Thank you for this little tutorial, but you know, when I do something, I know what I do and how I must do it. I do not even understand the purpose of your comment about the masks. I'm just adapting mapmodels/skins/... existing (that work perfectly and are very good). For the mask, I had two options: either remove the layers G & B, or, configured correctly their mapmodels .cfg to save me a lot of time in unnecessary work for a rendering and a weight exactly the same.

We will never agree on issues of folder structure and file naming! I stay in a rationalist point of view that, in fact, I have to adapt to the current structure of crap.
In the absolut, "packages" should be called "content" in which we should have a folder "gamemodels" and "mapmodels" ... So no, I do not agree: these mapmodels are perfectly positioned under the current structure of the folder "packages". And I want to make you notice that you are free to place them where you like.

Now for the names of folders and files, I do not personally see no problem because usually we do not load a mapmodel on the fly but it plans to use it in the ".Cfg" of its map. In this case, just copy/paste the declaration without worrying about the rest (the names of files/folders).
So yes, in this context, I opted to promote a clearer structure and ease of integration and improved handling. I could have your way indeed, but in fact, as you can see, me, the author of this package, do not have it done. Once again, free for you to waste your time to restructure the whole (which works perfectly) at your convenience.

Concerning the name of models, if I did not use the old "tri" of MD2 is for several very good reasons. The first is probably that (did you know), we are in the 21st century and that Sauer is, well, more advanced than Quake, that nowadays it is even more limited in 8.3 format, ...
The second is perhaps opting for a dual naming (folders and files) can subsequently manipulate the files more easily. A simple example would be an author creating a folder for the map but want to group all. "Md3" in a single file to manipulate / correct more easily. All appoint "tris.md3" would require to lose so much time to rename them one by one ...
And finally, the last reason is probably that, as you do not know maybe not, the MD2 bumpmaping do not support, so it avoids the confusion by trying to question him who would convert them to the old format.

What can i say you more except that you quibble about things without any real interest !

In any case, I have adapted exactly as they should be. And if, moreover, you can do better, then I invite you to do so and I would withdraw this node.

As the "masks" seems to have a higher value in the blue-channl than others, that's I notified you just because of that (since it's not logical plants have an environmental-map).

I never said anything about the "packages" folder, as models will always get searched in the "models" folder, so your not even able to load a model from the "packages" folder.

It takes just as much time to write down the model in a config as renaming the name of the model to tris, which can be even copy-pasted (so actually takes even less time and can take even lesser time, as there are programs to rename multiple files at once like "LupasRename").
It also looks way more professional if models all carry the same name, as they are all models (and just name the folder where they appear in to them).
I know bump-mapped skins for models which have a .md2 format won't get supported, but what does that has to do with your models? Seems like an irrelevant reason...

By the way, there is no such thing as "bumpspecmapparallaxworld" for models. The command "mdlshader" should actually only be used for any costum shader which isn't defined in "md3bumpmap" command or in the masks of the model.

Let me handle it, I'll be done soon...

Edit:
Done!
Took me like 20 minutes to do so. Everything has been renamed, config-files has been cleaned up, deleted the useless masks (as plants won't have much of that anyways and you've putted glow-map rendering off). It's downloadable just right here!
Next time, use a number from 0 to 100 when playing with the specularity, as 0.25 won't get rendered.

Yes, the command (mdlshader "bumpspecmapparallaxworld") is probably too much, but if it has no influence, there is, in fact, no problem.

I think you misunderstood my approach. The masks were done, so I had a choice between redo them, or ensure that the green and blue layers are not considered (glow and envmap). In the first case, I would have asked a lot of work for absolutely nothing, since in both cases, the result remains the same render and the same weight for images !

So yes, it is not "logical" that the mask has the information it does not need (glow and envmap), but it is not logical to brew wind for nothing when simply configure the .cfg file so as not to be considered !

To the question of "tris", once again, if you want to collect all your "tris.xxx" in a single folder by copying them to rework easier, not only will you be forced to rename them but in addition, you will have to rework the ". cfg". So this solution is, in my opinion, the worst option.

I know the masks were done (as I checked the website of course), so I haven't misunderstood your approach.
The problem was this time, the masks did not had any purpose at all as you've defined a wrong value to "mdlspec" (as it's not taking any half percent), which was resulting the masks where just rendered with "mdlspec 30" (as default).

Very true about when certain files "need" to be in the same folder (as the "leaf" and "trunk" skin). That's what I've forgot to tell as exception,
even it can be avoided by just added a folder as "leaf" and "trunk" and define them in the model's config-file. But keep names just short as it saves time to write down and it can save up file-size (when it's defined a lot of time, take for example if it would be defined in all official maps).

Edit:
Multiplied the models by 2 where the the duplicated models are rendered as "dry" models (just because it can :P). Saved up some more file-size as well...
To download it, the modelpack is just right here.

My mdlspec is wrong, it's true (bad copy/past), but it have not effect on the mask render (with the default spec. 30 relative to the mask). So, where is the problem ? Simply nowhere !

The short names are not important in file sizes, it's nitpicking. What we gain to make life difficult as you want it? 25 bytes per file? WOW, it's crazy ! Do like everybody : use the copy/past fonction of your text editor !
Anyway, nothing is perfect and it is certain that. Cfg can be improved / optimized, but personally, as it works perfectly well, I do not change anything. So if you want to do, do not hesitate.

There is no problem, but it just could be optimized (as the masks isn't necessary at all).

I know it's nitpicking but isn't meant like that. As a world like "bumpspecmapworld" would be copied and pasted over 100 times, you do notice an effect on file-size (as it's not much, it still does).

Tests done and see that the mask is useful for the fineness of bumpmaping !

If you copy, even 300 times a word of 25 bytes (7.32 KB), it is very far from the comparison with the gain provided by resizing an image and a single (skin, mask'n co). So it is definitely not on the same scale of gains/losses ;-)

Well, apart from that, do you think these will be useful mapmodels ???

I wanted to say that if someone wants to thoroughly rework this package, there are things to improve. For example, animate these models (it would be nice but I do not have time). Add to the last package mapmodel of the author (which I forgot to include). Repeat the texturing of the trunk of the palm tree. Add the fern bi-face. And so on...