Discussion:
text-decoration-skip-ink auto should continue past behavior - 30+ years of underline behavior changed by latest CSS draft
(too old to reply)
OwN-3m-All
2018-02-21 15:33:09 UTC
Permalink
I initially thought this was a problem with Chrome (since they seem to
be one of the early adopters - bug report here:
https://bugs.chromium.org/p/chromium/issues/detail?id=813256#c2), but
now that I've seen the actual spec, I'm shocked that the auto value
for the text-decoration-skip-ink property is to change the way
underlined text has worked since the beginning of computers!

https://drafts.csswg.org/css-text-decor-4/#text-decoration-skip-ink-property

Underlined text should always have the line over all characters.
Hanging characters should not be exempt. If you want to change the
default behavior of underlined text, don't force that behavior on us.
"auto" should be "UA must draw contiguous lines without interruptions,
even when they cross over a glyph.". Any other behavior is
NON-STANDARD.

Could someone please re-review the draft. It's crazy to think that
hyperlinks and any text with text-decoration: underline will soon be
showing up differently for no reason whatsoever!
Ambrose LI
2018-02-21 16:38:21 UTC
Permalink
Post by OwN-3m-All
Underlined text should always have the line over all characters.
Hanging characters should not be exempt. If you want to change the
default behavior of underlined text, don't force that behavior on us.
"auto" should be "UA must draw contiguous lines without interruptions,
even when they cross over a glyph.". Any other behavior is
NON-STANDARD.
Not really. Only non-standard on non-Macs, I believe. But I agree auto
should default to traditional, OS-specific behaviour.
--
Ambrose Li // http://o.gniw.ca / http://gniw.ca
If you saw this on CE-L: You do not need my permission to quote
me, only proper attribution. Always cite your sources, even if
you have to anonymize and/or cite it as "personal communication".
Myles C. Maxfield
2018-02-21 17:22:29 UTC
Permalink
Post by OwN-3m-All
I initially thought this was a problem with Chrome (since they seem to
https://bugs.chromium.org/p/chromium/issues/detail?id=813256#c2), but
now that I've seen the actual spec, I'm shocked that the auto value
for the text-decoration-skip-ink property is to change the way
underlined text has worked since the beginning of computers!
Yep. This change is intentional.
Post by OwN-3m-All
https://drafts.csswg.org/css-text-decor-4/#text-decoration-skip-ink-property
Underlined text should always have the line over all characters.
Nope. This is how computers have historically rendered text. However, historically, most high-typographic-quality examples which include underlines make the underlines skip over the descenders.

This is a progression, and improves typography on the Web.
Post by OwN-3m-All
Hanging characters should not be exempt. If you want to change the
default behavior of underlined text, don't force that behavior on us.
It isn’t forced on you.

