Moving a face by distance measured not perp to the face

Tom Foster

That's interesting.

if you first extend the roof beyond the helper entity

Or more generallly, 'if the roof's side-edge and the helper-Region's edge are non-coplanar' (which is most likely to be the case).

Tom Foster

@Hans De Backer said:> Now, we could change a clean and smooth command like pushpull into a complex one and add options ... (it certainly would make the command less straightforward)

No, absolutely not

Tom Foster

Tom Foster said:

this has little to do with freeform/curvy stuff like Zaha Hadid's, or Brics' 'make curtain wall from curves' feature

Hans de Backer said:

Well, your personal preference may be to ignore curved faces, but we can't

No no, please press on with the curvy stuff, which no other BIM attempts to tackle
(e.g. the curvy stuff is done in Mictostation then imported to AECOsim BIM for its data facilities; AECOsim itself can't touch it).
Well, the v big-boy architects use Gehry Technology's Catia-derived Digital Project.

I just meant to be clear that I'm not talking about that prestige freeform architecture, nor standard rectilinear architecture, but something forgotten and neglected, in between.

It's the difference between Zaha Hadid's 1990s Vitra Fire station
https://www.archdaily.com/785760/ad-classics-vitra-fire-station-zaha-hadid-weil-am-rhein-germany
which launched her career, making extraordinary rigorously geometriic shapes out of commodity straight flat materials, before digital curviness was even possible;

and her late production (until recent death) such as the Heydar Aliyev Center, stranded whale to the prestige of Azerbaijan's hideous dictator
https://www.archdaily.com/448774/heydar-aliyev-center-zaha-hadid-architects

If I can design and document it, ordinary builders can make a Vitra out of standard builders-merchant materials, at ordinary cost, for local communities - maybe not so wild-looking, but housing, workplaces, social buildings optimally shaped to site and function, not cuboid as the only option, and certainly not that digitally-prefabricated prestige curvy stuff.

If iconoclastic Brics isn't up for providing the tools I need, I can't see any of the other CAD cos doing it, stuck between, on the one hand, the fast-track utiltarian commercial, unloved and obsolete within 20 years, and on the other hand the lure of the prestige showpiece.

Well loved buildings are the ones that are shaped and rooted integral with their unique site, surroundings, landscape and social purpose, so are used and re-used. They are rarely the ones that are utilitarian-rectilinear, as BIMs assume, but are equally rarely the expensive freeform ones.

Brics devs - if you can make freeform/curvy tools in dwg, as well as rectilinear, I know you can make tools for the in-between - straight, flat but off-angle geometric precision - and make it easy like never before.

Michael Mayer

I agree.
But just one step after the other.
In studies I was told to start with simple straight modern architecture first.
Even all Hadids, van Berkels, Morphosis, Gehry showed that they can do
that very well before going fluid.
If straight modern architecture, which leads to enough problems and
special situations in detail level where different components meet,
finally works reasonably in Bricscad BIM, I am happy to see them
experimenting with applying multi-ply Compositions to THICKENed
NURBS faces.

But basically, your architectural style should already work with Bricscad's
current modeling options. Not as comfortable as simpler rectangular
geometries and maybe not yet as comfortable as it could be at one time,
but it looks possible for me.
I am not sure about Archicad, which is often able to do more special
situations as one would expect, but in Vectorworks your geoemetry
would often mean the need to leave the comfort of any standard
PIO BIM tools and going tedious custom modeling and manual tagging.

In this regard for me, Bricscad with Solid Modeling from the beginning
and easy BIMIFYing seems more flexible and comfortable.

And for Free Forms, there are already things like Curtain Wall Grid.

Tom Foster

basically, your architectural style should already work with Bricscad's
current modeling options. Not as comfortable as simpler rectangular
geometries and maybe not yet as comfortable as it could be at one time,
but it looks possible for me.

True, but when I finally started using 3D CAD
(first Bricswork 1995 which became Triforma, abandoned,
later AECOsim, arcane, few few tricks, workarounds all the way,
finally Brics incl BricsBIM, a complete, useable delight moving on from fluency with Acad 2D),
I actually found flat/straight but off-angle architecture harder than in 2D, surprisingly and disappointingly.

I really thought I'd be able to swing slabs easily into 3-axis (non orthogonal) geometric alignment with a complex of other dittos -
for example knowing it should touch this point, skim a virtual alignment of two other edges, and tilt just so as to come horizontal at that edge,
but not knowing actual dimensions or angle to be entered.
An easy geometry-solver, parametric without using parametric tools as such.
Having then trimmed/extended elements to meet, I would finally discover the dimensions and angles, instead of mind-bending trigonometry.

