From an indie mobile dev used to work with Ogre3D :
I am a super hero :
Extremely fast hands on the editor (level, BP, material, persona…) when starting from ground, i.e. with no previous UDK experience, with existing doc and tutorials. From zero to super hero feelings, the path seems short and easy.
The workflow is really really good! From CG model creation to C++ interface implementation and finally in-editor level setup, it’s really fast and flexible.
C++/BP interaction is really amazing!
Different editor parts interaction/integration with BP is really great.
AnswerHub is a great platform for exchange.
I am NOT a super hero :
A “corner-stone” is really missing in documentation for C++ dev. From day-one “learning UE4 with our first C++ project” to “getting the big picture” i.e. how UBT works, how UE4 C++ is really define under-the-hood and use it a right and safe way (UClass good pratices…), how the different modules (Core, GameFramework,…) and low-level Engine parts (game loop, RHI…) are layered and interact together, how setup .ini files and include module… there’s a really huge gap filled with … nothing! Of course everyhting is in the source, but without hints…
From indie POV, Blender missing support, like Clothing feature is really frustrating!
Yep Bsp editor is great. This is one reason why I chose Unreal 4. If so dumb and lazy artist as I can with the help of bluprint do a serious project, it means that bluprint is well done and does not require any improvements. Now I am without the help of a programmer can do ALL …
While it may be necessary to add wires ability to edit transparency and the ability by clicking on a function or variable the wire with which it connected to other variables became clearly visible, the other wires would be barely visible. Because sometimes I create a complex, large function that can not be divided into simpler functions. And inside of this function is awful a convoluted tangle of wires.
P.S. sorry for my english.
Hi Epic!
I especially love that you are working on Behavior Tree system, the AI module, navmesh system, PawnSensing module, PawnMovement, other AI modules …
I do not like that this is a slow process, but I understand that the AI development team is not large and therefore requires a lot of time for development.
I wait patiently improvements.
It would be desirable to increase the AI development team and development process would be faster.
I love your Blueprint, this is a very comfortable visual scripting system.
Also love your sensitivity to the community, you are open to communication, it is a huge plus.
I was recently asked to fix one problem and you fixed it in the new build, it’s gorgeous! I love you Epic, keep up the good work.
Special thanks to the AI development team and thanks you Epic, you are the best of the best. UE4 my favorite game engine. Sorry for my bad English. With love from Belarus.
I think you succeeded with this. I found Unreal easier to get into than Unity, for example, even though it’s probably got more functionality. I like the content that comes bundled with Unreal, too. It’s nice that you can try the big things without wasting time putting together a basic material, for example. Even when you decide you need a material that isn’t bundled, it’s easier if you have something you can start with.
Perhaps this could be taken further. Lots of games have common props like trees, foliage, furniture, crates. (I hardly ever see a crate in real life, but they’re everywhere in games!) If these could be bundled with the engine, or available for a small charge—as seems to be planned with the ‘realistic weapon pack’—I think it could save developers a lot of time.
It would be nice if there was an easier way of moving assets between projects. If you want to bring sample content into a project you’re working on, it seems to be rather involved. Unless I’ve misunderstood, you have to open the example project and export the content, rather than having an import function that you can use from your own project.
…
I also feel that there needs to be something between Blueprints and C++. I found the visual coding of Blueprints better than I expected, but I feel it works best as a mechanism for directing high level functionality. I’ve just coded a system that moves the player’s fingertip to a socket on an enemy’s back (so the player does a tap on the shoulder, effectively). At the moment it’s a blueprint, and it’s really funny seeing formulae expressed as graphical boxes. On the other hand, I’m still learning the Unreal API, and it’s handy that Blueprints gives me an easy way of searching for the function I want. If I went to C++ I would have the opposite. I would have conventionally written formulae but I would be scratching around looking for the right API calls to make.
I know the Java EE API quite well, and that is even bigger than Unreal’s. The thing that makes that bearable, though, is the IDE support. Even though Java EE uses a text-based programming language, you can search for API calls. The Java IDEs are very good, and being strongly typed, Java allows the IDE to make good guesses about the user’s intention. You can’t do that to the same extent with C++ (or Javascript, or Lua) because the type information is not there at compile time.
So my suggestion is that Unreal would benefit from Java or C# along with C++. What would be perfect is a function that compiles a blueprint to that language, so you can start by experimenting with blueprints, and then change to conventional code when you’re ready.
Incidentally, it’s common to assume that C++ is fast, and languages like Java and C# are much slower. In fact, my tests suggest that they’re roughly comparable. JITs have got a lot better, and they have some optimisation opportunities that are not available to ahead-of-time compilers: they can tailor code to the exact CPU which is in use; and they know which methods have been overridden, so they can avoid generating unnecessary indirect calls.
…
I’m not at the stage of writing HUDs yet, but as far as I can see, the support is quite basic at the moment. I was wondering how I could approach writing a HUD, and I had an idea. I never got into Scaleform because I never got into Flash generally, but it strikes me that the basic approach is a good one. So how about using HTML5 for HUD design? You could design your HUD using something like Inkscape, then script the resulting SVG using Javascript. Many people will already have the relevant skills because they will have worked in web development.
