Cannot snap to Parametric solid "through" filled shape

CE15.2

For years I've been annoyed that on occasion snaps didn't work but I could not figure out what the cause was.  If a filled shape exists in 3D space above a element in CE and you attempt to snap to the element snaps will not work even in wire-frame.  Under users>preferences>input> Locate interiors is set to never.  Is this intentional or an oversite??  Attached file for others to try maybe it's just me?

cannot_snap.dgn

-Grant

Parents
  • This has also been an issue for me - I've tried to get to the bottom of it with this thread; communities.bentley.com/.../594681

    ... but without a solution I'm still using v8i to remain productive.

    As a summary of the problem (using your file), see below. I really hope someone could provide a solution to this.

  • ^ Is this the case for everyone in this thread, i.e. trying to snap through a transparent filled shape in wireframe? 

    Does everyone agree that if the shape was not transparent you shouldn't locate/snap to elements behind it as you're not able to see them? This enhancement was made to locate for CE as there were numerous reports of undesirable behavior related to locating/snapping to geometry that is not visible.

    I can definitely understand a request to special case transparency...

    -B



  • Thanks for engaging with this, Brien. To answer your question - NO, I do not agree. I would like to select any point that I know exists, regardless of whether the concealing geometry is solid or transparent.

    I can see where Bentley was going with the logic on this - "if you can't see it, it doesn't exist" kind of thing. It's a 'real world' philosophy that makes sense, but I would argue is not appropriate in the cad world. As 3D modelers we work in a 3D world that transcends the physical world and as such would like to get the job done with as few clicks as possible. We can visualize the location of hidden geometry and should be able to snap to it without rotating views to see it first.

    As an example; say I want to draw a connecting line between the back corners of these two column plinths.

    In v8i this was easily doable with just two clicks. (ie. the plinth back corner point is immediately findable);

    In CONNECT, I have to scroll through the tentative snaps for the front column first before finding the plinth back corner point that I know is there. (ie. three clicks, just to find the first point - and other three clicks for the second column = total 6 clicks). This is a simple example - imagine much more complex geometry with multiple objects concealing the known target point.

    You've described this method as an "enhancement", but I see it as regressive - and would really like the option to return to the simplicity and predictability of v8i.

    Max

Reply
  • Thanks for engaging with this, Brien. To answer your question - NO, I do not agree. I would like to select any point that I know exists, regardless of whether the concealing geometry is solid or transparent.

    I can see where Bentley was going with the logic on this - "if you can't see it, it doesn't exist" kind of thing. It's a 'real world' philosophy that makes sense, but I would argue is not appropriate in the cad world. As 3D modelers we work in a 3D world that transcends the physical world and as such would like to get the job done with as few clicks as possible. We can visualize the location of hidden geometry and should be able to snap to it without rotating views to see it first.

    As an example; say I want to draw a connecting line between the back corners of these two column plinths.

    In v8i this was easily doable with just two clicks. (ie. the plinth back corner point is immediately findable);

    In CONNECT, I have to scroll through the tentative snaps for the front column first before finding the plinth back corner point that I know is there. (ie. three clicks, just to find the first point - and other three clicks for the second column = total 6 clicks). This is a simple example - imagine much more complex geometry with multiple objects concealing the known target point.

    You've described this method as an "enhancement", but I see it as regressive - and would really like the option to return to the simplicity and predictability of v8i.

    Max

Children
  • I also agree with Max and would make the argument why have a locate interiors option in the preferences and then ignore it.  If you want to tether filled shapes to the locate interiors options I have no issue with that, it would be more predictable then the present behavior.  What we have now just feels like the developer got half way there and took a coffee break decided never to get back to it.  If locate interiors is set to never then any filled shape rendered or other wise should never interfere with a snap.

  • What we have now just feels like the developer got half way there and took a coffee break decided never to get back to it.

    While I do enjoy my coffee way too much, this is certainly not the case.

    I also agree with Max and would make the argument why have a locate interiors option in the preferences and then ignore it. 

    It's not being ignored. Setting locate interiors to never doesn't mean pretend this surface doesn't exist and start picking things you can't see through it, it means only locate this element by it's edges, but still truncate the list of hits at this surface so that reset and tentative snap also can't cycle through elements that aren't visible.

    In the simple case of a single solid and line it might seem appealing to let you pick an element you "know" exists that is behind the solid, but Imagine an entire building in a shaded view, there can be thousands of elements under that roof.

    This wasn't changed in a vacuum one day for "fun", it's cases like that building and snapping to beams and elements on other floors that aren't visible and the feedback we received that prompted the enhancement to locate.

    You should need to see what you are picking, which could mean changing render mode to wireframe. As previously stated, it definitely would be nice to make a special case for transparency as that fits in with letting you pick what you can see.

    -B



  • Display depth and clip volumes are both valid tools for controlling how much a user has to click through in active modeling.  In my humble opinion the community in general hates CE for cases like this.  Going from v8i to CE things should have stayed the same with options to change functionality, not the other way around.  I understand why it was changed and even the consideration behind it, however that doesn't change the perception.  I've been using off on and starting at U0 in 2016, and full time for 2 years now, there are still many behaviors that feel off but I haven't had the time or energy to chase them down this is just one example.