Hacker News Reader

Comment by Macha | Hacker News Reader

> I think that the only hope for a good API is to actually have an API. Bukkit was the correct way to go from the start, forge is not an API.

However, most of the more interesting mods still could not possibly exist for Bukkit, due to it's server side status, so I can't see any MS API faring any better than Bukkit

View on HN

The current front runner in this niche is Sponge.

"Sponge is a combination of a new API (based off of Spout/Flow’s APIs) implemented on top of Forge, with assistance from other parts of Minecraft’s modding community (Glowstone, Cauldron, Spout, etc.).

It will be both a server and client API, and its target user base is pretty much anyone that wants to mod their game, including server owners. However, we may focus on the server-side portion first.

We invite any developer to help out."

https://www.spongepowered.org/

Development seems quite active on Github.

It might look nice from far away, but there is a vain of drama that runs through the body of the team. The team is indecisive and has that "ooh look, something shiny" mindset.

They are even making some of the VERY basic yet devastating problems that Bukkit made at the start. Again, there is more to go into than I can fit into the scope of this comment chain.

I think there are two things at play here. Firstly, I'd say that the reason Forge mods do such amazing things is because of the dedication of developers. Forge basically gives you an API to monkey patching and nothing else.

Bukkit failed because of copyright laws. Despite this, people still use bukkit. It was powerful, easy, and had an intuitive API. It made programming seem fun to me.

I guess my problem is similar to my problem with the new extensions model for Firefox. It might be much smoother for 95% of developers with the new API's. But the remaining 5% of cases go from "pretty hard" to totally impossible. With Bukkit's API, something like buildcraft's pipes with all its appeal would have been impossible. Or the only reason something like tree style tabs is going to be possible with the new fx extensions API is because it already exists so they know they need to make that possible. If Buildcraft or Tree tabs did not exist from the fully open ability for an add-on to modify their host app, I find it unlikely the APIs to support them would ever exist.

Well that is only due to the fact that there are limitations to a server-side only API.

A proper API done client side would look like bukkit, but have hooks into almost anything you could need. Look at GMod for example.

Reply on news.ycombinator.com