modding API might be server side only?

[Wooden_axe]

http://forums.somethingawful.com/showthread.php?threadid=3430195&pagenumber=221#post402411176

I might be more pessimistic than most, but he constantly calls them plugins, and says flat-out that if the modding API isn't good enough you can always mod the old-fashioned way.

I don't know about you, but that really seems like they aren't paying attention to client mods at all. Also doesn't help that a lot of the rumors and 'inside info (if you can even call it that) has said that they have no interest in client mods, they never have.

I know Technicpack isn't doomed, but should we get our pitchforks and torches?

I really really hope they don't turn into another Flowerchild piece of work where the only "mods" are server side and anything else would break it, that would doom technic.

[Jay?]

One of the plans for the future is to merge single and multiplayer, so it kinda makes sense.

[Arkalius]

Yeah I think they want to make SSP work on a localized client/server model to match the way SMP works to make modding (and everything else related to developing MC) simpler.

[Wooden_axe]

One of the plans for the future is to merge single and multiplayer, so it kinda makes sense.

ok, so that explains alot. but I'm still holding onto my pitch fork and torch

[Jay?]

ok, so that explains alot. but I'm still holding onto my pitch fork and torch

As you should, Those things come in handy a lot these days.

[Ilan321]

Wait.. if they intend on merging SMP and SSP:

1) Cacti is too retarded to jump on in SMP

2) Entity lag?

3) Is it possible to have internet lag from your own host?!?

4) bleh

[Jay?]

1) i have no clue what you are talking about.

2) unlikely. Entity lag is a result of the lag from data traveling over the internet. If the game is taking data from localhost, there shouldn't be any lag.

3)not unless it's coded by a five year old.

4)i disagree. I think it's exciting, since mods will most likely be both single and multiplayer by default with that setup.

[Ilan321]

I played a map called "t3c's Parkour map"

and he has a challenge with cacti, that you need to jump on them to get past it, but he says SMP isn't recommended because it's laggy. (lol)

Also, SMP has more bugs than the universe is big in cm^2

[jakj]

Believe it or not, there is actually some minor lag (even on excellent-quality computers) when running a localized server with nobody but you playing. Yes, Minecraft is coded -that- poorly. Hopefully this is no longer an issue with their Grand Design.

(Likely it has something to do with the game normally being so atomic on the tick, and that not being the case even on a super-duper-fast local server.)

[Arkalius]

Believe it or not, there is actually some minor lag (even on excellent-quality computers) when running a localized server with nobody but you playing. Yes, Minecraft is coded -that- poorly. Hopefully this is no longer an issue with their Grand Design.

In this situation, you still have a client process communicating with a server process through a networking layer. With what they are probably going for in the future SSP verison is a system where the client and server run in the same process and communicate directly without a network layer in between.

At any rate, once they do go to this model it will hopefully light a fire under them to get more of the bugs and quirks in SMP fixed up since it will then become the only way to play.

[Nphyx]

I think you'll actually see a net increase in performance (or at least perceived performance) for singleplayer gameplay on any modern machine because world-generation, NPC AI, physics, etc. will all end up being handled on a separate process (or thread, depending on how they do it) from the client concerns of actually displaying the game world and providing the user interface, meaning that it will naturally take advantage of additional cores. Imagine no more LSOD, no more framerates being cut in half during chunk generation, etc. without the use of Optifine. If you're on a multicore machine your experience may be similar to the performance gain you experience playing multiplayer vs. single player, but without the latency issues. I seriously doubt you'll notice the fractions of milliseconds of latency introduced by inter-process/inter-thread communication

Not to mention the fact that, even if the mod API ends up being insufficient for some of the more complicated mods, unification of the client+server code is going to mean all mods are automatically multiplayer (even if users and servers still have to manually patch them in). Even the additional complications / learning curve that may be involved in writing for the server are going to be much less than the general pain in the ass of porting code that prevents many awesome mods from seeing SMP releases.

At least from my perspective (speaking as a programmer but not a mod author), there are few things I loathe more than writing code once, then having to make a ton of little pissant changes to run on a different version/platform, and then doing that over and over again every time my code or any of those versions/platforms changes.

Also, I think critics of the potential of a mod API are overlooking the vast untapped potential for itty bitty mods that an API will make possible. Right now making a mod that just adds a couple decorative blocks or a single mob (or handful of mobs) or a single ore or new piece of equipment is a big waste of time, because it's way too much trouble to install mods on the user side to bother doing so for a million little tiny ones. With an API it will be come totally possible (and even probable) for third party build tools to allow artists, animators and other non-programmers to add new things to the game without learning Java, let alone the complexities of dicking around with decompiled obfuscated bytecode.

Imagine an editor GUI that allows you to plug in a texture image and tweak some properties in drop-down menus to create a single new block and then export that to a minecraft plugin, or a similar GUI for adding a mob or a weapon or a piece of armor. You can then stick that plugin on your server and automatically all your players get to use it. This could turn Minecraft into not just a game but a platform for making virtual worlds, in a way that is currently restrictively difficult to do. Those of you old enough to remember MUDs and MUSHes may have an idea of what I'm envisioning here. It's something an entire generation of gamers has largely missed out on, and I'm really excited to see it finally coming back.

[jakj]

Shit son, I already wrote the non-gui part of some of that. That's easy. If you're a gui programmer and know Java, we can get together and do what you just said already. Nobody needs Mojang anymore.

[Nphyx]

The key part is practical distribution without requiring end-users to dick around with patching jar files (or even using the really slick patchers out there). I'll be frank, I'm a highly technical person and even *I* surf away from any server that says I have to go out and download mods X,Y, and Z to use it. That's a real shame, but I just don't have the time to mess around with that kind of stuff, let alone the logistics of doing it separately for a dozen different servers. Spout claims it will eventually make this possible and if it does, I agree, who needs Mojang; but if the newly added Bukkit guys get the job done first even better since we won't even have to ask people to install spout

I hate Java (it's just so yucky), but if the Mod API manages to solve the distribution problem I'll definitely write some mini-mods. I don't do Swing, I do web applications, but I definitely do GUIs and might end up writing some web apps along these lines. We'll see what happens.

I see Minecraft becoming a huge ecosystem like the old MUD bases used to be, and I won't be surprised if MC derivatives are still around 10 years from now (hell, some of the old MUDs are *still* online). Whether Mojang or someone else ultimately makes it happen is up in the air. The community, as well as the necessary attitude regarding legal stuff on Mojang's end (if not an explicit open source license :/) is all in place.