Does mate sense make sense to everyone else?

traveler_hauptman

I have had very little time to actually use Onshape in the last 4 months and today was the first time I reviewed the limit mates.

It took me a little while to figure out why the mate offsets and limits were not working for me.

I finally figured out that while I expected the offsets and limits to be relative to the first part in the mate set, Onshape was using the second part as the base.

Has anyone else been caught out by this?

I know that I often cause a similar 'bug' in my own software when I'm writing code and drop the numbers into a transform matrix without stopping to think about whether I'm doing a body centered or world centered operation.

@support Is there a rationale behind the current behavior? I checked the help but the documentation does not mention how the offset is calculated at all. Nor is there any hint in the UI. Is experimentation (wasting time) the only way to understand how this control works?

andrew_troup

I agree that this aspect of limit mates is currently unduly cryptic, bordering sometimes on perverse.
I also agree it would be good to have an "under the hood" exposé.
It's hard to know what improvements to recommend until we understand what we've been given.

ben_basseches

@andrew_troup and @traveler_hauptman, This is a case where it could have been implemented either way and half of users would be initially confused.  

The current behavior makes sense if you think about "Mating one part (i.e. the first selection) TO another (second selection)."  If no offsets or limits were selected, the first Mate Connector (MC) would be moved directly to the second one.  In the case of an offset, you are telling it that you want the first MC to be placed X distance away from the target MC and the direction of this offset is determined by the axes of the target MC.  Same goes for limits.

I believe the current convention reflects practical assembly processes (e.g. securing a bolt into a threaded hole).  I agree, however, that this should be better documented. We will put in a ticket.

traveler_hauptman

@ben_basseches Confusion is caused by ambiguity and lack of information, but I agree that some people are body centered and others are world centered.

I was playing with the mates to get my head around my thoughts and was reminded that you guys have still not figured out how to handle ordered lists in your user interface. (ie, when editing the list you cannot change the order of the elements. If you delete the first element to select an alternate, the order changes and the whole mate is buggered until you delete everything and start over.) It's a separate issue though.

"Offset from second mate connector" and "Limits relative to second mate connector" would be an easy first step. Useful even as a tool tip if the designers are afraid of too many words on screen.

Highlighting the coordinate frame (mate connector) being used when entering numbers into the offset fields would be a nice visual feedback.

malay_kumar

traveler_hauptman  Thanks for the feedback. We are planning to show more visual feedback of mate measurements and its reference coordinate system. In the mean time we will improve documentation. The things are a bit complicated because of primary and secondary alignment options. The measurements are from second to first and resolved in first mate connector coordinate system. I also think a lot of time user are not paying attention to which one is selected first and second and so graphical solutions will be better. Also which part is fixed and which part is moving makes it difficult to track as what limit numbers to input. Most of this is just a convention and we picked one and want to be consistent about that across all mate types. 

traveler_hauptman

...huh I just explored the effect of primary and secondary alignment with the limits and it exposed another issue.

I am fine working with one coordinate system or the other (the first mate connector or the second) as the reference as long as I know which one to use. My natural tendency is to use the first element as the frame of reference but I understand bens line of thought and once the documentation or interface gives me a way to learn the correct frame of reference to use I'm happy with that.

But having the offset be relative to one coordinate system and the limits relative to the other coordinate system does not work for me.

Dealing with pose and transforms is a mind bender and I respect how difficult a problem it is to figure out a good interface.

SolidWorks solves the problem by letting the user choose which frame to work from (via the little "flip" button) and then puts everything, the offset and the limits, in that frame.

I think you guys should do the same. Let the user choose which frame is important to him and then have all user input in terms of that frame.