hoho

Can you reuse any components from existinc PC? E.g hard drives, PSU, monitor, case, ... Also, did I understand correctly that you'll be moving to a new place in 1.5-2y and you want to get the PC after that has happened?

I haven't actually paid any attention to this thread but I was wondering if you have considered asking Forge people to add hooks there for moving complex tile entities around in the world? I'm sure lots of mods would like adding ways to move blocks around and having a decent framework for it would allow to actually do that without having each of those mods re-implement moving special blocks from every single other mod. Of course that would assume that those other mod authors actually implement that interface in their tileentities but I'm sure it'll catch on eventually.

RAM will make neglible difference in FPS. I7 vs I5 is quite similar, at least in minecraft, and boils down to clock speed basically. The price difference is definitely not worth it. Also if you are not overclocking then getting the K versions over non-k is a waste of money. Biggest performance difference will come from craphics card and for those two the latter one is roughly twice as fast. As a general principle for gaming PC you'll get a decently balanced system if you put about twice the money under GPU that you put under CPU.

If I were Notch and just toyed around with code never expecting MC to become big I probably would have gone with C++ as I simply like the code in it more and quite likely I wouldn't have provided a decent API either. If I knew it would be an instant hit I'm not so sure any more. On one hand C++ has a few nice features and given decent API you can do almost anything with it. On the other hand Java gives people access to everything and extending things is far easier even when there is no API so I'd probably go with that. Obviously in both cases I'd have it be opensource so people can help fixing and extending stuff :)

First, as has been stated numerous times, notch coded far from greatly. Second, swapping languages WILL NOT bring a major performance increase assuming both implementations are written at roughly equal quality. Although I do admit that there are some things that are easier to optimize in C++ than in Java. Object pools and non-checked array access come to mind first. Pocket/XB editions of MC show that it doesn't really take all that much of computing power to get it to work well on mediocre hardware. All it needs is decent coders behind the desk. Simply writing an engine that would generate and render MC-like worlds isn't terribly hard or time-consuming. Adding all the content to make an actual game out of it is. [edit] Note: I've worked as Java coder for 3 years and C++ for 6 and written both for fun for much longer so I believe I have some experience on the matter of the two languages Yeah, there definitely is but difference between waiting 5s vs 15 isn't really that much of a deal. Also, you rarely need to do a full recompile of your project so C++ stuff isn't anywhere near as slow as you make it out to be. I'd even dare say loading up JVM every time you want to run your stuff vs straight running of .exe makes up for the extra linking time. That also completely discards the fact that I and quite likely every other coder spends extremely tiny fraction of development time compiling or even writing code vs figuring out how to solve problems in the first place. I've had numerous days where I've generated a whole total of <10 lines of code and have considered the day extremely productive Well, for anyone doing coding for living it is ancient. Newer CPUs easily compile stuff several times faster. Add in an SSD for another huge jump. I'd dare say it's still mostly single-platform if you dump all PC os'es in same pile and if they had staid with something like Allegro or SDL for IO and visuals then it would have worked on all three with just a simple recompile unless they had done something really dumb. The things running on consoles and phones are pretty much complete rewrites of the original. Not initially but yes, after they became a thing ID quickly helped them out.

Not really, as long as you write proper Java. If one would rewrite MC with same quality C++ code as is in it's current form I wouldn't be surprised if it ended up actually being slower (JIT compiler does work quite nicely actually and it's easy to blow your legs off with bad C++ code). Using the best coding for either language C++ might pull ahead by a few percentage points, not more. Hmm, where to begin here. Yeah, compiling is slower but as long as you use anything half-decent (= anythng but MSVC) it's nowhere near as awful as you make it sound. My ancient quadcore Core2 recompiled Linux kernel in under 5 minutes and that's several orders of magnitude more code that make up MC. Compiling speed of C++ is absolutely not a problem. C++ is NOT single platform thing. Not more than Java with LWJGL that MC uses is. Unless you are complete retard you WILL NOT have to recode almost anything to get the thing running on most platforms. Modding would have been a bit harder, yes, but not impossible had there been a proper API from the start. Though even then it wouldn't have been impossible: @see original Quake mods. A proper API would have resolved this. Now that's definitely a language to use if you plan to cut your performance lower a few times. No amount of good code helps there, unless you do everything in C/C++ modules and just call something like runGame() from your python code. TileEntity ticks are bad but nowhere near as awful as the rendering system as a whole. Basically Java is good enough for most stuff (fast enough when used properly) and rather awesome for some things (runtime modification of existing stuff) so rewriting it in another language wouldn't really make much sense. Though rewriting MC is kind of needed anyway to fix the awfulness of what Notch made, it's nearly impossible to salvage it. Good thing that they are already doing that with the "soon" released modding API + renderer rewrite.

