[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lmi] [lmi-commits] master 3da2e92 5/5: Follow a w3.org recommendati
|
From: |
Vadim Zeitlin |
|
Subject: |
Re: [lmi] [lmi-commits] master 3da2e92 5/5: Follow a w3.org recommendation although it seems weird |
|
Date: |
Mon, 15 Feb 2021 16:39:45 +0100 |
On Mon, 15 Feb 2021 15:11:48 +0000 Greg Chicares <gchicares@sbcglobal.net>
wrote:
GC> On 2/14/21 11:30 PM, Vadim Zeitlin wrote:
GC> > On Fri, 12 Feb 2021 11:53:50 -0500 (EST) Greg Chicares
<gchicares@sbcglobal.net> wrote:
GC> [...]
GC> > GC> validator.w3.org warns against "https" in a System Identifier, and
GC> > GC> recommends "http" instead.
GC> >
GC> > Just in case: this is not really weird because the URL here is just a
GC> > unique string identifying the doctype, the browsers don't actually follow
GC> > it, they just recognize the string by matching it literally.
GC>
GC> https://w3ctag.github.io/web-https/
GC> | The Web platform should be designed to actively prefer secure
GC> | communication — typically, by encouraging use of "https://";
GC> | URLs instead of "http://"; ones (although exceptions like
GC> | "localhost" do exist).
GC>
GC> So I guess this is another exotic special-case exception. I just don't see
GC> why they didn't resolve it by permitting "https" in System Identifiers.
It's certainly not worth arguing about this, but this URL is not actually
used for "communication", so it's pointless to make it secure. It's just a
unique identifier of the HTML format.
GC> > Also, this might be a somewhat bigger change, but there is no reason at
GC> > all to bother with this cryptic DTD stuff since quite a few years, in HTML
GC> > 5 it's just "<!DOCTYPE html>", see
GC> >
GC> > https://html.spec.whatwg.org/multipage/syntax.html#the-doctype
GC>
GC> Is there any reason to migrate lmi's '.html' files to HTML5?
Well, getting rid of the PUBLIC "-//W3C//DTD HTML 4.01//EN" stuff is IMHO
a good enough reason on its own. And HTML5 is the standard, so IMO the
question should be why _not_ use it rather than why use it, when using it
doesn't seem to require any special efforts. I might have frightened you
with the "somewhat bigger change" above, but this "bigger change" is still
pretty small. AFAICS it would involve just:
1. Simplifying DOCTYPE as already discussed.
2. Simplifying <meta> tag to just <meta charset="utf-8">.
Of course, further improvements are possible, e.g. using HTML5 semantic
tags such as <section> or <footer> could be a good idea, but this is
optional. In theory, <tt> would would need to be replaced with <kbd>,
<var>, <code> or <samp>; and <big> -- with <hN>, <strong> or <mark>, as
these elements are obsolete in HTML5, but they are still rendered correctly
by all browsers, to the best of my knowledge, so I don't think this needs
to be done (although it's not like doing it would be very difficult
neither). It's also possible that I've missed something else from skimming
https://www.w3.org/TR/html5-diff/ quickly, but I don't think I could have
missed that much, migrating from HTML4 to HTML5 is really not a problem.
Regards,
VZ
pgpaUz7qi84Rc.pgp
Description: PGP signature