assign experimental/private codepoints; leave out MAY algorithms

[TJ-91]

fixes #83. this is an alternative to PR #86.

I only assign IDs to "MUST" and "SHOULD" algorithms. In cases where a referenced ID is not assigned, I write TBD (Full Algorithm Name) to keep the reference in tact.

In cases where there are multiple IDs listed and the IDs are only there for convenience and the meaning of the table is still clear, I only write "TBD" to keep the draft more readable. For example in Table 12: ML-DSA parameters and artifact lengths in octets.

[TJ-91]

We still have codepoints left, so we can just use the next one for SHAKE if it comes to that. I'd be fine with both options.

[falko-strenzke]

We still have codepoints left, so we can just use the next one for SHAKE if it comes to that. I'd be fine with both options.

Yes, I think we can worry about that later.

[wussler]

May sound silly, but what about using 105-109 instead? Most experimental algos I know of are already using 100-102 and this could be a source of conflict

[falko-strenzke]

May sound silly, but what about using 105-109 instead? Most experimental algos I know of are already using 100-102 and this could be a source of conflict

Then that makes sense I think. Clearly, none of these solutions are robust – as soon as anyone else arrives at the same solution there will be a conflict...

[TJ-91]

I changed it to 105-109. I'm not sure if it's really relevant but I also don't have any problem with the IDs starting at 105.

[dkg]

Thanks for this, @TJ-91 ! I think this is the right way forward.