Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:3 as permitted sender) client-ip=2620:137:e000::3:3;
Date:   Fri, 3 Nov 2023 15:46:26 +0700
From:   Bagas Sanjaya <bagasdotme@gmail.com>
To:     Ken Moffat <zarniwhoop@ntlworld.com>,
        Jonathan Corbet <corbet@lwn.net>
Cc:     Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Linux Documentation <linux-doc@vger.kernel.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Akira Yokosawa <akiyks@gmail.com>,
        Stanislav Fomichev <sdf@google.com>,
        David Vernet <void@manifault.com>,
        Miguel Ojeda <ojeda@kernel.org>, James Seo <james@equiv.tech>,
        Daniel Vetter <daniel.vetter@ffwll.ch>,
        Federico Vaga <federico.vaga@vaga.pv.it>,
        Carlos Bilbao <carlos.bilbao@amd.com>
Subject: Re: [PATCH RFC RESEND 0/4] Documentation: Web fonts for kernel
 documentation
Message-ID: <ZUSzYtBpvAmM3ZRs@debian.me>
References: <20231102123225.32768-1-bagasdotme@gmail.com>
 <874ji48658.fsf@meer.lwn.net>
 <ZUQ-K7MXzHZ_oyVK@llamedos.localdomain>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha512;
        protocol="application/pgp-signature"; boundary="6kmgiKCzrjAfMsus"
Content-Disposition: inline
In-Reply-To: <ZUQ-K7MXzHZ_oyVK@llamedos.localdomain>
Precedence: bulk


--6kmgiKCzrjAfMsus
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Nov 03, 2023 at 12:26:19AM +0000, Ken Moffat wrote:
> On Thu, Nov 02, 2023 at 10:35:47AM -0600, Jonathan Corbet wrote:
>=20
> Jon, some slight nit-picking below, after comments on the stated
> problem.
>=20
> > Bagas Sanjaya <bagasdotme@gmail.com> writes:
> >=20
> [...]
> > >
> > > The solution
> > > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> > >
> > > Uniform the font choices by leveraging web fonts. Most of people read=
ing
> > > the kernel docs should already have modern browser that supports this
> > > feature (e.g. Chrome/Chromium and Firefox). The fonts are downloaded
> > > automatically when loading the page, but only if the reader don't
> > > already have ones installed locally. Subsequent docs page loading will
> > > use the browser cache to retrieve the fonts. If for some reasons the
> > > fonts fail to load, the browser will fall back to fallback fonts
> > > commonly seen on other sites.
> >=20
> Bagas,
>=20
> If loading the web font fails, you will get whichever fallback
> fonts are enabled by fontconfig and whichever fonts you, or your
> distro, have installed.  If those fonts are not generally adequate
> you should complain to your distro, or install different fonts in
> ~/.local/share/fotns and perhaps change your fonts.conf entries.

I beg to differ.

That's depending on font-family rule. For example, if I write it as:

```
body {
	font-family: "Liberation Sans", Helvetica, Arial, sans-serif;
}
```

browsers will try loading the first three fonts, in order. If a font
isn't available, they will try the next one until they can. Only then
when all other options are exhausted, generic fallback font will kick
in.

And yes, I do copying all fonts from my Windows installation (since
I dual-boot both it and Debian), then configure GNOME to use Segoe UI
as UI font (as it looks nicer to me). I also drop in font substitution
rules in `~/.config/fontconfig/conf.d` since I'm not the fan of
(obviously non-free) S=C3=B6hne and substitute it with Inter and Source
Code Pro.

>=20
> > So my immediate response to this is pretty uniformly negative.
> >=20
> > - If you don't like serif, tweaking conf.py is easy enough without
> >   pushing it on everybody else.
> >=20
> > - I'm not thrilled about adding a bunch of binary font data to the
> >   kernel, and suspect a lot of people would not feel that the bloat is
> >   worth it.
> >=20
>=20
> Jon,
>=20
> As I understand it the (woff) fonts would be downloaded on request
> by the browser if this went in.  So not a bunch of binary font data
> in the kernel, but a download from google (adding to the popularity
> of the font) and yet more font data in the browser cache.  I don't
> have any desire to see woff fonts referenced in the docs, just
> nit-picking about the details.

But I wasn't considering people using terminal-only browsers (like
Lynx).

>=20
> However -
>=20
> > - The licensing of the fonts is not fully free.
> >=20
>=20
> AFAICS, the SIL OFL allows everything except changing the font name.
> If you have the right tools you can apparently fix things like "that
> specific glyph looks ugly" or "you put a latin breve on a cyrillic
> letter" (apparently they should differ) or "You mismapped this
> codepoint to the wrong glyph". What you cannot do, if those changes
> are not accepted by the font designer/maintainer, or if the font is
> no-longer maintained, is fork it and provide it under the same name.
>=20
> You can fork, but the font name has to be changed (e.g. LinLibertine
> -> Libertinus and then the serif forked to CommonSerif).
>=20
> Oh, and you cannot sell the fonts by themselves, but you can bundle
> them with a distro or embed them.
> https://www.tldrlegal.com/license/open-font-license-ofl-explained
>=20
> Question: is that not free enough, or is that site wrong ?  If not
> free enough, is there a better licence for fonts ?

Yet Debian distributes OFL fonts in its main archive...

For me, for the fonts, I'd like CC-BY-SA instead.

Thanks.

--=20
An old man doll... just what I always wanted! - Clara

--6kmgiKCzrjAfMsus
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----

iHUEABYKAB0WIQSSYQ6Cy7oyFNCHrUH2uYlJVVFOowUCZUSzXwAKCRD2uYlJVVFO
o2kSAP9OiWMN4avk4kAZdm2oGRch0M+/NeExTW/cph9LdbqHhwEA7qIWmTFHufcs
9Mlih2xOkjGFDJfmqDPv5jeAgI6CYAw=
=qrkE
-----END PGP SIGNATURE-----

--6kmgiKCzrjAfMsus--