andreuparri

BabaJ
But what's the advantage, other than personal preference which I can't 'argue' with, of referencing it to a wrangle vs. local file in $HIP ?

I suppose it is in a way a personal preference but in another way I just think it makes more sense to be node based since is the houdini philosophy in the first place.

Advantages I see wrangle vs text file on disk:

• Selfcontained. I cannot be lost that easily. If it exist in the scene is much easier to find by other artist.
  
• Easy to have multiple version. Its easy to test diferent versions by simply duplicating the wrangles.
  
• Can use presets which hopefully get even more functional in future versions.
  
• Can easily shared with hdas.
  
• Using relative paths inside node networks can easily duplicate both the definition wrangle and the wrangles calling the functions. This way would make the dependencies a bit more flexible.
  

BabaJ

andreuparri
In the "feature" I was suggesting the definition of the functions leaves in another wrangle instead of a vex file on disk.

Which has the same advantages as using wrangle vs files.

But what's the advantage, other than personal preference which I can't 'argue' with, of referencing it to a wrangle vs. local file in $HIP ?

To clarify I'm always thinking of local functions that are used in multiple wrangles in a single hip file. When a single wrangle with function definitions is updated, multiple wrangles calling those functions in the same hip file get updated automatically. Like a mini embedded hda but much simpler.

It does add anything new is just more convenient, faster and self-contained.

Another advantage is that you could have hdas that simple hold vex functions and loaded when you need them. Yes you can have this with included files but not as tidy imo.

For global functions to use in many diferent hip file it does make more sense to have them in a file in an include folder.

BabaJ

andreuparri
Including functions from other nodes would also allow for importing vex functions from hdas which is something I many times wished was possible.

But you're still having to define where and what you are referring too.
So the idea of doing away with 'dependancy management' is not eliminated. You still have to manage the references(if they change), although now within the hip file itself; But it's still management.

Yes, it just makes more sense to me to work with nodes (even if they have dependencies) since is just the way Houdini works.

I would personally use this very often where I need the same functionality (sometimes with small variations) in many branches but is not generic enough to make and hda.