This. If you have a cell phone not older than some 5-ish years you already have a good music player. Bonus is you don't even need to get another pocket to fit it in.

Multithreading is easy assuming the code is built with that in mind and is not horrible. Minecraft fails on both accounts.

So how do you know it's better than alternatives if you have no experience with them?

I believe railcraft has blocks that re-link mine carts. You could probably use them to link your stuff back together after they pass through portals.

I think I could find use for 32G. At the moment I'm using only 8G (DDR2 is expensive) and task manager is telling me I'm using 92% of it. That's even having just 2G for minecraft and not running any virtual machines. To run them I need to start closing stuff down. If you don't care too much about other games and MC is your main priority then you should go with relatively expensive CPU and mediocre GPU. I5 or I7 with high clock speed is going to be the main thing driving the game. GPU can be anything but I wouldn't go lower than 560TI. Also, NVidia has vastly superior OpenGL support and will likely have less problems running MC.

Eloraam doesn't allow anyone to just create compatibility with her mod. If you want to do that you'll first have to be lucky enough to actually get to talk to her and then even more lucky for her to allow it.

GTX 560 here so GL4-level. I'd say ditch anything <GL 3.0, it's just not worth the effort. Unless you have tons of extra time to waste instead of doing something useful Yeah, it kind of works automatically but as long as you are not writing games using 10-year old technologies (no offscreen rendering) scaling to >1 GPUs WILL suck horribly. Unless the ONLY thing you do is crank up anti-aliasing without changing anything else. And yes, they kind of try to keep v-ram synchronized and it's physically not possible to pool either the GPU processing or RAM anyway when GPUs have vram access bandwidth in hundreds of GB's/s while buses are just a tiny fraction of that. That's why you need custom-coding to make multi GPU worth a damn.

I got into gaming at around '97. My favorite by far was X-Com: UFO defense. I must have spent thousands of hours on it over the years.

Considering there is no TMI in the modpack it's quite expected. Switch NEI from utility/recipe mode to cheating and it should work just fine. If not then screenshot of your game with inventory opened would help diagnose the problem.

It should run it but far from max settings. The not-so-awesome GPU is integrated in a not-so-awesome CPU. It'll surely be better than your old PC though.

From what I've understood the rewriting is required mainly for rendering and collision detection. Everything else should work just fine. Problem is MC way of rendering blocks and performing collision detection is, well, sucky.

New and fancy way of teleporting EU would be to combine MFSU cart with portals.

Yeah, double-transformers are quite definitely cheaper. I've mainly used batboxes with miners like that at start. Main thing it gives me is that I don't need to worry too much about keeping my generators fed since having 4+ batboxes between them and miners give me pretty decent buffer.

You can get unlimited range lossless transmission by inserting energy storage devices at certain distance. For copper a batbox after every fourth cable piece will do just fine. Only problem is it's rather expensive and cumbersome to relocate without loosing energy stored in those batboxes.

I've made maybe around 20-ish nether portals and have yet to encounter any serious problems or land on FU-isles. Guess I'm just lucky.

I guess they will add a few more wood colors and I think they enabled placing stairs in some new way. As for performance, as long as your CPU can handle the MC tick fine I think it should run better than the current version considering that rendering will be offloaded to a different core. Of course if you have a sucky CPU without two or more physical cores then it'll be slower than it is at the moment.

Anyone knows how far is buildcraft 3 SMP port?

I'm not saying unifying SSP and SMP is a bad thing, I'm just saying modders who haven't ported their stuff to MP yet will be having hard time getting their stuff work with 1.3

1.3 will not contain SSP. The "new" SSP will be a local SMP server with lipstick on it.