:root {
text-decoration-skip-ink: none;
}
Post by OwN-3m-All
"auto" should be "UA must draw contiguous lines without interruptions,
even when they cross over a glyph.". Any other behavior is
NON-STANDARD.
The CSS specification defines what is standard and non-standard, and the CSS specification states that “auto” is the initial value. So, indeed, the behavior you propose is non-standard.
Post by OwN-3m-All
Could someone please re-review the draft. It's crazy to think that
hyperlinks and any text with text-decoration: underline will soon be
showing up differently for no reason whatsoever!
See above. Not crazy; it’s a progression.
Myles C. Maxfield
2018-02-21 17:27:27 UTC
Permalink
Post by Myles C. Maxfield
Post by OwN-3m-All
I initially thought this was a problem with Chrome (since they seem to
https://bugs.chromium.org/p/chromium/issues/detail?id=813256#c2), but
now that I've seen the actual spec, I'm shocked that the auto value
for the text-decoration-skip-ink property is to change the way
underlined text has worked since the beginning of computers!
Yep. This change is intentional.
Post by OwN-3m-All
https://drafts.csswg.org/css-text-decor-4/#text-decoration-skip-ink-property
Underlined text should always have the line over all characters.
Nope. This is how computers have historically rendered text. However, historically, most high-typographic-quality examples which include underlines make the underlines skip over the descenders.
Or, stated differently, underlines cross descenders in existing software because it was convenient for software authors writing code. However, we’ve done research in underlines through the ages (way before computers were invented) and the best typographical samples always use skipping underlines. This is a situation where changing behavior on the Web doesn’t break content and automatically improves typography for everyone. (And it has an opt-out mechanism if for some reason you don’t like good typography.)
Post by Myles C. Maxfield
This is a progression, and improves typography on the Web.
Post by OwN-3m-All
Hanging characters should not be exempt. If you want to change the
default behavior of underlined text, don't force that behavior on us.
It isn’t forced on you.
:root {
text-decoration-skip-ink: none;
}
Post by OwN-3m-All
"auto" should be "UA must draw contiguous lines without interruptions,
even when they cross over a glyph.". Any other behavior is
NON-STANDARD.
The CSS specification defines what is standard and non-standard, and the CSS specification states that “auto” is the initial value. So, indeed, the behavior you propose is non-standard.
Post by OwN-3m-All
Could someone please re-review the draft. It's crazy to think that
hyperlinks and any text with text-decoration: underline will soon be
showing up differently for no reason whatsoever!
See above. Not crazy; it’s a progression.
Alexander Dekker
2018-02-21 19:02:22 UTC
Permalink
Post by Myles C. Maxfield
Post by Myles C. Maxfield
[...]
Post by OwN-3m-All
https://drafts.csswg.org/css-text-decor-4/#text-decoration-skip-ink-property
Underlined text should always have the line over all characters.
Nope. This is how computers have historically rendered text. However, historically, most high-typographic-quality examples which include underlines make the underlines skip over the descenders.
Or, stated differently, underlines cross descenders in existing software because it was convenient for software authors writing code. However, we’ve done research in underlines through the ages (way before computers were invented) and the best typographical samples always use skipping underlines. This is a situation where changing behavior on the Web doesn’t break content and automatically improves typography for everyone. (And it has an opt-out mechanism if for some reason you don’t like good typography.)
“I am a Web developer with a degree in typographic design and I approve
this message.”

Also, back in the day we used to have tiny aliased serif type on gray
backgrounds. Almost illegible, but it was a thing. I applaud efforts to
make type on the Web less terrible (also known as more readable).

My two cents.


Kind Regards,


Alexander Dekker
deidee
Allan Sandfeld Jensen
2018-02-26 23:47:35 UTC
Permalink
Post by Myles C. Maxfield
Post by Myles C. Maxfield
Post by OwN-3m-All
I initially thought this was a problem with Chrome (since they seem to
https://bugs.chromium.org/p/chromium/issues/detail?id=813256#c2), but
now that I've seen the actual spec, I'm shocked that the auto value
for the text-decoration-skip-ink property is to change the way
underlined text has worked since the beginning of computers!
Yep. This change is intentional.
Post by OwN-3m-All
https://drafts.csswg.org/css-text-decor-4/#text-decoration-skip-ink-prop
erty
Underlined text should always have the line over all characters.
Nope. This is how computers have historically rendered text.
Glyphs with descender parts (eg. pqjgy) must overlap an underline
CSS Test: 'underline' decoration painting order and descender
http://test.csswg.org/suites/css2.1/nightly-unstable/html4/painting-order-un
derline-001.htm (you need to download
https://www.w3.org/Style/CSS/Test/Fonts/Ahem/
and install Ahem font
AHEM____.TTF 2017-01-31 20:55 22K
to view that test)
Post by Myles C. Maxfield
However, historically, most high-typographic-quality examples which
include underlines make the underlines skip over the descenders.
Or, stated differently, underlines cross descenders in existing
software because it was convenient for software authors writing code.
However, we’ve done research in underlines through the ages (way
before computers were invented) and the best typographical samples
always use skipping underlines. This is a situation where changing
behavior on the Web doesn’t break content
If textual links have descenders (or a blank space), then it may look
like there is 2 links and not 1 link. On 1 hand, it will be easier to
read (typographically speaking) the textual link but it may confuse the
user (or lead him/her to hesitate) in thinking that there are 2 textual
links.
Those two arguments really should end the discussion right here. It is
documented CSS 2.x behavior and it is confusing for users given existing
behavior.

