No announcement yet.

Why it is so difficult to really get started with UDK

  • Filter
  • Time
  • Show
Clear All
new posts

    Originally posted by Mark Rein View Post
    If you could send me some info of what bits are out of date I'd appreciate that. Email it to mrein at (remove space and change "at" to @)
    It was mentioned in another thread, and that it was being taken care of at that time, and that it'd take a day or two to propagate. No need to bother the big guy, but the prompt attention is awesome!


      Let me share my own recent experience that reads like a summation of this thread:

      I'm in a senior-level (4th year uni) "Practical Game Dev" class, wherein the focus is mostly on AI implementation in games. We have a group project to finish by the end of the semester (December). One of the first considerations made was as to the engine we'd use.

      Someone on the first day suggested UDK, and over last weekend I decided to dive in (as I'm sure some here can attest to, given how I bombarded the Tips forum with questions ). I immediately understood where Alan is coming from in the OP.

      I've rolled my own engines, worked with engines from c++ to actionScript, etc, and this one is different. The tutorials seemed all focused on level editing, Kismet, etc. It was frustrating.

      Today in class, someone was complaining about this and said that we should just use a 2d engine, so as not to get bogged down learning a complex engine.

      That's ****. If you want to make video games, or work in the world of CS at all, you'd better get used to crawling through codebases. There are enough basic tutorials here to show you syntax. Once you've learned that, open UnCodeX and get to studying.

      The point is, once you've overcome the sharp learning curve, the world is your oyster. There's a fully fleshed-out physics engine, camera system, network code, the list goes on. And it's enterprise level.

      I guess the tl;dr is this: it's hard for a reason. If you have the determination to overcome the learning curve, you end up with a ridiculously powerful tool for rapid prototyping and development of release-worthy games.

      You could say I dig UDK.


        And then there is also this yet unconfirmed observation Elliot was kind enough to point us toward on the Wiki yesterday:

        I personally doubt that you could program efficiently with such a system or that things will take at least twice as long as with conventional methods and that most employers would probably always prefer an "oldschool" coder to someone who has to rely on such a thing, but maybe it can help to make some coding a bit more transparent to newcomers who just want to take their first steps in the UDK and make a little game for testing. But in the end you will probably end up raisin your hat to the Deutsch Limit.


          Originally posted by Crusha K. Rool View Post
          And then there is also this yet unconfirmed observation Elliot was kind enough to point us toward on the Wiki yesterday:
          There's probably no such thing as Kismet 2. The k2 classes are for visual programming inside the AI Tree feature they're developing. (right click in content browser & create a new AI Tree to check out the editor). Also, AITree.uc extends from K2GraphBase, which also confirms what the k2 classes are for.


            Hmm, seems like an major misestimation then…


              UDK is a good engine for anyone who wants to start game dev with. It has a learning curve, yes, but it also has as much tutorials to get you over it. The real test is time and patience. Do you have the patience to learn how to use this engine properly? Before I started the scripting, I made entire levels using kismet. Then, I started looking at anims, sound, materials, etc. If you are willing to learn, then this engine can be very beneficial!!


                My two cents...

                Having a programming background in OOP, it generally was easy for me to pick up unrealscript.. I understood how to look through the example engine/core classes and to figure out how to build based off of those... and I went from there. Having a strong base in OOP previous to learning Unrealscript helps a lot.. as im sure others have said.

                One thing I was annoyed at was finding some quick samples of how to do something with unrealscript, when the only outcome was people using Kistmet.. which is great for what its ment for.... but frustrating when you cant use kismet.

                Another tough learning curve for me was figuring out how the engine handles the objects and how to interact with other objects and classes.. but generally I picked it up in 6hours.


                  I would say game making is hard on geenral. Even if some program are just more intuitive or more easy to understand in th eother and they offer less power or less control. For instance in modeling. Maya is hard for some reason, harder than sketchup but I can model anything with sketchup because it lack of controls. UDK is not hard if you consider that you can create a AAA game with and this is a tool use by team made of several professionals. You are not one that's the reason it's hard. If it was easy you'd seen more game made with the udk already.

                  If you have no experience in game development you are starting from ground up so basically don't expect to do any really impressive and playable demo before at least 6 months.


                    maybe epic can put something similar to xna's starter kits into future xdk releases? have a starter kit for rts, platformer, racing, shooter, etc.

                    that way, folks who are interested in a certain genre can build upon a specific starter kit framework and not spend too much time figuring out how do i change a fps based engine into a platformer based game?

                    it would ease new folks into the udk-way of things. :P