Hi mushroomgod - yeah I definitely agree that a right-click on the hide should be leveraged.

But I just want to make sure I understand correctly - clicking twice on the button will end up doing a "show all" (show all hidden objects so nothing is hidden anymore) which is different than what you are talking about, is that correct?

You mean making the right-click do what could be called an "isolate selection", which hides everything else but keeps the selected objects visible?

- Michael

Hi Michael...

I am agree with you for irregular shape and group selection, just show editing bonding box base on xyz.
First, I know we have talk many times about constraint base, sorry just for your inspiration in the future :) :


Thanks

yeah.

the whole clicking twice thing confuses me no end!

Basicaly while modeling stuff I have curves and old bits of model all over the scene, most of the time/all of the time I just want to work in few tiny objects. And what I end of doing is selecting those objects, invert the selection and hide all the irelivant objects.

Personaly speeking I would be happy to lose the double click buisiness (on the hide button) and have it so simple even an idiot like myself can figure it out.

So......

Click the "hide" button with left mouse hides what you have selected
Click the "hide" button with right mouse hides the inverted - so select the object you want to work on and hide the stuff you dont want to work on

Ofcourse this brings up the problem of getting all your hidden objects back...im not a fan off adding buttons neadlessly, but ever 3d app iv ever used has a "hide" and a "show all" button.

Maybe is you added a "show all" button the same kinda thing could happen

Click the "show all" button with left mouse un hides everything
Click the "show all" button with right mouse inverted whats being shown - hidden objects become un hidden and and visible objects become hidden?

I agree with Burrman -

Even though MoI is intended as a fast concept modeler rather than a CAD system, even an artist will have certain base dimensions in mind when drawing a car, a building, or a product.

Let’s say I'm designing an iPod and I start with a 2" by 3" LCD display. I want that LCD rectangle to show as 2 x 3, regardless of its orientation. Now I want to play with my design and try another screen size - so I edit the screen rectangle to 3" by 4". I would want to select the rectangle, change the size, and not worry how my drawing is positioned.

Same idea would apply to changing the size of doors and windows in an architectural drawing to elements having standard dimensions.

For me, the bounding box dimension would have little meaning in these situations. For odd shaped objects or multiple selections, then the bounding box dimension is the one to use.

Hate to suggest adding complexity, but perhaps a way is needed to show both dimensions, or set a preference option to show one or the other.

Ed

"I agree with Burrman -

Even though MoI is intended as a fast concept modeler rather than a CAD system, even an artist will have certain base dimensions in mind when drawing a car, a building, or a product.

Let’s say I'm designing an iPod and I start with a 2" by 3" LCD display. I want that LCD rectangle to show as 2 x 3, regardless of its orientation. Now I want to play with my design and try another screen size - so I edit the screen rectangle to 3" by 4". I would want to select the rectangle, change the size, and not worry how my drawing is positioned.

Same idea would apply to changing the size of doors and windows in an architectural drawing to standard dimensions."


...quoted for agreement

Michael,
I did the test with the cad and its as you set with moi!

Where I would be coming from is more of a "scene Managment" type creation where my goal is building specific things at specific locations. It's a sub-d thing.


I understand the bounds data and the need and use for that.

In your previous post to Danny, you answered my question and gave me my workflow.

With Nurbs we have to remember that a "solid square" is actually made up of curves and surfaces, as opposed to poly we think of a square as a square only (that build to exact dimensions thing), even though poly is vertices and surfaces too, we still think of 2 inch box as 2 inch box no matter where or how its placed.

I can work as I like by remembering when I want to numerically resize, I work from the edge level, which reports the size I expect.

A minor change in thinking to work with NURBS Data.

Thanks Danny, (My little friend George)
The CPlane answer will also handle (I think) when I need to resize say a "curved or odd" shape where I need to make sure it is still "2 inches long" at a certain plane angle instead of sizing world bounds data.

I guess I have to finally learn how to use CPLane (I know I'll be happy when I have!) and then be wanting the "Setup and save" CPlane for complex operations!

Sounds like Homework assignment!

Thanks for the help.

Burr

Edge and CPlane Handle what I was after!

FYI

>"I agree with Burrman -

>Even though MoI is intended as a fast concept modeler rather than a CAD system, even an artist will have certain base dimensions in mind when
>drawing a car, a building, or a product.

>Let’s say I'm designing an iPod and I start with a 2" by 3" LCD display. I want that LCD rectangle to show as 2 x 3, regardless of its orientation.Now
>I want to play with my design and try another screen size - so I edit the screen rectangle to 3" by 4". I would want to select the rectangle, change
>the size, and not worry how my drawing is positioned.

>Same idea would apply to changing the size of doors and windows in an architectural drawing to standard dimensions."


I also agree - bounding box size is not much use to me. Never even thought of it that way (bounding box that is) :-)

Hi Burr,

> I did the test with the cad and its as you set with moi!

Yeah, I figured it would be set up in the same way...