As I see it the only way we can have the underline skip the descender and not
look like two links is by making the skipped part very short, that is
underlining a glyph partially, but that would require font support and would
in that case be a matter of the font dictating how it is to be underlined and
no longer a CSS matter.

Note also that some fonts have and underline position below the descenders
which makes the skipping even worse.

Regards
'Allan
OwN-3m-All
2018-02-21 17:34:37 UTC
Permalink
Post by Myles C. Maxfield
This is a progression, and improves typography on the Web.
Sorry, but that's not progression. You're plain wrong. Apple should
not be able to dictate what is progress and what isn't. Such a
garbage argument.
Post by Myles C. Maxfield
It isn’t forced on you.
Yes, it is. The way text-decoration works has been changed, and now
to get it to behave the old STANDARD way, I have to add a CSS style.
Are you kidding me? If you want to rewrite the way underlining works,
you should have to be the one to add your CSS override. This change
breaks the way it worked before causing people who care (like me) to
have to revisit ALL WEBPAGES and add CSS rules to make sure they
display the old way.
Myles C. Maxfield
2018-02-21 17:43:48 UTC
Permalink
Post by OwN-3m-All
Post by Myles C. Maxfield
This is a progression, and improves typography on the Web.
Sorry, but that's not progression. You're plain wrong.
By this argument, no browser should ever fix any bugs because doing so would cause behavior changes.
Post by OwN-3m-All
Apple should
not be able to dictate what is progress and what isn't. Such a
garbage argument.
The CSS specification states this behavior, not Apple. You are posting on the CSS Working Group’s mailing list.
Post by OwN-3m-All
Post by Myles C. Maxfield
It isn’t forced on you.
Yes, it is. The way text-decoration works has been changed, and now
to get it to behave the old STANDARD way, I have to add a CSS style.
Are you kidding me? If you want to rewrite the way underlining works,
you should have to be the one to add your CSS override. This change
breaks the way it worked before causing people who care (like me) to
have to revisit ALL WEBPAGES and add CSS rules to make sure they
display the old way.
User-stylesheets are a way for users to modify every webpage on the web without creating a detriment to any other user. I suggest you use them (or a browser which supports them, like Safari).
OwN-3m-All
2018-02-21 18:03:02 UTC
Permalink
@Tadeus,

The same could be said to you. Could you please cite some legibility
or typographic material in support for the change? Either way, it's
technically an opinion. I prefer the other style (the old way). I
can't get used to the new change and personally think it's wrong.
However, since underlining has worked the way it has for so long, why
is the default being changed? The configurable option should be
enough to allow people to choose the way they'd like for it to be
styled.

More choice is good, but changing the way it's worked for so long is a
bad idea in my opinion.
OwN-3m-All
2018-02-21 18:52:28 UTC
Permalink
Then, you are experiencing a culture shock. You will get over it in due time.
That's a bad argument. It's best to leave the default the way it is
and add these supposed "enhancements" as available features through
addition of CSS code... not the other way around.
OwN-3m-All
2018-02-21 19:04:50 UTC
Permalink
If every browser had chosen to not cut through descenders, would you had any problem with the current CSS strengthening the default to not cut through descenders?
Thankfully, we didn't go that route in the beginning. I prefer the
line cutting through the descenders after seeing both styles. Since
I've seen it that way all of my life, I think that's the way it should
be done. That's my whole point. A default has been established, and
we should keep using that as the default.

Open any word processing app like Microsoft Word. Underlined text
works the way I describe. It's only recently that someone had the
idea to change it, and why? What was wrong with the old way?
Nothing.
OwN-3m-All
2018-02-21 19:41:06 UTC
Permalink
But, see that Gérard has raised an objection on the new behavior with merit in the other e-mail.
My objection has merit because the default behavior has changed (a big
no no in development), and the new behavior is confusing. I don't
need to provide anything else.

