Can someone clarify for me please? I want to prepare some clothing models for sale as objs. I’ve retopologised so I have neat rows of quads. But … if I have add-ons, like flaps or other added strips of material - is it normal practise to leave these as separate objects? With pockets, this seems to be the case. If I merge vertices of extra fabric layers along seam lines, this messes up the topology flow. Is this simply counter-produtive and not what someone would like to find in their purchase?
Welcome to BA 
For paid clothing materials, pockets and other such additions should be one object with the base. Also, you should retopologize in a way that you still have good quad flow and no overlapping geometry, even with the pockets. Good quad flow, no overlapping geometry, and no loose parts is the bare minimum standard for a paid 3D object, at least one intended to be animated. If I bought a piece of clothing to animate and it had loose parts or bad quad flow, I would ask for my money back immediately
Thanks so much for replying, very appreciated. Where I’m still getting confused is, if I attach a piece of ‘loose’ geometry, such as a flap or pocket, to the base - by merging vertices, then this messes up the quad flow. You get six faces meeting at one vertex.
I might be using the wrong terminology of course, but by ‘loose’, I mean pieces that have not been merged with the base, rather than just ‘joined’. It’s still seeming to me that if you model a complex garment, either it’s all merged, and with quad flow messed up all over the place. Or it has separate ‘bits’, each with good quad flow. What am I missing here?
This figure may explain it better. If the outer plane, representing a sewn-on piece of material, is merged nicely with the lower (main) one, the quad flow ends at that edge.
When it comes to animation and marketing objs, is ‘parenting’ each part to the main one, an acceptable solution?