Split from the Pablo Dobarro Sculpt Mode thread.
Started with a discussion about Radial Array, then evolved into a discussion about Everything Nodes.
Initial link:
10 Likes
This just makes me sad. Yet another straightforward, sensible, and easy to use fix for one of modelings biggest pain points in Blender (radial arraying) will get thrown to the trash because a developer wants to push this long overdue and much needed feature further out into an imaginary future implementation utilizing particle instancing.
Perfect is the enemy of the done, but worse than that is virtually having perfect AND done at the same time literally given to you and ignoring it because it could, in one persons opinion, be a little bit better - and Iâd even argue against that.
Radial arraying belongs with the array modifier, not instancing. If thereâs a reason to instance vs. array, just add that too. Sorry for the rant but really tired of seeing solution after solution for this binned time and time again and kicked even further down the road.
26 Likes
Oh for Peteâs sake⌠I had to do like 6 radial arrays yesterday, the choice was either to dig up some addon I havenât used for a while or use an empty as a control object. So clumsy and convoluted for such a common modeling task. Thatâs just sad.
Well, I guess we just wait for modifier nodes then, as if thatâs just around the corner.
To stay on the sculpt topic though, does anyone know why the vertex colors are being put in as experimental?
9 Likes
Yeah that sucks indeed. And itâs not that we need instancing stuff for everything, I might just need a quick circular array for a temporary modelling subtask that in the end may have nothing to do with the final model. This is something weâve been asking for for ages, itâs done and ready to be initially implemented, instead weâre supposed to wait 10 more years for the new system to be completed. When the new system gets completed, maybe then we can start evaluating if array/circular array stuff can be deprecated.
7 Likes
Personally Iâd rather have a fund raiser to pay 2 or 3 developers make an addon that adds 2 or 3 different brand new radial array modifiers that copy the exact behaviour and options of good radial array systems from other software. By the time the Blender Foundation devs finally make their brand new systems that underpin their preferred radial array implementation those community solutions could be rock solid and with a code base good enough to easily be added officially to Blender.
Imagine if we could raise an obscene amount of money for this relatively simple feature. Perhaps some people at the foundation would get the feeling of âHey, thatâs a hell of a lot of money the community chose to spend elsewhere instead of donating it to the foundation! Perhaps we should take another look at the communityâs most requested small and simple features.â
8 Likes
Maybe some amount of percent of the money donated could be earmarked to community chosen features? Or when donatin there could be a priority checklist ranking to follow the money.
1 Like
I for one would be more than happy to pitch in to get someone focusing on something that I need. Even if it has to be conform with the current development plan and code review guidelines.
For example, texture painting is not even on the roadmap. Iâd rather put my subscription money into getting it done for Blender than to pay for another year.
Seeing what Pablo could do with sculpting (Not the multires. All the other stuff.), Iâm sure a lot of people would get behind a feature theyâd need. Itâs a tricky thing though, agreeing on how a feature needs to work with 100 people whoâre all investing can probably become super chaotic.
2 Likes
Seriously, developers really thought that would not take years to pass to modifier nodes.
And they have legitimate concerns about exposed patch.
About Codyâs WIP, a radial array should be demoed with an asymmetrical mesh to show orientation of copies, not with an infinitely symmetrical object like a sphere.
I would add an option to make copies pointing to center of array.
I can understand the desire of using local space, global space or view space.
But a custom object space that forces user to use an empty makes no sense.
The whole point of adding such panel is to avoid the use of an empty.
Current Object Offset is already able to handle what would be done with a custom object space.
@CodyWinch 's WIP is benefiting of work on UI.
I always found mock-ups of proposals about a radial array, horrible, making array modifier twice longer than what it was.
But now, that array values are sub-panels.
I would not bother anymore and I think that developers could have same feeling.
They have a clear roadmap. They know that we have still years to wait before modifiers nodes.
Cody tweeted that he will submit his patch for review. If it works, I donât see why it would not end-up in 2.91.
