX. - Even know these are walls they are still FORMs.
SO extend any side to another any side should work. BUT IS DOESN't SO----
The fix is to allow form tools Side tools work too on any wall. The WALLS only as WALL TOOLS is a handicap and force you into a very limited boxy design.

SO Fixes to take sections a step toward completion.
Fix Centerlines
Allow ANY FORM Side Manipulation on compound Assemblies

Would we agree that no.1 on wishlist is Connect tools that work on all Form faces equally?

ABD's cuboid concept assumes that the only faces that need a L- T- or X-Connect, are Ends - so only pairs of Ends can trim/extend to each other.
'Trim/extend to each other' is the key thing, the key to a competent Building modeler - but it's cripplingly limited in ABD, allowing End-to-End only. That does work beautifully; it even allows lap-priority.
Connect Form is the prime ABD modification tool, the fastest, simplest and most reliable.

The second-choice ABD modification tool is Modify Form, lacking any 'Trim/extend to each other' ability. In Modify Form, only Top and Bottom faces (not End faces) can Trim/Extend directly to another Form, but without affecting the other Form - exact same effect as the T-option in Connect gives on Ends.
Modify Form is slow, tricky, internally inconsistent and often unpredictable in result.

Just imagine - if all Form faces - Top Bottom and Ends - would work equally, then L- T- or X-Connect and Modify>To Form would merge together into one easy Connect tool.
Optional multi-select as well as single-select would allow optional multi-facet as well as single-facet result.

Just that, would turn ABD into a geometrically competent Building modeler, free of its dumb cuboid limitation. For a start - of course there's more.

Side faces alone would disallow multi-facet result: a Form is hardly a Form if not parallel in thickness.

Well said Tom.
Been saying this for years as things like roofs and anything that is not a cube causes slow creation.

Once Connection tool for any type of form.

Though you'd like it Eric

To be clear, Connect tools that work on all Form faces equally means not just Top to Top and Bottom to Bottom, but all combinations equally - Top to End, Bottom to Top etc as well. An end to trying to keep track of which is Top, Bottom, End etc.
For the great bulk of routine work, which really is strictly cuboid and ABD handles so productively, nothing need change.

If that's all that Bentley's mainstream big corporate Architect customers demand of ABD - utilitarian automation
of routine cuboid commercial building production, and
of the boring routine bits of their more ambitious buildings,
then that's because those corporates have other Building modeler software to do the more imaginative bits that make a 'signature' building. That superior Building modeler software is their highly expensive investment, and its capabilities are guarded as their competitive advantage.

Cos like Fosters have long had their own sophisticated in-house mdl aps, which can do sophisticated geometry (BricsWork (Belgium, 1993) started as one such mdl ap, which Bentley bought and re-named Triforma, but never developed beyond its cuboid capability). Others have paid megabucks for Gehry Technology's Catia based Digital Project, and now Dassault are directly introducing Catia Building modules. Jon Hirschtick’s Belmont may be following close behind. Rhino and Bentley GC are in the mix too. These softwares specialise in NC-manufactured curviness that only the big boys have the key to.

The result of this software stratification is our cities today – a zoo of all the limited tricks that Revit can do. If it can’t be done quickly and profitably in Revit, it won’t happen.
Blocky buildings compete by façade variation, with stuck-on features and eye-catching roofs.

Low grade commercial buildings have always been like that, but until the 80s very many quality Architects loved to find new geometric/spacial solutions in response to social, technical and townscape needs. I feel so sorry for new Architects, with nothing between Revit façade-wallpapering on one the hand, and for the elite, Rhino exhibitionism on the other hand.

There is no 3D CAD-supported middle ground, for good provincial Architects to return to designing radical buildings,
creating new freer geometric/spacial solutions in response to social, technical and townscape needs, and
detailing them for buildability by ordinary builders,
out of common materials.

ABD does not support that – it’s typical of what has been lost to commoditisation, ever since the Reagan/Thatcher years.

The answer is as simple as ABD Connect tools that work on all Form faces equally - that's all it would take to turn ABD into a geometrically competent Building modeler, free of its dumb cuboid limitation. For a start - of course there's more.

Though you'd like it Eric

To be clear, Connect tools that work on all Form faces equally means not just Top to Top and Bottom to Bottom, but all combinations equally - Top to End, Bottom to Top etc as well. An end to trying to keep track of which is Top, Bottom, End etc.
For the great bulk of routine work, which really is strictly cuboid and ABD handles so productively, nothing need change.

If that's all that Bentley's mainstream big corporate Architect customers demand of ABD - utilitarian automation
of routine cuboid commercial building production, and
of the boring routine bits of their more ambitious buildings,
then that's because those corporates have other Building modeler software to do the more imaginative bits that make a 'signature' building. That superior Building modeler software is their highly expensive investment, and its capabilities are guarded as their competitive advantage.

Cos like Fosters have long had their own sophisticated in-house mdl aps, which can do sophisticated geometry (BricsWork (Belgium, 1993) started as one such mdl ap, which Bentley bought and re-named Triforma, but never developed beyond its cuboid capability). Others have paid megabucks for Gehry Technology's Catia based Digital Project, and now Dassault are directly introducing Catia Building modules. Jon Hirschtick’s Belmont may be following close behind. Rhino and Bentley GC are in the mix too. These softwares specialise in NC-manufactured curviness that only the big boys have the key to.

The result of this software stratification is our cities today – a zoo of all the limited tricks that Revit can do. If it can’t be done quickly and profitably in Revit, it won’t happen.
Blocky buildings compete by façade variation, with stuck-on features and eye-catching roofs.

Low grade commercial buildings have always been like that, but until the 80s very many quality Architects loved to find new geometric/spacial solutions in response to social, technical and townscape needs. I feel so sorry for new Architects, with nothing between Revit façade-wallpapering on one the hand, and for the elite, Rhino exhibitionism on the other hand.

There is no 3D CAD-supported middle ground, for good provincial Architects to return to designing radical buildings,
creating new freer geometric/spacial solutions in response to social, technical and townscape needs, and
detailing them for buildability by ordinary builders,
out of common materials.

ABD does not support that – it’s typical of what has been lost to commoditisation, ever since the Reagan/Thatcher years.

The answer is as simple as ABD Connect tools that work on all Form faces equally - that's all it would take to turn ABD into a geometrically competent Building modeler, free of its dumb cuboid limitation. For a start - of course there's more.