Re: [I18ndir] I18ndir early review of draft-ietf-dispatch-javascript-mjs-07

[John C Klensin <[e-mail]>]

--On Monday, May 11, 2020 12:55 -0400 John R Levine
<[e-mail]> wrote:

>> At this point, I'm afraid the ship has sailed. Introducing
>> yet another MIME type *now* is unlikely to gain adoption in
>> the HTML Standard, browsers, or on the web (because it
>> wouldn't be backwards compatible). Let's not publish
>> specs that intentionally conflict with both the HTML Standard
>> and web reality.
> 
> It actually would be backward compatible because of the advice
> to use Accept media types to tell when to send which media
> type, but if browsers aren't going to do it, I guess we put a
> warning in the security section and give up.

Well, just a minute.  Despite approval by DISPATCH for further
discussion and processing, this is basically an individual
submission, not a draft developed by a WG.  With the
understanding that what I'm about to say is not I18n-specific
and therefore should perhaps go on the agenda of the next
DISPATCH meeting or wait for IETF Last Call, the IETF has, as
Patrik has pointed out, both a collection of standards-track and
BCP specifications in these areas and experience with what
happens when other approaches (like reliance on name suffixes)
are used instead.  Recommending something else is not just a
matter of Security Considerations and, fwiw, I note that at
least one document has been stuck after IETF Last Call completed
in September because the IESG had DISCUSS-level concerns about
telling implementers what they should do if they chose to not
conform to the IETF's normative recommendations.   If this is
really just a description of what browser vendors are doing or
will do now and in the future, then we have a long history of
publishing "for the information of the community, this is what
some vendor or cluster of vendors are doing" descriptive
Informational documents.  They are not standards track, they do
not use normative language (2119/8174 terminology included),
they just describe what is being done and are quite explicit
about that.  

So, if the decision criteria for what belongs in this I-D is
"whatever the browser vendors are going to do whether the IETF
thinks it is a good idea or not", then let's see it recast as a
"what is being done in practice" description, with the normative
language removed or transformed into "if you want to do what the
vendors are doing, then this is what you need to do" (or even
"should" do - noting lower case).  If that means updating the
registration for text/javascript (which, I note, is in the
registry at
https://www.iana.org/assignments/media-types/media-types.xhtml#text
as "OBSOLETED in favor of application/javascript" -- exactly the
opposite of which this I-D proposes) and application/javascript
to treat them as vendor registrations (applying the exception
for non-faceted names allowed by Appendix A of RFC 6838) then so
be it... because that is, apparently, the truth of the matter.

Sorry, but it seems to me that this document is trying to have
it both (or all three) ways: to be a normative document without
meeting the higher bar the IETF usually applies for standards
track documents, to be a vendor specification without making
that clear in the text (and noting that vendor specifications
are still expected to have accurate Security Considerations
sections), and to make requirements that are at variances with
significant community experience about bad results when similar
things have been done in only slightly different areas and still
expect IETF consensus for the document.  I don't think that
works.  YMMD.

And...

>> It's not entirely harmless. The following JS program
>> behaves differently if it's preceded by a BOM:
>> 
>> # !/usr/bin/env bash
> 
> I really don't want to know when self-running shell scripts
> became part of Ecmascript.

Presumably, by an extension of the comments that caused the rant
above, whatever "part of ECMAscript" means, they are a
consideration because some or all browser vendors decided to
treat them as if they were.

    john