SIMPE DEFORM, ARRAY, CURVE, BOOLEAN, BEVEL etc
Simple deform and curve
Bevel
Array + curve
Array
A bunch of modifiers needs to be fixed after, so yes, the merge modifier is really needed.
If it’s easy to write, please do it as fast as possible!!!
Not to undermine your enthusiasm -I would like something like this, too - but it’s been discussed at length before and I believe Campbell (if I remember correctly ?) said something along the lines of “the modifier system is not designed for that kind of small operations since it creates an entire copy of the mesh for every modifier” (paraphrasing from memory). However recent changes to derivedmesh may have changed this. Not sure.
Campbell isn’t a modeler, he doesn’t see the point of this modifier.
We use a lot of modifiers, one more is no big deal.
Or at least they should add a merge threshold in the current modifiers like I asked several times.
This Merge Modifier is really needed, more than any other modifier IMO.
Non destructive modelling is the new day modelling, it is fun, efficient and gives users superior control over their mesh with ability to go back and tweak parameters easily, stress-less and always knowing you can go back to to previous parameters without needing to worry about applying modifiers until you are 100% sure with the design. If blender is not taking this path it will be left behind.
Currently we are in a very good spot and we can do a lot, but merge modifier is one of the key aspect of this workflow that is missing. I wouldn’t need to apply modifiers anymore and run remove douples (merge distance) I could simply add this modifier and keep editing my mesh parameters and having it perfect until it would be verified for final touch. I encourage you to learn how modifier modelling works and you can cut your daily modelling time by half…but we need merge modifier to complete this workflow.
I guess once more procedural workflows start to appear in blender all of this will be needed (edge groups, face groups, merge node etc) - i think jaques even did some documents about this topic.
But as always its hard to project future requirements …
No need, ND workflow is already there and we need merge modifier more than edge groups etc
I love the idea of unifying the various edge marking systems into edge groups (read: multiple seam sets and sharp sets, in addition to multiple bevel weight sets), but I can see why narrower implementations have stayed on the back burner.
Are you thinking of a merge modifer as a stand-alone modifer, or something in the vein of the Sculpt Branch remesh modifer’s boolean operations? It introduces the concept of sub-modifers in a nifty way.
I’m not aware of what is going on in the Sculpt Branch remesh modifier and sub-modifiers. I’ll have a look.
Here’s a quick screengrab:
It’s a neat way to incorperate secondary operations. It could also serve as a way to limit the scope of a merge sub-operator to the vertices affected by the main operator, without the user creating vertex groups.
There is an edge split modifier. Using merge modifier influenced by a vertex group with animated weights could allow to easily animate tearing of a mesh.
Or merge could be triggered by edge marks.
This would allow animation of closing of an envelope made of several parts.
Think about edge seams existing between cloth pieces.
On the opposite, disabling the merge would allow animation of hole/door opening.
An animated merge by distance setting could be used as a kind of decimate for LOD.
Globally, it could be a kind of step value for any kind of animation of modifier from Generate or Deform category.
Influenced a vertex group and after a dynamic paint effect, we could animate any kind of asymmetrical collapse of any kind of asymmetrical mesh.
We could animate disappearance of part of a mesh in another way that with mask modifier that produces a hole or boolean that produces a cut.
As pitiwazou said as a basic modeling operation, it could be used in many cases to clean-up meshes.
In sculpt branch, there is an OpenVDB remesher as an operator and as a modifier that handles several boolean operations.
Most of remeshing methods are producing a lot less or a lot more polygons than original mesh.
The idea of a merge modifier is pretty far from that. Because most of time, the idea would be to produce a slightly different mesh topology or a topology that looks like obviously generate from original basemesh or a polycount close to it.
I don’t see a merge modifier as merging two different objects into one as a better solution than a boolean modifier or an OpenVDB remesher.
Existence of a simple Merge modifier does not go against the possibility to use it after a boolean or a remesh modifier.
Edge groups would also be a great addition. Bevel is a clear example of a modifier that would need that. Different Bevel edge groups for different Bevel modifier.But it is true for any use of edge marks.
Different Seams groups for different UVmaps.
Different Creases groups for different Subdivision modifiers.
Different Sharp groups for different Edge Split modifiers.
Different Freestyle groups for different linesets.
We can all agree that we would find uses for that.
Edge groups could be a game changer in overall modeling workflow and can be a huge step forward in productivity - both modeling for games or movies (high poly, mid poly) but also concept art when changing things really fast in fly (non-destructive) is helpful and sometimes important.
The reason is that non-destructive nowadays is crucial. Having possibility to assign bevel width for specific line would be amazing, because thanks to that I can control object bevels more freelly. Bevel weight is cool but not everytime. Controling different type of bevels through angle is bad and you need to be careful to keep good angles values on mesh.
Below is simple example how this could look like
I know that there are some addons that allows to make your work easier and faster, but still first, it’s often based on features that Blender already have, and second thing I don’t really like complex addons because they are often writen in own artistic thinking way. We should make Blender more consistent.
Edge group is important but it will take more time to do than an easy merge modifier that will take at least several hours max.
Another Example Howard, the merge on the screw modifier doesn’t work.
Some modifiers have merge option, but it is not working or not entirely.
That’s why a merge modifier is needed, absolutly needed!!!
I think the merge on screw is only merging edges that meet when the angle is 360 and there is no offset.
I guess there are two choices: make all the modifiers mentioned here have a merge option that really works; or make a merge modifier. I think ideally we’d do the former (because then you can have it only affect vertices resulting from what the modifier did); but of course the latter is much easier.
I vote for both options ^^
1 - Fix the current merge on modifiers
2 - Add a merge modifier
The Merge modifier is like the Weigthed Normal, even if the Bevel have some options, I prefer to use the Weighted normal to be sure that the normals will be ok on the resulting mesh.
Merge modifier is also better just because it eliminates need to copypaste the same code into all other modifiers and maintain it everywhere.
Maybe it will be good idea to implement it as “merge by distance” method in decimate modifier, for me it makes sense to put it into “that thing that reduces geometry complexity”.
Today on Blender Today stream we must spam in chat that we need merge modifier 
Good idea we most do something
What if add a more powerful clean-up modifier? https://blender.community/c/rightclickselect/smdbbc/
That’s a good idea!
I firstly thought that merge could fit well in the edge split modifier*, but putting it into decimate as an option makes more sense to me.
*edit: the modifier of course would be renamed something like edge split/join 
