Fake user is part of the automatic garbage collector mechanism.
Garbage collector is an integral part, since blender can pack external data.
Fake user is part of the automatic garbage collector mechanism.
Great. Thank you for enlightening me. I’m so stupid.
Do we really need this thread up in flames, already?
Fake user is what should go to the garbage
No, it works in opposite - fake users datablocks do not go to the garbage.
I agree, they need to be exposed. The new trend of hiding important icons is not reallly convenient. Such as in the modifiers, apply icon is unexposed. Applying was the most common thing we do :((
Apply and Duplicate for modifiers have shortcuts, so you don’t need to go to menu OR aim for the button. Simply hover over the modifier and hit the key combo. I guess it could be annoying if you don’t have keyboard handy under the off hand though.
Until you are told that you need to hover over the modifier during learning software.
ok. i believe you only use blender, so you’re not aware of whats available out there, therefore you might think that is impossible to live without this archaic fake user system. and for that i forgive you
and just so you know, even the devs at some point not so long ago considered to remove this fake user trap-system entirely from blender. too bad i cant find the thread, it was fun
it’s a shame they didn’t killed it when they had a chance
No need to speak for everyone.
A fake user is a very handy thing. Don’t delete it without thinking a replacement.
Allows you to keep the necessary information in the library file without duplicating geometry and a lot of unnecessary.
Why does new blender change so often make it worse? We’ve broken a great wireframe, now we got to the user’s flick.
It is very noble of you that you did not force them to do this.
Otherwise we would be forced to suffer purging trash manually every time, like in other software, where is no ability even to pack external data into file, which we would like to avoid…
Can you further clarify what you mean by it being part of the Automatic Garbage Collector mechanism?
Does this have to do with blender’s memory management (I’m not a computer scientist) or do you simply mean that “garbage” data is being collected instead of deleted?.
Here’s a link to the task for saving some datablocks by default, for those who are interested: https://developer.blender.org/T61209
For now, I think the fake user icon should be always visible as it’s now. Or at least when the datablock has no real users. Otherwise it becomes even easier to lose data, as mentioned already.
The problem is not that Blender deletes datablocks that aren’t in use any more, per se. The problem is that there are datablocks where the user expects the data to reside in memory, logically. I think nobody here argues that datablocks that are really unnecessary should be discarded to keep the files smaller and more preformant. The problem is in how Blender handles and communicated the datablocks and deletion thereof.
For example: If a user deletes a mesh then they’d usually expect the geometry data to be gone as well.
With animation clips and materials this is a completely different story, though. These are datablocks which work in conjunction with visible data but not on their own like a mesh item. It’s sometheing a user often expects to create a library which resides in the current file to apply onto an existing object, rig etc.
An object which is deleted usually signals (logically) that the data associated with it isn’t needed. Creating an animation library on the other and usually rather means to switch something or have iterations or variations being applied.
Especially in cases like exporting an object with multiple takes for a game engine or creating NLA takes, using different materials for different render passes or different engine or pipeline exports …
If a vertexmap isn’t used by anthing specific at the moment the user wouldn’t expect it to be deleted on exit, either.
And since Blender does do datablock management with more control for the user there should at leat be a way to set the default for the way the user likes to work or a way to manage large amounts of datablocks.
Better feedback, options to set the default behaviour for datablock types: delete meshes on exit but keep materials and animation clips by default … something like this in the options, maybe.
(edit - clarification)
Make Single User Copy and pressing the duplicate icon would do the same for data-block selectors. Neither Brecht nor Bastien could see a reason to keep the former if duplicate is available, so it was removed (actually just disabled for now).
There should be a default duplication operator though, so it’s available for all data-block selectors.
It’s been proven over and over again that branches do not get sufficient testing for representative feedback. Only people who actually want a design change will test the branch. Experience tells us, the only reliable way to get less biased testing/feedback is making things default in master. Designs can be iterated on based on feedback - that is why we create these feedback threads.
Blender 2.92 is in an alpha phase. Using it won’t always be the most pleasuring experience. Designs have to go through phases of evaluations and iterations until they are ready. The release cycle is built in a way that allows that.
Oh yes, agreed that they are redundant.
I really miss the number of users.
The workflow behind garbage collector is easy - during work a huge amount of temporal data condenses into files, and garbage collector removes unused data that was not marked by user as important (as a fake user).
As practice has shown, it is much easier to mark useful blocks among the garbage than vice versa.
This allow to keep files clean.
Well, there is an issue - there is no time given for proper redesign and forming proper feedback.
This problem became especially acute under William, when almost 50 completely different topics were created and pushed almost simultaneously.
So you open the software, you find a ton of problems, and you have to simultaneously form opinion and redesign, for example, vertices selection, mesh display, background images behavior, normals detection, statistics display, collections management and outliner design, otherwise the train will leave.
I am project manager and workflow designer, so my job is to make sure that software we use will be compatible with unique complexity workflows our company perform, and this is too unbearable burden for my health.
Designs have to go through phases of evaluations and iterations until they are ready, otherwise it is a high chance to get critical workflow issues, but for the past two years, I just don’t know what problems I will have to solve the next morning.
I don’t know yet how to keep a sustainable business this way.
Uncertainty is a very big problem in the current development of the 2.8x branch.
Blender is redesigned so hard, like it was completely impossible to use it in any way, which is far from being true, since it contains a lot of unique perfectly balanced UI UX solutions.
An attempt to sit on two chairs, or maybe three or more.