W3C home > Mailing lists > Public > spec-prod@w3.org > January to March 2018

Re: [Restyle] Auto hyphenation can make the spec ambiguous

From: Marcos Caceres <w3c@marcosc.com>
Date: Wed, 14 Feb 2018 18:45:42 -0800
Message-ID: <CAAwChxMc1iYNgEU=yhaEM1PiXjJupxXe5rz7qyLr278OMO+DOg@mail.gmail.com>
To: fantasai <fantasai.lists@inkedblade.net>, spec-prod@w3.org, Chris Lilley <chris@w3.org>
On February 14, 2018 at 11:07:33 AM, fantasai
(fantasai.lists@inkedblade.net) wrote:
> On 02/12/2018 07:16 PM, Chris Lilley wrote:
> > A bug was raised against a CSS specification, because it seemed as if a hyphen was allowed
> in a keyword ("open-type" as well
> > as "opentype").
> >
> > https://github.com/w3c/csswg-drafts/issues/2307
> >
> > It turned out not to be a bug in the grammar but to be introduced by the styling. Furthermore,
> that bug is in base.css and
> > thus affects all W3C specs, not just CSS WG ones.
> >
> > The rule in question is
> >
> > body { hyphens: auto}
> >
> > Hyphenation is fine in general, but not when it alters the technical content of a specification.
> Code should be wrapped in . Then you would not run into such problems. :)

The following might be ReSpec's additional styles maybe screwing up -
I've not investigated yet! Nevertheless, wanted to highlight a few
things I've seen that relate to this, that perhaps we can fix in

I've encountered two similar problems, which can be seen in [1]:

When code (specially IDL) is too long, it can overflow and get hidden.
This is not great, because it can lead to situations where it looks
like the interface defines an "optional SecureContext" (yikes, thats
both invalid and bad!), as seen in [1].

And two, also shown in the second screenshot of [1], instead of
hyphenation, words wrapped in `code` also get broken broken up in two
(thankfully with no hyphen! but still not great either).

> We can turn off hyphenation on W3C specs altogether if you prefer, though.

Might be a good idea, just to be safe.

[1] https://github.com/w3c/respec/issues/1237
Received on Thursday, 15 February 2018 02:46:08 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 15 February 2018 02:46:08 UTC