So you had originally asked about seeing the exact dimensions so you knew you had the right accuracy for your CAM work - the bounding box dimensions given by MoI currently are actually giving you the right information for that, correct?

If I had set it up to give the same answer between an "unrotated" and a "rotated" box, then that's when you would actually be seeing a different result in the reported bounds in MoI versus your CAM app, it seems like that would be worse off instead of better for "CAM accuracy" ?


> Where I would be coming from is more of a "scene Managment"
> type creation where my goal is building specific things at specific
> locations. It's a sub-d thing.

I think you've got the idea already, but the way that you can deal with this in MoI is to set your construction plane to that particular area where you want to do some specific located work.

That basically sets up the whole drawing environment for working in that particular local area.

It involves several things:

The grid moves to attach to be oriented and centered on the location that you pick

The Top/Front/Right views switch to become plan views of that particular location and orientation, like the Top view becomes a plan view looking down directly on that local area so you can draw some points in there with the mouse and they will be plotted on that local area.

Even typed in coordinates will be in that local frame of reference, like if you type "0" it will give the origin of that local frame of reference, and things like typing in 5,2 means 5,2 in that cplane not in the world coordinates.

Then on top of that, the bounding information that is shown in that properties area will be bounds relative to that orientation as well.


So there are a lot of things that happen when you apply a construction plane to make it possible to work very specifically in that particular local area.

I haven't really had a good chance to describe too many details about the cplanes yet - I had meant to do some more description when they first came out but haven't managed to do it yet.


> I can work as I like by remembering when I want to numerically
> resize, I work from the edge level, which reports the size I expect.

Yeah that should work well, when you select an edge that is a line it will report the length of that edge which is not dependent on any kind of orientation.


But I think overall your concern about the CAM accuracy is well covered by how things are currently set up...


- Michael

Hi Tony,

> I also agree - bounding box size is not much use to me.
> Never even thought of it that way (bounding box that is) :-)

The basic idea is that it gives you some general idea of the size of your object.

Like if you see sizes over there of something like 2000 x 1000 x 5000 , you can have an idea that your object is generally a few thousand units in size.

If you see something over there like 0.01 x 0.06 x 0.02 you can get an idea that your object is something around 0.05 units in size.

It's not an unusual thing for people to want to have some general idea of what kind of scale the object is set to, that's the general idea behind the bounding box readout there.

Otherwise when you have the grid turned off, it is pretty hard to get an idea of an object's size by just looking at it on the screen.


For special objects like lines and circles, it is possible to give a more precise and "orientation free" dimension result like a length or radius value. Those also work to give you an idea of the scale of the object as well, so those fit in nicely there.

If there is not an "orientation free" analytic value that can be used, then you just get the bounding box. If you want to control the frame of the bounding box that is used, that is possible to do with by setting your cplane first.


Anyway, that is some of the basic ideas and uses for this new widget.

- Michael

Hi Anis,

> First, I know we have talk many times about constraint
> base, sorry just for your inspiration in the future :)

Actually, that does give me a good idea for something - I think when you have 2 adjacent control points selected I should give back a length value in that case rather than a bounding box value.

That would allow you to change the distance between those points easily, basically allowing you to set the size of one part of a polyline when edit points were turned on.

But the other constraint parts there that you were showing with the bottom part of that irregular outline moving when you set the distance actually seem like they could cause problems instead of being a good feature for working with irregular outlines...

- Michael

Hi Ed,

> For odd shaped objects or multiple selections, then the bounding
> box dimension is the one to use.

What is the definition of an "odd shaped object" though - is it anything other than a rectangle?

Does taking one corner point of a rectangle and moving it to the inside a bit change the rectangle to then be an "odd shape"? If so then that means that moving a rectangle's point by a tiny amount in one direction would cause the whole shape to report a completely different size than the previous rectangle's size, even though it was just a tiny edit.

I think that's a pretty bad problem... I just don't think that method is viable, it's too brittle and not general purpose enough. It would not really be of much use unless you were only modeling with rectangles and no other shapes - that is a pretty big limitation.

The current system that I'm using is general purpose and doesn't have that kind of "morphing the width by height dimensions after small change" problem.

There is another alternative that would be more in line with what you are asking, but it involves a fair amount of work and various technical difficulties and side effects. That system would be to keep a "local axis" defined on every single curve that was created, so that for example if you edited the rectangle and moved one of its points over a bit, it would still remember the original rectangle's orientation and report its size in that orientation so there wouldn't be a sudden shift.

That involves a whole bucket load of problems though, it probably requires methods to see and edit those local axes, and typically having an local axis showing up right in the middle of a bunch of curves is a rather weird thing in many aspects, it kind of takes away some of the simple "2D drawing" type behavior, like it is just strange to be working on a drawing made up of a lot of lines and have all these special "axis markers" in the middle of each individual line segment, filling the screen with odd marks... But if you don't show those axis markers prominently then it becomes hard to know what directions are being used for the bounds reporting. Another related way is to make all curves live only on a "sketch plane". That can prevent proliferation of tons of "local axis" marks all over the place, but then tends to have the bad effect of restricting your freedom to just draw curves or lines in space at arbitrary 3D locations.

