I like the skilltree, or at least a set of packages of courses, so you know which part of the engine or path you’re looking into. Also currently I’m using the Like button to mark the courses I want to see, so probably a To watch later marking button would be the most helpful now.
I don’t know about the portfolio, I like the quizzes now and how there’s a text with a tip when you fail a question.
@Hourences I think the content that you guys should start with are the essentials that keeps the game running. Like how to set up a save system and a title screen before you start diving into any particular genres.
@Hourences
Given all the changes that happened for it - the Engine needs:
This is the most important thing as it directly affects new comers, and experts alike.
To use fast path, you have to think differently.
While you cannot really teach people how to think, You can start by pointing out that almost ALL the official samples do it incorrectly. And providing ways to do it “right”.
Also, while on this topic…
Fixing up the animations distributed along with the mannequin so they are actually correct and do not purely rely on re-target settings to work may actually be beneficial to Everyone.
This has been factually broken as shown by various tutorials since before 2015:
https://www.youtube.com/watch?v=NYFeFUxU6DI
(Meant to share the full playlist which eventually even shows the dropped shoulder error present everywhere).
If the team is interested in a proper “learner progression”, I think that maybe first things have to be scripted up so as to follow your own standards. No?
No. Then how about the standards that work providers REQUIRE ? like fast path ? 
Thats good to hear, however, are there any links to their work, their tutorials, their youtube channels etc?
Because, for example, this weekend, i needed to create a 2D array, a 3D array or a 4D array… and it was a nightmare to get info, about how to do it, because… you cannot do it, if you use blueprints, which i do.
So after a long search, i found a code, that used structs to build any higher than 1D array, and of course, i send my thanks to Mathhew Wadstein once again, because he again had an explanaintion about how to correct use structs (and especially, that you have to create them first in the editor, cuz that was, was made it impossible to import that code at first).
Sooo… i would have loved to get that info from the array documentation, because i found plenty questions, that asked exactly the same, how to get 2D or any multidimensional array.
(here, that was the code i found, and use right now, enhanced to 3D array too: Make 2d Array posted by anonymous | blueprintUE | PasteBin For Unreal Engine 4)
Wow I didn’t know this. Thanks!
Can you please (PLEASE!) link any resources for the correct optimization fast path? I don’t want to learn the wrong way! 
It’s a lot of trial and error involved, but you can start with the regular docs.
The other important thing to be aware of is to just use the animation BP for nothing at all (Which is literally opposite of all of epics resources and tutorials).
The (arguably?) best way is to store the reference to a casted animation instance (mesh > get anim instance > cast > promote to variable) on begin play within the character.
You then set variables directly from the character since there is no need to parse things twice this ends up improving performance slightly too.
The nasty bit is the fact you need to refactor logic for transitions to essentially work on a Boolean basis. You cannot check the value of a float before entering a state machine for instance, you would have to store that “boolean” result and use that directly
(at least currently? They did expand some things like !not, so perhaps there’s a bit of wiggle room for future changes there).
If you need further help, make a topic in animation with questions and I’m sure many of us on here will chime in with specifics.
Thank you!
As experienced engine user, at this point the only content I’d want to see is in-depth familiarizations with new features, modules and systems of the engine.
Backtracking to 2014, I would perhaps prefer a branchy skill-tree like tutorial series.
Set it up like a tree.
First, teach people the core basics of using the editor, making a new project, installation and all that.
From that start, they can then choose a path and each path breaks up into different paths.
Obviously, on the game path, you would make a third person character that can walk, run, jump with WASD/Mouse. From there, people can dig deeper into gameplay, maybe go first person, or explore combat, or enter the AI path, the animation path, or the sound path or the environments path. In the environments path, they can go for the lighting path, or the landscaping path.
Different paths would of course combine later on, so if you are in a team, you can focus 100% on your task, if you are an indie, you can pick and choose what is required for your goal.
And please, do consider allowing the community to discuss the content, either underneath it, or on a hub link, because some will require things in a different way, may not understand everything or have improved solutions after tweaking things.
I want to know if the video tutorial will be uploaded to Youtube afterwards. I am not very good at English. When I was on youtube, the automatic translation function was very convenient, which helped me a lot. Or does the new website have similar functions?
One of the problems with any online learning (it happens everywhere) is dispersion of information. They have hundreds of hours in live trainings with awesome useful stuff that isn’t organized. Maybe these videos can be added to the learning portal in its own category with a summary of topics
Live learning videos aren’t targeted enough to be actual learning materials - with some exceptions.
In the last year, most of those videos have largely been a waste of time where they crack non-funny jokes instead of explaining how things work - with some notable exceptions, and a lot that go the opposite way where they have guests or employees who don’t even know the thing they are showcasing or how it works.
The dispersion of information is also due to Epic axing the community wiki. Info was once organized.
Then epic decided to take over with the learn tab, so the community wiky could no longer exist, or people wouldn’t visit said learn tab.
And there you have it…
Hopefully when they add the new supposed tutorial section to the forums, some order can be restored…
Yeah, it certainly feels like they’ve been picking topics at random. But it shouldn’t be too hard to reorganize an YouTube playlist, lmao
I remember the wiki not being updated as often as the reference documentation is now, and it was also confusing as it was made by a number of different people with different styles. I think a forum is a better format for text stuff where you can reply and discuss, unless it’s official documentation
The wiky was killed in its infancy…
All they had to do was add some official technical writer to review/approve/deny changes. Like Wikipedia does it.
Anyone can write their 2 lines of BS.
And anyone can flag their 2 lines of BS as BS for removal.
Then someone paid to, goes and makes a decision.
Moving forward, hopefully, the forums will work in this simple way…
You might want to consider the start in a backwards fashion.
Instead of choosing a start towards everything, create goals instead and break them down into their individual parts.
Example: How to create a huge photo realistic open world. First, we would be introduced to the main parts and then each can be expanded into smaller segments. Each segment could be a very short tutorial, both written, with images, as well as videos. In the accompanying article, links could be provided to the documentation.
Depending on the user’s experience, one could choose where to begin to reach that final goal.
For each path, there could be underlying parts all the way down to how to create an EPIC account, download the launcher and install the engine. Plus, required underlying parts, such as how to use nanite, lumen, quixel and all that realism stuff.
For animations, there would be different underlying parts, and yet, the first underlying part would of course be total beginners; sign up, download, install and from there, enable Control Rig etc.
"The (arguably?) best way is to store the reference to a casted animation instance (mesh > get anim instance > cast > promote to variable) on begin play within the character.
You then set variables directly from the character since there is no need to parse things twice this ends up improving performance slightly too."
Wait, this is news to me. Is it thereby a bad idea to have booleans on both the character and then fetch them in the AnimBP and promote them again in the AnimBP?
Well, you are by definition doing something twice.
Is it necessary?