Peter Alexander
Peter Alexander
This makes sense, but I'm trying to think of scenarios where this would change the outcome. The only one I can think of is where the client e.g. support v2...
> In that vein, will extensions be binary only (present / not present), or can they fine-grained a la current MCP capabilities? For example, could an extension be declared with...
> Those extensions will have their unique URIs, but the client has no way of knowing what those extensions are. I assume this is about _humans_ looking at a server...
@tulshi Yes. Will clarify.
@tadasant that's a good call-out. I don't have a strong opinion except that we should try to be consistent. I don't know widely the implied `_meta` naming conventions are already...
@patwhite Yes I saw that doc, not opposed to the idea, but the purpose and scope of this SEP is just to try and formalize the existence of extensions and...
Oh, just noticed that some (but not all) pages already have a e.g. `**Protocol Revision**: 2024-11-05` box at the top. Will remove those.
I think @dsp-ant was hoping this could just be a PR, but I'm happy either way. I'm slightly leaning towards a SEP as well to get some more discussion, but...
I think that's reasonable. I kind of wish we just got rid of URI so that we don't have this problem :)