Blender - what is wrong?

Regarding the weirdness of Blender’s input (mouse and keyboard) mapping, it’s the horse that’s been repeatedly beaten to death over the years, then buried, then dug out and beaten some more to the point where it’s pretty much turned into dust.

Everyone knows 2.79 and earlier input mapping has been dumpster fire for most of the people, and even Blender developers know it. It’s already being fixed and there’s 99% chance it will be completely sorted out in final 2.8 release. There’s no need to complain about it anymore, simply because not only is it already acknowledged, it’s also being fixed literally as we speak.

Let me, however, point out some wrong assumptions in the beginning of your post:

  1. You assume that Pro users do not demand good looking UIs.
    False: If you look at the very same user interface every day for 8+ hours straight, you want it to look good and pleasing. If you have to sit down to some Windows 95 looking chimera of 30 different panel styles (3ds Max) every day for years, then it will push you very close to burnout syndrome.

  2. You mentioned Eevee as part of the “bells and whistles” intended to attract noobs.
    Again false: Eevee is a great help for anyone working on a realtime content, such as game assets. You no longer have to preview the assets in Cycles. You don’t have to interact with non-realtime Cycles renderer at all. You can create your realtime game asset from start to finish without doing a single render and at the same time constantly looking at pretty much the final rendered result in the viewport in real time at realtime frame rates. You can also do PBR texture painting, in realtime. It’s basically substance painter for free.

You repeat a mistake many people tend to do. You assume that if you don’t use certain feature then it’s either useless or it’s there for noobs only. While it’s easy to make such a mistake, after 18 years of professional experience in the industry, I’d expect a bit more restraint and wisdom :wink:


I am a veteran blender user, but of course I have also used many other proprietary software for forced cause needs …
and I can tell you that I share what you say about the situation on the current status of the outliner …
especially the fact that it is not functional to the contextual selection of the objects and tools available in the property panels … but I think the situation will change quickly.

Thanks for reply. Maybe I was not so precise when I was writing the topic you are referring to.
“You assume that Pro users do not demand good looking UIs.”
There is so much work to do on Blender, to become pleasure to work with it. This UI thing is the last priority on my list. If Blender would have some toxic color palette, and crazy big icons, the yes - this would be the issue. But you have pretty vast possibilities to customize UI look, and therefore I do not have any problem with it… I do not care about Blender UI. But I care about what is underneath - workflow, logic, priciples… If it fits together well, and all the necessary controls are well spreaded across the UI, and in good reach, then the only thing I need is dark color palette, and good readability of icons and fonts. Output matters… If you are digging a grave, you do not need gold shovel. You need shovel that fits you, and will help you to dig it in the shortest possible time with least energy spent.
“You mentioned Eevee as part of the “bells and whistles” intended to attract noobs.”
You said it, and you did not understood my text correctly. I wrote, that I have mixed feeling of it. EEVEE looks amazing, and it is undoubtedly great thing. I wanted point out, that developers are able to do such great thing on one side, but still there are horrific workflow flaws on the other side. This led me to think, that Blender wants to attract more beginners than pros. Otherwise, they will primarily focus on more serious issues, than those “bells and whistles”.
Is it clear now?

1 Like

Hi nokipaike

Good outliner is just a top of iceberg. Next thing is schematic view, which is tragically missing in Blender.

Just quick example:
Some time ago I was working on a commercial in which the train should transform into running robot, and then back to train.
Here is test render:

… and technical test:

This job would be impossible to manage without good schematic view, which actually looked like this:

You can clearly see the train structure, rig, and all the necessary things. All easily selectable, with great overview over individual realtionships… Doing this without schematic view would cause some serious headache.



I may be wrong, but I think this would be part of the “Everything Nodes” update that they are working on.

I was also convinced that "“Everything Nodes” it is the first step that will open the door to a better contextualization of the functions

I never saw this thread and I think I may be the one that key Blender people should read first.

I tried to give Blender a chance like 3 times in the past few years, I even gave money 2 times for the foundation, but it always ended up the same way. Giving it up and hoping it would become better later.