Iâm not sure why weâre talking about Array modifier in this thread, but since this is being discussed, do any of those new patches allow working with Empty elements, therefore with instances?
Well, I hope that when the nodes thing is done, Array node will be able to work with âEmptyâ and âinstancesâ. Even the possibility of Array on objects generating instances instead of copies, even if there are some limitations later.
2 Likes
No I donât think so. A modifier is taking a mesh ( or a curve), modifies it and outputs a new one, that serves as input for the next modifier in the stack. That concept doesnât work well with empties as that would rather spawn separate objects. Nothing wrong with having an array solution on object level or for instances, but that has not that much to do with modifiers.
Yes, I know there would be limitations if it worked with instances. But there would be many other cases where this would still be useful, in addition to using fewer resources. For example, Array and curve modifiers without deformation, or Array with Wave or Cast modifier without deformation object. In my opinion âAs instancesâ should be an option, and live with limitations if any. This is useful in many cases.
Yes, sure are there useful scenarios for arrays like with instances, the thing is just that incorporating that directly into a modifer would dilute the modifier stacks basic concept. What you are asking for is a different level in the hierarchy, it would no longer modify the âcontentâ of an object. Internally these solutions can share code where it makes sense, but spawning objects doesnt fit well into the modifier stack.
Iâve seen this reasoning too many times. Agree âbetter is the enemy of goodâ.
1 Like
Yep. Having better systems are something to strive for, but ignoring perfectly good solutions that work right here and now are just foolish. Imagine having that mindset with everything in life. Nothing would ever get done in a timely manner and would be a giant waste of time.
I believe thereâs a sweet spot to aim for between idealism and pragmatism⌠I mean Blenderâs a tool, not a statue
3 Likes
Those cases are covered by Instancing panel.
You just add a Simple Deform, Wave or Cast modifier on parent of object in a separated modifierâs stack like Debuk is explaining.
Having a dedicated Array object type would probably be more discoverable.
But that would not be fundamentally different.
With animation nodes, you may use a dedicated Grid node or a mesh node as Input of array.
But that is same principle : basis of arrangement is another entity than instanced object.
cool idea! I just read that thread on Right-Click Select 
What about a âBlender Bountyâ website where money is crowdsourced into the top 20 or 30 blender ideas that are actually achievable (like Google summer of code- Only a few trusted people would select the projects)
at the end of each year any project that is fully completed gets looked at by a Blender developer and if it gets put into blender official verison the âbountyâ money is released to the developers who contributed âŚwith 10-15% of money paying the Blender Foundation. Projects which are uncompleted after 2 years are removed from the âBlender Bountyâ website and all that money goes to the Blender foundation.
Like Right-Click Select users could upvote the projects they like most. The least liked 1 project of the 20-30 projects after a period of 1 year would be abandoned (or relegated like a Football league 
) And another fresh idea instead would be promoted in its place. (that would keep things interesting)
Names of website âBlender bountyâ or âBlender- league of Codeâ
People tune into football league games because they want the team they support to win. So a bit of competition might attract people to a new Website 
I am not a developer -so Iâm just putting random ideas out there - just ignore this if its useless 
3 Likes
Sounds like a neat idea, but I donât think serious developers would work like that unless theyâre payed for milestones and the usual sprints.
Projects like summer of code would be cool, Iâd be happy to pitch in as well. Itâs about finding a project youâre keen to see making it into Blender and backing it. The community is big enough I think, itâs just the management/review part thatâd introduce serious overheads.
I feel like the sculpt thread is getting somewhat derailed though, might be worth to pick this up in a separate thread, that is if people think itâd make sense to do so. The idea is good I think, but itâd probably need to be discussed with the BF, realizing it might be more complex, itâs better done right if itâs feasible at all.
3 Likes
So weâre about sixteen posts into offtopic-land. Come on guys, it literally reads âsculpting and paintingâ in the threadtitle.
Listen to the man. Oh and by the way, there have been several such threads around here already. I can tell you now in advance: It wonât lead to anything meaningful.
greetings, Kologe