So there is kind of a whole load of baggage that comes along with such things that I'm not particularly hot to dump into MoI...


> Same idea would apply to changing the size of doors and
> windows in an architectural drawing to elements having
> standard dimensions.

This would probably be better handled with a kind of "component" system like SketchUp.

Instead of trying to stick a "local axis" on every single curve and individual line object, it would be better if it was only something like a "component instance" or "block instance" type object would have such an axis. That would be more natural for such an entity to prominently display the axis when it was selected, it would not be as weird as seeing a line drawing with a zillion axis markers in the middle of all the lines...

So I'm thinking it may not be until that type of mechanism is available in MoI before things would work totally in the way that you are asking about. I'm not really sure when that will happen, there are some other kinds of basic object organization mechanisms that I'm probably going to try to get working first before that.


The way the current system works, you can see and edit the proper dimensions that you are talking about though if you set the construction plane to be oriented along that plane.

So for example if you want to edit some curves for a window, set the construction plane to the wall that they are on, then when you select those curves the bounding information that you see and can edit will be the bounds oriented in that wall. That should give you a way to make those kinds of edits until the time that a component type system is available. This system also works with curves of any shape, not just rectangles.


Anyway, those are some of the issues involved as for why the system works the way it currently does.


- Michael

Hi mushroomgod - re: hide / show options

One problem with the setup that you are mentioning there is not having any way to show just some of the objects that were hidden instead of all of them.

That was a pretty heavily requested function for hide/show capabilities.


That's the reason why there is some added complexity to the current hide/show mechanism - when you click twice to show all, there is also another alternative, which is to click once, then you see all the objects that were hidden, and if you want to only show just a couple you can pick those and then push done (or click the hide button or right-click in a viewport which are aliases for pushing "Done" in this case) and only those particular ones will be shown instead of all.


The request for this came up so frequently that I made a pretty difficult decision to incorporate this kind of additional power directly into that hide/show mechanism instead of just doing a one-click "show all". It is a case where I decided to sacrifice some simplicity to gain more functionality and power.


It's a pretty significant UI design benefit to have only one button for hide/show, in the future I'll probably want to also add a "Lock/unlock" mechanism, and having the full mechanism encapsulated in one button makes that a lot more possible - having 2 buttons for each of these kinds of functions makes it much harder to incorporate additional ones.


I'm certainly up for making right-click do something, and I think there are actually 2 separate things that I can do for right-click - I can do one thing if you right click when there is a selected object (isolate it by hiding all the non-selected), and if there is no selection I can do some other action when you right-click, which could be used for a "one click show all" I suppose, if that would help.


- Michael

>> Anyway, that is some of the basic ideas and uses for this new widget

Thanks Michael. Not really the accurate way of driving the size I first thought it was. But, as you have said that's a whole different ball game and will have to come later.

Cheers.

Hi Tony -

> Not really the accurate way of driving the size I first
> thought it was. But, as you have said that's a whole
> different ball game and will have to come later.

Actually if I understand what you want correctly, you should be able to use this feature as it is currently set up for getting and editing the exact size that you want.

With angled objects, you will need to set your construction plane to that particular object's plane first though.

Once you have done that, the data that you see and edit there will be the sizes in that particular orientation - that should provide you with exactly the stuff that you were asking about.

Another way to do it is to select a linear edge instead of a whole object - when you select a linear edge the length of the edge will be shown and editable, and that is not dependent on orientation so you do not need to set the construction plane first even for angled objects if you use that method.

That's another way that the current system can be used to get the precise information that you are looking for.

So there are a couple of different ways that the current system can be used to do the job that you want, unless I have misunderstood what you are looking for?


I guess the thing to watch out for is if you rotate an object, the world bounding box of that object will be reported until you set your construction plane to the same orientation as the object.

The part that should get better in the future is with things like "instances", those will likely carry around their own axis information as part of the object itself, which will make it easier for those objects to report size in its own axis directions immediately without relying on the construction plane's axis directions.

- Michael

Thanks for the tips on using this new function, Michael. It looks like the bit with the single edge is what I'm after.

Cheers.

If I dissapear from the forum it's probably because this has got the better of me.

I am tooooo old.

Brian

Hi Brian, I'm sorry I'm not at all following what your image is showing...

You've got a big red arrow pointing to the name "Options" that is the title of the Options dialog?

Is there something wrong with the title?

- Michael

The key board options don't!

And, I am not interested in complex methods to reinstate.

Sorry. I exist one day at a time. (LW9.5 is in the rubbish bin)
I should not need to call on the likes of Burr to be my saviour.

Brian

Being odious---3DC is on it's 41st V3 Beta update since Sept 16th----have not had to do a single "Brian".