If I get to the point of writing a HUD and Unreal isn’t offering anything tempting, I’ll probably use this: Google Code Archive - Long-term storage for Google Code Project Hosting. . It will render a web page to an image, so I think it should be possible to make a special kind of Unreal texture which is actually a web page! With that I should be able to implement my menus and HUD in a very straightforward way.
Again I think I should get away with it speed-wise. Scaleform works okay, and rendering Flash must have a similar cost to rendering HTML and SVG.
I wasn’t joking, but clearly what I said was incorrect.
I was trying to say that IDEs tend to give better suggestions, more quickly, for languages like Java and C#. Perhaps this is because generating the suggestions is less work. Having said that, I’m wondering if I’d say the same if I’d spent as much time developing in C++ as I have in the other languages. It’s possible that I’m just biased in favour of the familiar. (I learnt programming in the early '90s, so I started out on C and C++, but it’s a long time since I’ve used them extensively.)
Is there anyone who has experience of both—perhaps doing projects in both Unity and Unreal? If so I’d be interested to know your thoughts.
Of course this is always very difficult because different benchmarks say different things. The benchmarks I looked at were these: http://benchmarksgame.alioth.debian.org/u64q/benchmark.php?test=all&lang=gpp&lang2=java&data=u64q . I don’t know if it’s been changed, but a while ago I found that these benchmarks tended to understate the performance of JIT-compiled languages, because the tests were too short. When I increased the length of the test, the gap narrowed, though didn’t disappear altogether in the case of Java, IIRC.
@anonymous_user_f724fc9b, believe me, comparing java to C++ is like comparing bike to airplane. Both of them will get you from A to B but one will get you much faster and much more comfortable. I do not want to take part in yet another “language war”, so please give yourself and me a favor. Try to use C++ (modern C++, that is C++11 and onward) for three months. You will never look back. And the speed will blow you away compared to java and C#.
But there is more to C++ than just raw power and speed. There is a beauty of expressing yourself as no other language allows you to do so. And the freedom. Believe me, you’ll never look back.
THE LOOK! Finally Global illumination, the Materials, it all looks just great! Working for architects this is the first time I see a real time renderer that can actually compete with the common render-engines like vray etc…
THE INTERFACE: For a Game-Engine Noob like me it was very easy to get into, having absolutely NO programming skills.
THE COMMUNITY: For most questions, there will be a reason, people are helping each other, great.
THE PRICING: Finally a up to date Game-Engine that is affordable!
What could be improved:
IMPORT: More Formats (like Obj. or 3DS etc.) Support or plugins for other 3D Programs like Cinema 4D, Rhino, etc. Especially for UV’s and Smoothing groups.
MAPPING: Possibility to easily change mapping of materials in Engine (Cubic, Cylindric etc.)
MESH: Possibility to make simple changes on importet meshes (moving points, edges, faces etc,)
PROJECTS ON SERVERS: Support for migrating objects to a project that is not saved on harddisk, but on a server or an external HD.
I just wanted to take a moment to say thank you for all of the solid feedback! It’s been fantastic to see what folks who have been enjoying, and it’s equally as nice to get more info on what everyone feels still needs love and attention. We have a long road to travel, so don’t be shy about telling us what you feel isn’t where it should be. =)
A person or 2 in this thread has already mentioned it, but don’t forget that our UE4 roadmap is available for viewing and voting : Trello
If you have a chance, please take some time to make sure the improvements/features you are looking to see are getting voted for if they are listed.
Lastly, I know quite a few other folks at Epic are reading responses, and are as appreciative as I am that you are taking time to talk with us. Please keep the feedback coming! And please keep the WIP and finished project posts coming as well. Seeing the all of the cool stuff people are able to create is a big part of what keeps our batteries charged and helps us push forward.
Thank you all again for being an awesome community!
Interface, congratulations to all, able to turn a great tool, something easy to use, this is a masterful job.
Frequent updates
Ease of use
Almost all items have explanatory tooltips (it is perfect, it is faster to read the tooltip, than keep looking in the documentation.)
Points to be worked
Documentation
UX, interface and interaction are perfect, plus I add this point, so that the work in this area is kept, the interface is clean and straightforward, this is excellent. Main functions, have large and prominent, easy to find buttons. Very nice. But I think some areas need attention such as:
Cascade
Matinee
Both visual consistency, as associated with the shortcut keys in the various editors. Check this post.
I think the tooltips, could have a link or some way to be pointed to the documentation (so, who want deeper explanations, be able to find it easily).
Some way could be notified when a post is quoted (allows us to know more easily when someone answers a question).
The documentation is constantly growing and evolving. We know there is still a lot of ground to cover, and we aren’t going to be slowing down anytime soon.
As for Cascade and Matinee, it’s perfect that you should mention them specifically, because they are on our radar for significant improvements in the not-too-distant future.
