Hi Peter,

> On 28 Aug 2026, at 08:42, Peter Bex <xxxxxx@more-magic.net> wrote:
>
> On Tue, Jun 16, 2026 at 03:30:46PM +0200, Peter Bex wrote:
>> Hi again,
>>
>> I don't *really* want to bring this up but I think it's important to
>> at least acknowledge that the RFCs are officially deprecated in favor of
>> the WHATWG "living standard" document.
>>
> [...]
>> Anyway, enough about my personal grievances with WHATWG.  My point is
>> that this is the spec the browsers use and some modern URI parser libs
>> follow (at least partly), so the SRFI should probably mention how it
>> relates to that spec, even if just repeating the problems I see with it.
>
> The current draft still doesn't mention the WHATWG spec.  I think it'd
> be good to define how/why it relates to that spec and the RFC, which is
> considered deprecated by many.
>
> Cheers,
> Peter

Thanks for that. I wasn't really aware of the WHATWG before your email a couple of weeks ago.

Anyway here's my current draft text, to be put immediatey after the rationale section:

"RFC 3986 and 3987 (2005) are the latest standards proposed by the IETF, but the WHATWG also publish the URL Living Standard [14]. The most dramatic change is to eschew the formal grammar defined in RFC 3986 in favour of a long-form text description of a parsing strategy. This arguably makes it difficult for implementors to select a processing strategy most specific to their implementation (data structures, parsing techniques). It is difficult to understand how RFC 3986 and the WHATWG specification diverge, and the WHATWG approach represents much more of a moving target than the IETF processes of deprecation. Finally, there is no successor proposal for IRIs, which remain current in recent standards such as JSON-LD [15]. These issues are discussed here [16], and the proliferation of different, competing standards are discussed here [17]."

[16] is your blog post, [17 is:
https://datatracker.ietf.org/doc/html/draft-ruby-url-problem-01

Emphasis is primarily on it being cumbersome for implementors (this is a sort of technical report), hence reference to [17] and explicitly highliting WHATWG is a moving target.

I know you mention security issues with respect to parsing, but I wasn't sure how to paraphrase you fairly.

From my reading so far,[17] points to a couple of documents particularly for IRI which may clarify issues in RFC 3986/3987. So it seems like there might be parts which could help with the current SRFI proposal.

Thanks,
Duncan