I personally (a pro user) will use and fund blender when:

  • the area of the interface will have tabs (one editor per panel is a total workflow bottleneck) and creating/deleting areas/tabs with be done by an icon or a right click menu
  • the axis orientation will match the industry standard (Maya and more - for modeling and rigging)
  • the viewport and the outliner will be in sync
  • the outliner selection will work the same way as any file explorer in the world… including that clicking in the void deselects… seriously
  • the basics shortcuts will match the industry standards (Q WER by example, tab for creating a new node, left click selection by default (even this obvious basic universal detail took like 5 year of discussion to be done), left click shift ctrl alt ZBrush navigation hotkeys for scultping)
  • the properties editor will not be shared for completely unrelated concepts (tools settings, scene settings, item settings are supposed to be different types of editors… consistency in the structure is important)
  • the gizmos will behave la like maya ones, and not forcing to apply the change right after the hotkey is pressed
  • the linear workflow will be default and better made
  • the render pass setup nodal (I hope the project that has the name everything nodes means -everything- literally and not -most of- so Blender can be used like Katana from Foundry)
  • …basically when it will match the behavior of the industry standards, to FIT into the (bigger and more important than Blender) software environment of what is already leading and established.
    Because this annoying manner of doing things always differently, and alone, it leads nowhere.

You say regularly the foundation funding is so low and you need more developers, but you literally reject and discourage 80 percent of the CGI community (and 99 percent of the pro one) by keeping doing the things your own (bad) way and refusing to go standard.

Yeah Autodesk sells softwares way too expensive and far from perfect, yet people/studio with big money give the money to them… Why, because they are what serious cgi artists are working with. Because when you get standard, it binds people together in a team way more efficiently.

Not being mean here, just exposing the truth guys.


I was searching for “everything nodes”, and unfortunately,I was unable to find anything more than node based particle system video. And, of course, this is something completely different. Schematic view is different kind of scene overview, compared to outliner…

Bill - maybe you’ve got some more in-depth information about everything nodes - will this solve schematic view as well?



It’s often better to go against muscle memory and relearn how to do something differently. You might not want to and that’s fine, just keep using what you’re using or work on your own version of Blender to customize it as much as you can.

Blender is not 3ds Max or Maya and while it’s good to acknowledge that it doesn’t mean Blender is going against everything. It’s not a good idea to just try molding it into industry standard. Standards don’t have to be correct and certainly in Autodesk’s case the standard is mostly one feature glued on top of the last one and everything is kind of falling apart all the time.

I think it’s kind of established by now in many fields that while Blender is going a different way, it’s seriously competitive and stating that 99% of pro users are discouraged is not correct… In fact I know many people who gradually switch over to Blender over the past 3 years and while they are annoyed by some stuff generally they are happy with how it all fits together.

Blender’s approach is pretty good and while most of the features come later, they come in a much tighter and stable package and there’s a lot that other industry standard software could learn from…

Also, first reason why studios use Autodesk SW has a lot to do with Autodesk pushing their software for free to universities which obviously leads to people being skilled in 3ds Max and establishes what people use even before they have a chance to look around and choose. They know it very well that it’s a good idea to push their SW to all universities. You like what you know but you can only really judge after you know both.

Second reason is also the age 3ds Max has been used and the level of veteranism it has. Lots of top ranking CGI people used 3ds Max since it’s infancy but that doesn’t mean it’s better today. It’s what they know and they want young people working in that software because they want to teach them and have control.

Last reason is that for decades thousands of inhouse little scripts and plugins have been done for 3ds Max by all these studios and they have all the 3D models done in it, all the materials, rigs … It’s just a big hurdle to change software for older studios.

It may have been so in the past but today it’s certainly not because 3ds Max can do something Blender can’t at least in 90% of cases. There are some specific types of work Blender doesn’t pull off that well, true.

I have a lot of respect to Discreet for what they’ve done, those were the brightest people in CG by far back then but honestly since Autodesk bought them it all started to crumble and today it seems they are unable to get it together anymore. It falls apart a bit with every new version and the classic user experience of blinking, disappearing, crashing with panel on top of panel is still there.

On a related note I was super surprised that Autodesk let Softimage’s XSI die, that was a very bad choice as that was a relatively new SW and had tremendous amounts of potential (reason for acquisition).


Just a few answers. Not necessarily to persuade you, but simply to provide information.

We considered adding this during the Code Quest. The short version is that this requires a lot of work to work well. Some apps that have this behaviour, such as Unity, allow you do drag out Editors and drop them into borders to dock them, or to just float them above the viewport as independent windows. This interface has some smaller drawbacks, like eating a bit of vertical space for each editor. We could add this in a later version.

