they also seem to have this drag-out-cable behaviour. blender just did an update to their geometry nodes and also implemented drag-out-cable-release starts nodebrowser. seems to be a standard interaction now.
edit: i really like the category columns navigation in resolume. this gives a better indication where you are and enables you to switch categories much faster. the chosen categories in gamma are in a weird place… bottom right. this never made sense to me. ignoring existing (working) concepts of navigation and reading direction.
this gesture seems to make sense for environments that use connecting links via drag, but this is not the case in vvvv. in vvvv you can just start typing while having a link at the mouse.
the drag a link gesture would be a bit alien in vvvv, and i always have the feeling that it is breaking my fingers when working in an environment that uses that gesture, especially when linking across a bigger patch.
but sure, many come from being used to dragging links, i wonder whether both could work at the same time, like depending on where exectly you release the mouse… but that could also be too many roads to rome.
hehe, i’d argue that i’m very very used to vvvv’s interaction, using it since 2004, but recently discovered that the dragging out the node browser of a pin seems quite a natural way of doing things.
how would this work when you want to see node suggestions of a pin ?
yes, it would, right now, when you try to drag out a cable from a pin, you start a selection rectangle, which does not make sense
yesterday i learned a faster way to setup pads in a record, just typing the name and then choose pad. maybe list pad first when being in a record ? that would speed up things. or is there a faster way then looking for pad down the list.
my biggest concern is finding nodes, often it feels like i have the wrong strategy, lets say i want to change the circle color and i don’t remember fill. the hint would be skiapaint. but when i type this into the nodebrowser i end up with the wrong stuff. this is just one example of many. what would be a better way to look for ?
this depends on a new UI framework and moving everything to .NET6, which is a big undertaking. this is the reason why it is moved to later this year. we still need to evaluate whether the new .net MAUI (which isn’t released yet) or Avalonia is the way to got. Or even a combination of both…
we did some research with Avalonia, and we liked it because it has strong Linux support and can even run in “frame buffer” mode, without any Linux desktop system installed. that would make it perfect for kiosk apps on small machines with a basic Linux kernel, even Raspberry Pis… so you can produce hundreds of embedded systems at a low cost. that’s how the devvvvs dream…
not the patches, that is custom code developed by us, it is about the framework of “the things around it”, the connection with the operating system. it is windows forms at the moment, which isn’t cross-platform.
Unfortunately they don’t seem to have a wayland backend (yet) and X11 is on its way out. Afaik most major distributions are now using wayland by default.
That would also entail moving to Vulkan I suppose?
With a desktop system DX11 could work with wine. My latest tries getting a compiled exe to run using bottles / Wine 7.1-1 under Manjaro have failed though.
yes, vulkan is definitely the way to go. which should be an easy switch, as skia and stride already support it. but with every change of technology, there will be bugs and small changes in the feature set. thats unavoidable…
back to intelligent node filtering, this is definetely already helpful, in my example “fill” appears in the list which is really cool. but there are suggestions which simply don’t work, i’d say these should not be in the list. like shown in the gif. can’t the system check wether a connection makes sense or not and only show meaningful options ?
I have been asking myself as well recently if the node browser should also call it “Process” and not “Node” in the node browser, as it would always create a process node, right?
Or call the process in general a “node” if you want to keep that naming in the browser…
what about having the automatic node filtering automatically enabled when dragging a cable out of a pin. for the traditional patcher it would not change anything because he never drags out cables (like Tonfilm said) and for people like me, who are missing this interaction badly (muscle memory from unreal, cables and unity) it would be really helpful.
Isn’t it the case, that in all the other tools you can drag a cable but you mustn’t and also can click on a pin like in vvvv?
Also, I think the contexts of the mentioned tools differ quite a lot: for example the amount of nodes/datatypes, how low- or high-level they are in general, what and how many categories/use-cases are there, and so on. So, different tools lead to different UI and UX.
Personally, I feel arguments that are muscle memory from other tools is more an opinion than an argument. I mean, everybody has his personal preferences an personal muscle memory and I guess this venn diagram has no center.
The point here is to develop a UI at some point that is best suited to vvvv - if common ui-patterns can be used, it’s super good, but that shouldn’t be the starting point or goal for this journey.
And, last point here, a UI thing (nodebrowser) behaving differently depending on how it was invoked is a big source of confusion imho.
off-topic
Ok, one more point, but a bit off-topic:
What is the focus and what needs most devvvv priority? Is it the nifty ui details (and yes i also don’t think the current UI is perfect) that need attention? I personally would opt for other things on the roadmap like .net 6 what will also lead to changes in the ui. So, in that sense, I am totally happy about the roadmap.
nope, check for yourself. and i’m not proposing to change everything but have an alternative, additional way of starting a connection and automatically open the browser. it speeds up things, trust me :) but you don’t have to use it.
again, i’m not saying “change everything” , but you can learn nifty things from other tools and adopt principles to improve vvvv were it is suitable. and surely its an opinion and personal experience and not a theoretical paper on UI design in form of a forum post :)
agreed
and last but not least, why all the fuzz (if you are interested), in light of the “mastering vvvv gamma for teaching” course i collected all the things which i noticed during my teaching experience at university. to give feedback what went wrong, what are the things people (and me) struggle with. what devs make out of it, how they prioritize improvements is up to them. i tried to communicate all points in a constructive friendly way. and regarding the feedback, there seem to be a couple of valid points.
