two connected meshes question

I’ve noticed that in a lot of tutorials and Blend files from blend swap when two objects are put together in a model most artist tend to just stick them together without connecting the vertices into on solid object. Is there a reason for this and does it affect low poly modelling for games? Thanks.

Are both of these methods acceptable when designing models? What advantages does one have over the other? Thanks.

The advantage of first and second images will make smart rigging because you will be able to move the smaller box with a bone (surely if you rig with the right connection) while the second will pull some vertices of the bigger box.

It’s mostly because of what yassiramry says but there also other reasons you can consider:

  1. Polycount:

In the first example you have 24 triangles (12 quads, but people mostly use triangle count) (you can even go to 22 triangles if you optimize)
In the Second Example you have 28 triangfles (14 quads)

  1. Selectability:

Sometimes I disconnect some parts of my mesh so that I can select them using L-key. Useful for when you don’t want to select the entire model but just a small part.

  1. Unwrapping complexity

Ties in directly with selectability, it can make your texture map less complex because you don’t have to deal with a lot of stretching

That makes a lot of sense. I spent most of my adult life working as a carpenter, so the idea of having a box floating within another kinda freaks me out. I wasn’t sure if these models were done this way as a time saving measure, but it appears to have a real purpose. Unfortunately for me, I have to modify these meshes so that I can use them in Unity. Is there a simple way to seam the two meshes together, or do I have to do it manually?