BTW, you can right-click on editor dividers to split/join. Also, make use of Workspaces to minimize UI setup.

There’s no complete standard here. Most architectural apps, as well as 3DS MAX, use the same approach as in Blender. I haven’t heard of any project to allow for changing this - is this really a problem in practice?

We should really do this, or at lease allow an option (probably on by default) for synced selection. The current system is really confusing and counterproductive in most cases.

See Industry Compatible keymap - solves this issue.

If we would split it apart, it would mean that users would need to have more Editors open at a time which takes up more space. Not keen on that.

This isn’t the case. They don’t require you to apply the changes.

I believe Blender has a fully linear workflow. What are you specifically looking for?

Just to add a counterpoint, various DCC apps already work somewhat differently from each other. Zbrush is very different from Maya, which is quite different from Modo, which is different from Houdini.

Blender doesn’t aim to be exactly like Maya, for example. If you are looking for Blender to be free Maya, that is not the goal.

However, we have (and will) taken steps to make Blender easier to integrate in heterogeneous workflows. That’s why we use more standard wordings, and why we are adding the Industry Compatible keymap, as well as improved support for left click selection, among other things.


Thanks both for the answers,

-The tabs stuff, yes it is needed and there is a trick tor the little space drawkback, like houdini and nuke do, add a little icon on the side that allows to shrink this tab space into a few pixel line (usefull for the timeline by example).

-The axis is needed because it allows to export things to other dcc without dirty tweaks to make it work out of the box. Also Z as depth concept is stuck in many pro artists brains. This should at least an available option. Also the rigging tools should have options concerning that.

-Thanks for agreeing on that, and clicking in the void to deselect, and getting rid the the b hotkey stuff for multi-selection (shit/ctrl click instead), as well are important. This is literally universal behavior I describe here.

-That was example, not asking for Maya one necessarily, just less alien stuff, by example 1 and 3 in maya for lowpoly/subdiv is not the best good choice neither.

  • I stand for the properties editor that has to be split, this would work well in conjunction with the area tabs I told about. I really think the current design is a concession, and many people try to make is more clear like the space between vertical tabs but it definitely looks more like workaround than real clean solution.

I don’t want Blender to become a open maya project, but that Blender grab the main concepts to seduce pro users and stop doing things a different way for ultra common stuff. Look at what Houdini did between version 8 and 9, they literally said, we go Maya like UI and workflow, yet it is far from being Maya copy and everyone like using it. Clarisse did the same, and it’s popular too.

Some softwares like ZBrush work in a very different way, and almost everyone in the vfx hates this choice, even the sculpt artists, but it’s so so so much the best tool on the market that people deal with it. Doing alien stuff work when you are the only king in your domain.

And concerning AD, I know that a big part of their ‘success’ if due to hardcore lobbying, but don’t forget that to convince their user base which is number 1 by far, you need to go their way a bit.

‘It’s often better to go against muscle memory’, hum no, it’s not. Not only for learning the software, but also for switching between apps at work. It makes the switch butter smooth when you use the same conventions between softs.

Concerning the scripting, I think blender python module is a well made, that’s why I don’t moan about it.
Way better than what is available in maya ( I don’t know 3ds well enough).

If I write those lines, it’s because I have faith in Blender and I just want to share what is obviously preventing it from being the alternative that you want to be. I have many pro friends a and colleagues that share the exact same point of view, and pros don’t bother giving a chance to softwares when they are not used a lot in production + not standard.

Thanks for the hard work of what’s already working really well.

1 Like

The Outliner could use a bunch of improvements, such as:

  • Synced selection between Outliner and viewport by default
  • ability to drag multiple objects onto others to parent, not just one at a time
  • ability to hold shift and select all items between two selections
  • ability to easily throw objects in and out of the active mode
  • possibly things like special operators to rename many items at a time


Now that the Outliner is more tightly integrated for dealing with Collections. There is a lot of incentive to make these kinds of improvements.


sorry for my english
MarekHolly, Hi. I want to thank you for your opinion. You do a good job, compare the blender with other professional programs and show the Blender flaws. It is very important.
Our developers are great guys, but they are not artists (not regular artists). They need our feedback. Because a good solution for them is not always a good solution for us.

