Update WIT and binary import/export syntax

[alexcrichton]

This is an update to WIT.md and a minor update to Binary.md to reflect the changes discussed in #193. Lots more discussion is in the issue, and a summary of the changes is:

• WIT use statements are almost entirely replaced. Old pkg and self-style paths are all gone.
• WIT documents are gone as a concept, a package is a collection of interfaces/worlds now.
• The externname binary format production has been updated to either be a kebab-name or an "ID". The URL field has been dropped.
• Imports and exports of interfaces in WIT worlds is now done through IDs instead of kebab-names.

[alexcrichton]

I'll also note that I'm in the process of implementing these changes. Current work is:

• https://github.com/bytecodealliance/wasm-tools/pull/1027
• https://github.com/bytecodealliance/wasmtime/pull/6390

[guybedford]

With regards to the ESM integration, it would be great to handle this in this PR as well since those behaviours do need to be worked out as part of this. My suggestion would be a note along the lines of the following bullet points:

• When importing components in the ESM integration, ESM-integration-compatible components are those that do not use namespaces but instead use kebab names in the outer (exposed to the esm integration) import and export positions, or optionally URLs in the outer import position.
• When converting components into esm-integration-compatible components, namespaced exports may be provided by convention as interfaces nested under the namespace and package interface names. In addition, if there are no conflicts against other exports, the interface being exported may be treated as a direct export of its own name in this convention.

[alexcrichton]

Ok I'm going to go ahead and merge this, if at least to have the current spec reflect what's implemented now in tooling. I talked with @lukewagner and he'll update Explainer.md shortly afterwards and I think any further minor updates can happen in subsequent PRs. Thanks all for the discussion here!

[guybedford]

@alexcrichton if I had known that the ESM integration upgrade wasn't going to be made in this PR I likely wouldn't have approved it. There are not "further minor updates" but fundamental details of the model that we cannot ignore.

[lukewagner]

ESM-integration updates aren't "minor", but I do think this and other language integrations questions can be done as follow-ups without changing what's in the AST and binary, so it's a good idea to merge now and iterate forward. I'll be working to update some other parts of Explainer.md shortly.