Some of these intersections I could do with 2D tools, but very far from it in 3D. Instead a lumbering inflexible requirement for absolute dimensions to be entered mostly. T and LConnect, Extrude to limit etc are great for the final trimming but help not at all with the geometry solving.

Even the super-simple case in this thread requires workaround, despite being streets ahead of AECOsim (which is v similar to all the other BIMs).
This case just scratches the surface of what's needed.

Michael Mayer

I actually found flat/straight but off-angle architecture hard

OK.
I understand the wish for your current :
"Move Face along normal but input dimension in world axis"
comfort, to avoid the need for trigonometry and helper objects.

For the parametrics,
I would also like to Bricscad have the global coordinate/rotation/scale
Readout, with direct editing option, that Cinema4D or Blender have.
(like selecting a Wall and typing a 15° banking angle)

Modo's "Absolute Scaling" allows the same but is far more tedious to use.
Vectorworks has it perfectly for 2D Objects and partly for PIOs,
But nothing for Solids and you can even edit one at a time only.

I am still fighting with things like Bricscad's (or Autodesk's ?) old Rotation
or Alignment Tools logic etc. But I am fine with Bricscad since Manipulator.
But all the latest DM and BIM Tools makes it really powerful and comfortable
for me.
Connect to Nearest in Microstation or VW would mostly mean a few steps of
boolean operations and/or some manual cleaning up for me.
I mean that meanwhile Connect to Nearest works although your Wall
already intrudes the Wall's Face to connect to - is great!
BIMDRAG works magically as expected in nearly all cases, without crashing
Bricscad.
Not 100% sure but with correct Modeling Settings, I think if you rotate one
of your Slabs (or its Faces one by one) , it may already adapt/cut connected
Walls (?) in one go.
Or if you change a Wall's banking angle while design process, connections
can be repaired relatively quickly.

Roy Klein Gebbinck

@Michael Mayer
What is the meaning of 'PIO'?

Tom Foster

Really appreciate your appetite for discussion Michael - no better way to clarify my thoughts, than to write responses, I find.

Just want to re-emphasise -

T and LConnect, Extrude to limit etc are great for the final trimming

So your praise, above, for such tools in Brics is right and interesting

but [they] help not at all with the geometry solving

that I'm talking about in this thread.

Michael Mayer

@Roy Klein Gebbinck said:
@Michael Mayer
What is the meaning of 'PIO'?

Sorry,
Vectorworks calls all parametric building Tools
PlugIn Objects (PIO).

And there is a heavy use of funny acronyms. like :
(DLVP) = Design Layer Viewport
(SLVP) = Sheet Layer Viewport
(OIP) = Object Info Palette (Properties Palette)

Tom Foster

@Michael Mayer said:
it should work by selecting top+bottom Edges, which allows DMmoving along WCS distances, while still following the Slabs slope to keep edge integrity.

@Michael Mayer later said:
Moving/Pushing Edges didn't work in my current test .... Dimension input was along correct axis - but Face or Edges moved to another distance anyway.

Trying this now anyway - but how to make the dmMove to be along WCS distance?

Michael Mayer

I think that seemed to work when you are in a 2D Side View
and the trick to select both edges before starting DMMove.

Tom Foster

@Michael Mayer said:
But I also tried to move the sloped face by manipulator, by rotating the manipulator from Face Normal to a World Axis.
But this will fail because when I input the distance in world Axis, it will move the Face for that distance - but in Face Normals Direction anyway.

Yep, and for why, I ask? what's the point of being able to realign Manipulator, to WCS or anything I like, if it makes no difference? The distance typed in WCS direction isn't applied in WCS direction, but stubbornly perp to the selected face, as if I'd left Manipulator in default orientation.

Being able to reorient Manipluator seems to be a most potent facility which, if it actually did anything, just might obsolete my wish for

an optional subsidiary UCS that can be defined, that works only to set the plane/normal (incl SHIFT to lock snap tracking) of

• the distance of the Move,
• or the Move's snap-intersection with plane, edge or point, or with a projection of plane or edge,
• or the Move's perp-to-ref-plane projected alignment using 'ignore entity snap elevation' and 'entity snap to negative Z'.

or as Michael puts it

to separate Extrusion Vector from Dimension Vector and do the math for you.

Tom Foster

duplicate deleted

Tom Foster

@Tom Foster said:
how to make the dmMove to be along WCS distance?
@Michael Mayer said:
I think that seemed to work when you are in a 2D Side View and the trick to select both edges before starting DMMove.

I got dmMove OK along WCS now - in 3D as well as 2D side view.

The difference is:
a) Select eave face>dmMove, the eave face moves normal to itself, preserving side faces, but the distance moved is 300 normal to itself, even tho the 300 distance was entered in WCS plane.
b) Select both top and bottom eave edges>dmMove, the edges move in WCS, properly by the 300 distance entered in WCS, but this parallelograms the whole solid.

