We are. MIME types in the `m` tags and Nostr-specific classification in the `M` tags.
Discussion
Examples:
I agree with nostr:nprofile1qyghwumn8ghj7mn0wd68ytnvv9hxgtcpzemhxue69uhhyetpd3ujumtvv44h2tnyv4mz7qpqfjqqy4a93z5zsjwsfxqhc2764kvykfdyttvldkkkdera8dr78vhsfa5ckl, implementing numeric kinds was a design misstep in the Nostr protocol. Numeric identifiers are inherently more difficult to memorize, standardize, and use correctly, also they create confusion.
A better approach would have been to implement standardized mime types or namespaced types instead of numeric kinds, which would have created a more intuitive and self-organizing system.
while I absolutely think you should use the `m` tag with your use case, I also think that we're early enough in nostr that we can still add mime-types as a replacement for kinds.
I actually like the kinds, as they're an easy way to find the relevant spec for some particular implementation (just search for all of them listing the kind number), and they allow numerous similar implementations to occupy the same space, without having to agree on a spec.
But the M space they occupy should be defined. Also, kind numbers only need to be unique _within a particular M space_, which means the devs working within that space need to keep a list of the kinds. Nobody needs to keep a full list.