Yes, the problem with not cutting through the descenders is that links
look like multiple separate entities. It actually makes it harder for
me to read. We shouldn't cater to every disability. The default
should remain the way it was, and those with disabilities who find it
easier to read with the new way can use their precious user style
sheets.
OwN-3m-All
2018-02-22 03:51:50 UTC
Permalink
CSS 2.x
Appendix E. Elaborate description of Stacking Contexts, E.2 Painting order
https://www.w3.org/TR/CSS21/zindex.html#painting-order
https://www.w3.org/TR/CSS22/zindex.html#painting-order
states that underline text-decoration is painted first and then glyphs are
painted. Therefore, glyphs with descenders (gjpqy) will paint over an
underline. Since underline and glyphs often use the same color, this will
give the visual impression that underline cut through descenders.
That sounds right! Not rendering the underline under or above (I'm
fine with either as long as the line isn't broken apart) certain
descending characters makes no sense and is inconsistent with how
underlined text has ALWAYS worked in the NORMAL world.
Myles C. Maxfield
2018-02-21 18:02:54 UTC
Permalink
[...]
Post by Myles C. Maxfield
Post by OwN-3m-All
It isn’t forced on you.
Yes, it is. The way text-decoration works has been changed, and now
to get it to behave the old STANDARD way, I have to add a CSS style.
Are you kidding me? If you want to rewrite the way underlining works,
you should have to be the one to add your CSS override. This change
breaks the way it worked before causing people who care (like me) to
have to revisit ALL WEBPAGES and add CSS rules to make sure they
display the old way.
User-stylesheets are a way for users to modify every webpage on the web without creating a detriment to any other user. I suggest you use them (or a browser which supports them, like Safari).
@Myles: then you admit that it is being forced on everyone since now
those who like to have the old behavior have to do something to get
the old behavior back, no?
Not “forced on everyone.” Instead, it is on by default. (“Forced” implies that there is no way back.)
@OwN-3m-All: as Chris Lilley wrote in the other e-mail, it is best to
present a technical argument for the merit of the old behavior since
otherwise your argument is just due to kind of culture shock that
everyone will get accustomed to after sometime.
--
Best regards,
Tadeus
Myles C. Maxfield
2018-02-21 18:05:01 UTC
Permalink
Post by Myles C. Maxfield
By this argument, no browser should ever fix any bugs because doing so would cause behavior changes.
There was no bug that needed to be fixed. This change has caused a
bug. Though some individuals may consider this underline change an
enhancement, it's also technically a BUG since it changes default
behavior. The change could still go in as long as it doesn't change
the way underlined text has displayed for years.
Post by Myles C. Maxfield
User-stylesheets are a way for users to modify every webpage on the web without creating a detriment to any other user. I suggest you use them (or a browser which supports them, like Safari).
The purpose of CSS is to make a website display the way the AUTHOR chooses.
This is false. Engines which accept CSS are called “user-agents” in the CSS specification precisely because they operate on the user’s behalf, not the content author’s behalf. This is fundamental to the Web (indeed, it is fundamental to personal computing in general).
Tadeus Prastowo
2018-02-21 18:14:21 UTC
Permalink
Post by Myles C. Maxfield
Post by Myles C. Maxfield
By this argument, no browser should ever fix any bugs because doing so would cause behavior changes.
There was no bug that needed to be fixed. This change has caused a
bug. Though some individuals may consider this underline change an
enhancement, it's also technically a BUG since it changes default
behavior. The change could still go in as long as it doesn't change
the way underlined text has displayed for years.
Post by Myles C. Maxfield
User-stylesheets are a way for users to modify every webpage on the web without creating a detriment to any other user. I suggest you use them (or a browser which supports them, like Safari).
The purpose of CSS is to make a website display the way the AUTHOR chooses.
This is false. Engines which accept CSS are called “user-agents” in the CSS specification precisely because they operate on the user’s behalf, not the content author’s behalf. This is fundamental to the Web (indeed, it is fundamental to personal computing in general).
+1.

--
Best regards,
Tadeus
Chris Lilley
2018-02-22 14:21:11 UTC
Permalink
There was no bug that needed to be fixed. This change has caused a
bug. Though some individuals may consider this underline change an
enhancement, it's also technically a BUG since it changes default
behavior. The change could still go in as long as it doesn't change
the way underlined text has displayed for years.
I looked into best practices for underlining. Bringhurst barely mentions
it (I checked both the first and the second edition), but briefly notes,
under underscore "To clear descenders, a repositioned version of the
character is required" (page 286).