I work for freelance more than 5 years. I work together with 3d max and Blender. Lately, more often with Blender. And I have the opportunity to compare these two programs. I see the flaws of these programs.

I want to ask you about the basic selection tools and manipulator in Blender.
I made this feedback, because these are the very first and most important modeling tools:

Do you agree with this, сan you add something to this?

Hi Bill.
Great news. Concerning the rename - work smarter, not harder. As usually, XSI example of smart (m)ass rename:


Surprise - there is no rename function.

On right side, you can see outliner (explorer). Upper one, is standard object properties window. There you can rename object. But if you select multiple object, there is (naturally) no name in “name” field. Once you write-in some new name, all the selected object will get this name with appropriate postfix number.

Of course, this can be definitely pushed further - options for custom prefix, postfix, definable numbering range, etc…

I am pretty sure, I am annoying a lot of people here. But I want Blender to become beautiful. So I will keep on posting things like this, to show smart solutions. And this renaming is most definitely beautifully smart :slight_smile:



Well, you cannot easily modify properties of multiple objects in Blender without holding Alt or other things. But there’s a task for that here, only requires two small changes:


Hey that’s not fair - you put in quotation marks something I never said :stuck_out_tongue:

The above is the reason for left click select and industry compatible input.

You are not annoying, keep posting please. :slight_smile:

1 Like

Hi, I thought I would give my 2 cents here, since I’m always working with assets for games to be exported for 3rd party game engines, like Unity.
Yes it is a problem in practice, specially because the FBX exporter of Blender is not a true axis-changing solution.

Picture this scenario: You create a cube. On blender, the cube has the Z pointing up.
You export it as FBX, not changing anything.
The exporter, knowing that blender has this particular issue, allows you to reassign the axis of the cube, to the more acceptable standard of Y=up and Z=forward:

image of the export default


When you put the cube in the game engine, even though it’s visually correctly orientated, here’s what it looks like:

As you can see, The cube has a -90 degree X rotation. It doesn’t matter if you change the FBX export to +Z forward, or ±Y forward. The cube will always appear correct, but have a 90º deviation in it’s core rotation.
Funny thing is, if you put the Cube as a child of some other object, the object will have 0/0/0º correct rotations, but the child will have the 90º rotation.
Now when you put yourself in the shoes of the programmer that has to implement thousands of 3D assets on the game engine, and work with this wrong coordinates, you understand how nightmarish this can be.
Programmers on some studios that I’ve worked were usually pissed off that all of our blender assets had wrong rotations… they asked us to fix it, but how could we fix it?
The only way the Cube will go with 0/0/0° rotations to the game engine is if you:
-Rotate the cube on object mode -90º on X
-Apply the rotation
-Rotate the cube 90º on X
-So now when it goes to unity, the 90º cancels the -90º compensation and it becomes 0/0/0º.
Obviously this is not a solution for rigged characters, thousands of assets, and other scenarios.

Maybe it’s an FBX exporter problem? I know that it’s not a Unity problem because you can import the Cube on Blender itself and it will have the -90º X rotation.
What I know is that if Blender had the option of naturally having Y=up and Z=forward, the exporter wouldn’t have to compensate, and this problem wouldn’t exist.

As long as you have bake_space_transform set to true in your export settings, that issue you’re describing should exist only for armatures.
Somehow bake_space_transforms doesn’t affect them yet. Maybe that’s something the devs can take care of in the future…?

Static meshes, skinned meshes, bones, empties and so on are imported into Unity with 100% correct transforms.
Armature objects, by default, are currently imported into Unity with an x rotation of -90 instead of zero. And the first bone in the hierarchy (usually the “Root” bone of your rig) automatically compensates for that with an x rotation of 90, so at least this issue doesn’t really break anything. Animations play just fine.

Blender’s internal coordinate system is never going to change. Not a chance. That’s what the devleopers told me. It’s just too deeply rooted in the software’s code.

I’d recommend writing a simple python script that automatically takes care of the rotation of your armatures, or wait for the bake_space_transform code to be improved. Personally, this issue never bothered me, though. The two objects affected by it are usually never manipulated by game code, only their parent is, so it has basically no negative effect on your work whatsoever.


When will Blender 2.8 come out as stable industry ready version?
(My boss keeps asking and I always answer around Januar or Februar 2019)