So tho the latter obeys the WCS in which the distance was entered, it's still not what's reqd.
What is reqd is the face to move as a), by distance as b).

Michael Mayer

Maybe I selected the Face when failing ....

If I get that right,
you basically want to e.g. go in a 2D Top View,
grab your upper eave face's edge (maybe both edges if mandatory)
and pull it in X or Y axis only over your Wall end,
grab it again and move it by another 300 mm input in X or Y axis only
over the Wall.
But still restrict the Face to keep its angle and its movement
restricted to its neighboured faces and edges.

No clue if that is easy to program and just about offering
optional input vectors and doing the math to convert to
final sloped vector ...

Tom Foster

Yes, exactly, and it's exactly same effect when using Manipulator reoriented to WCS instead of using dmMove.
Except that I'm doing it in oblique 3D view so I can see what's going on. To do it in a flat top view would take a lot of trust!

@Roy Klein Gebbinck said:
I am surprised that no one has considered using temporary tracking...

That got me excited but I can't see that it could help. Yes it would be the way to make one step out of

grab your lower eave face's edge and pull it over your Wall end,
grab it again and move it by another 300 mm input [beyond] the Wall.

But how can it help? with

to separate Extrusion Vector from Dimension Vector and do the math for you.

Roy Klein Gebbinck

Here's a temporary tracking example.
Note 1: The UCS has been set up properly: in plan the X-axis is perpendicular to the bottom edge of the roof.
Note 2: To get the tracking line along the direction of the roof you only need to hover over the 2nd endpoint.

Tom Foster

Hey that works! with Temporary Tracking the distance is entered horizontally and the eave edge complies!

What an education, in realms of Settings, background colour, marker and aperture sizes etc, to be able to see what's happening on my screen and duplicate Roy's, without spurious snapping causing vectors to disappear etc - even after deciphering what Roy was actually doing.

So this Temporary Tracking is powerful stuff - but more and more, Brics 3D, even 2D, seems to rely on constant adjustment of snaps and other things - rather than just leaving all snaps on as 2D Acad seemed to haplily manage.

This could adequately achieve the effect of what in Michael's words I was suggesting -
"I think it would be easy for Bricscad Devs to separate Extrusion Vector from Dimension Vector and do the math for you"
It turns out this is already the case - but only with Temporary Tracking. How wierd.

Trust Roy to suggest what no-one else has, neither users nor the Brics guys.

Now to test it on the case of
https://forum.bricsys.com/discussion/34057/projecting-plan-to-sloping-plane#latest

Tom Foster

In Roy's gif, 2 posts above, he gets, as he says,
"Note 2: To get the tracking line along the direction of the roof you only need to hover over the 2nd endpoint."
But I can't get that tracking line to appear. Settings are as attached, with as much dynamic dimensions etc switched as poss switched off, and 'Press SHIFT to acquire entity snap tracking path' (instead of just hovering over the 2nd endpoint).

Also, in the gif, Roy's final step, projecting vertical from the second temporary tracking point 200 out from the wall, only works when it's in plane to exactly hit the lowest/end vertex of the roof solid. Why can't it work just by hitting the lowest edge? I'd a thought 3D Snap to Nearest shd do it?

ScottS

I used to rely extensively on temporary tracking points, now less often since modeling in BC. Roy's is a good example of how it remains useful.

However, I was having the same trouble you are, Tom, unable to get the blue-violet tracking line to appear. After several tries, it would occasionally show up randomly, though unreliably, upon hovering over that 2nd endpoint of the roof. But then the tracking line would fail to reappear at the critical final step making that vertical projection up from the 2nd tracking point.

Maybe you are already doing this, but one way I got it to work is to hold down SHIFT just before hovering over the 2nd endpoint. (Another way is to enter "tt", then select that vertex itself as a temporary tracking point.) The blue-violet tracking line appears as I move away from the endpoint vertex.

To avoid losing the auto tracking line once it shows itself, I press SHIFT again, and the tracking line turns red and stays visible. See "Using Snap Track Lock". If I keep it in this red mode, wherever I position the cursor, the coordinates are projected perpendicularly onto that red tracking line. This could be useful for other purposes, but not in this case.

So then I press SHIFT one more time, which releases the red tracking line, before typing "tk" to begin the temporary tracking point selection process. Although that original tracking line disappears while I select the top of wall and horizontal offset, it seems to consistently come back now when I need to snap to it in that final vertical projection. As if temporarily "snap track locking" it in red prioritizes it in BC's memory. It works consistently for me in this example now; I need to test the method in a more complex model.

Will also think more on your last question, if the base point and tracking points are in different planes.

Tom Foster

Brilliant, ScottS - will try that.