Dowding on the other hand is more forthcoming. Having first warned
against underlining in principle: "Underlining as a method of
emphasizing a word or phrase in a displayed setting is a crude relic of
Victorian typography. Emphasis can, & should, be achieved in other ways"
(page 77) he then gives advice on how to do it: "If one is compelled to
underline then the underlining must be properly done. The rule must be
close to 'the line' or feet of the letters. Where there is a descender,
the rule should be set on each side of it. When  a descender comes at
the beginning or end of a word in lower-case that has to be underlined,
set the rule fairly close to the feet of the non-descending letters
only." (page 78)

There follows a figure, where underlining on the word "Begin" avoids the
descender of the "g", and also where underlining on the word "many"
avoids the descender of the "y" (and is also cut off at an angle, to
match the slope of the descender). My edition was published in 1995; the
first edition was 1954. The claim of "30 years of underline behavior"
can clearly be put to rest by 64 years of typographic commentary.

Bringhurst, Robert "The Elements of Typographic Style". Second edition,
1996.
Dowding, Geoffrey "Finer Points in the spacing & arrangement of Type".
Third revised edition, 1995.
--
Chris Lilley
@svgeesus
Technical Director @ W3C
W3C Strategy Team, Core Web Design
W3C Architecture & Technology Team, Core Web & Media
Chris Lilley
2018-02-21 17:49:06 UTC
Permalink
Post by OwN-3m-All
Post by Myles C. Maxfield
This is a progression, and improves typography on the Web.
Sorry, but that's not progression. You're plain wrong. Apple should
not be able to dictate what is progress and what isn't. Such a
garbage argument.
Please keep your arguments on a technical level, with courtesy to other
participants.
https://www.w3.org/Consortium/pwe/

Your point appears to be either

a) that underlines which cut through descenders are preferable. If so,
please cite some legibility or typographic material in support of this
assertion.

or

b) that underlines which cut through descenders are both commonplace and
desirable, depended upon, and should not change. In that case, please
provide some evidence that this is desirable (as well as just being common).
--
Chris Lilley
@svgeesus
Technical Director @ W3C
W3C Strategy Team, Core Web Design
W3C Architecture & Technology Team, Core Web & Media
Stefan Schumacher
2018-02-21 18:07:21 UTC
Permalink
Post by OwN-3m-All
Sorry, but that's not progression. You're plain wrong. Apple should
not be able to dictate what is progress and what isn't. Such a
garbage argument.
Please use your real name in discussions here.

There are always different opinions, but keeping a clean and helpful
language is the only way to change things, if there is good reason.

Thank you
Stefan Schumacher
fantasai
2018-02-28 03:55:16 UTC
Permalink
Post by Stefan Schumacher
Post by OwN-3m-All
Sorry, but that's not progression. You're plain wrong. Apple should
not be able to dictate what is progress and what isn't. Such a
garbage argument.
Please use your real name in discussions here.
There's no need to use a real name in discussions here.
Post by Stefan Schumacher
There are always different opinions, but keeping a clean and helpful
language is the only way to change things, if there is good reason.
Remaining courteous, however, is indeed a requirement. :)

~fantasai

Ambrose LI
2018-02-21 18:07:28 UTC
Permalink
Post by OwN-3m-All
Post by Myles C. Maxfield
This is a progression, and improves typography on the Web.
Sorry, but that's not progression. You're plain wrong. Apple should
not be able to dictate what is progress and what isn't. Such a
garbage argument.
It *is* progress. In traditional typography underlines don’t cross
descenders. Apple totally did it right; it’s the others who have been
doing it wrong.

The problem isn’t standard or not standard, but that underlining is
not visual decoration. It can mean something. And changing the default
changes meaning if meaning is intended.
--
Ambrose Li // http://o.gniw.ca / http://gniw.ca
If you saw this on CE-L: You do not need my permission to quote
me, only proper attribution. Always cite your sources, even if
you have to anonymize and/or cite it as "personal communication".
OwN-3m-All
2018-02-21 18:15:11 UTC
Permalink
Post by Ambrose LI
It *is* progress.
In your opinion. I don't agree.
Post by Ambrose LI
Apple totally did it right; it’s the others who have been doing it wrong.
Again, opinion. It's not a matter of fact. The underline change is
opinion based.

The new way is inconsistent and looks bad to my eyes.

You really prefer the way Chrome renders this text as oppose to Firefox?

Loading Image...

I don't.
Brad Czerniak
2018-02-21 19:29:50 UTC
Permalink
Another non-member chiming in here with some fun background:

IIUC, a big push for text-decoration-skip-ink includes a popular medium
design post
<https://medium.design/crafting-link-underlines-on-medium-7c03a9274f9>
about their work with underlines, which was spurred on by a change to Chrome
<https://bugs.chromium.org/p/chromium/issues/detail?id=338148>.

One important argument for skipping descenders by default is that it's more
accessible for users with dyslexia and other differing abilities (3-7% of
Post by OwN-3m-All
Post by Ambrose LI
It *is* progress.
In your opinion. I don't agree.
Post by Ambrose LI
Apple totally did it right; it’s the others who have been doing it wrong.
Again, opinion. It's not a matter of fact. The underline change is
opinion based.
The new way is inconsistent and looks bad to my eyes.
You really prefer the way Chrome renders this text as oppose to Firefox?
https://s17.postimg.org/qqhjdkyfx/chrome_underline_ugly.jpg
I don't.
OwN-3m-All
2018-02-22 01:27:55 UTC
Permalink
Finally, note that accessibility advocacy groups and experts formally discourage using underlining of non-clickable text and items. You want to reduce confusion, guesswork, hesitation of users.
Yep, I do this already. However, with the recent change, links are
now confusing due to W3C changing the default look of underlined text.
It looks like W3C ignored those advocacy groups in making this
decision.
OwN-3m-All
2018-02-22 21:17:49 UTC
Permalink
I'm glad you are all experts on underlining (Is there even such a
thing? If there is, that is pathetic)! Links have ALWAYS been
underlined as best practice for web development. Breaking them apart
for hanging characters all of a sudden is unreasonable. I don't care
what any supposed "expert" says. Common sense says you don't change
something unless it's broken, and it wasn't broken. No one asked for
this change. If someone did, it most certainly should NOT have become
the default behavior. It should be on the person wanting the change
to have to add CSS styles for this CHANGED behavior. Instead, I'm now
being punished by having to get rid of this stupid change by adding
CSS styles to all of my websites and pages (which are numerous)
because of some silly unnecessary DEFAULT behavior change.

Change the "DEFAULT" way its displayed for decades if you want... but
it totally doesn't make you right! It's a question of how something
displays, and I prefer the old default behavior which was with us for
decades.
Alan Stearns
2018-02-22 21:26:59 UTC
Permalink
On 2/22/18, 1:20 PM, "OwN-3m-All" <***@gmail.com> wrote:

I'm glad you are all experts on underlining (Is there even such a
thing? If there is, that is pathetic)!

This is not an acceptable way of engaging with this community. Please keep to a technical argument, and do not use sarcasm or insults.

I prefer the old default behavior which was with us for
decades.

You have stated and restated your preference several times in this thread. If you have nothing more to add, please consider your message received.

Thanks,

Alan
Ambrose LI
2018-02-22 21:50:59 UTC
Permalink
Post by OwN-3m-All
I'm glad you are all experts on underlining (Is there even such a
thing? If there is, that is pathetic)! Links have ALWAYS been
underlined as best practice for web development. [stuff deleted]
This is white supremacy. In some languages underlines are punctuation,
and underlining links is,technically speaking, grammatically wrong.

And if you’re not aware, on the topic of not changing the default (as
opposed to the new default being an improvement) several people are
actually agreeing with you (including myself), even though you
continue to disagree with us. I'm not sure what this means.
--
Ambrose Li // http://o.gniw.ca / http://gniw.ca
If you saw this on CE-L: You do not need my permission to quote
me, only proper attribution. Always cite your sources, even if
you have to anonymize and/or cite it as "personal communication".
Allan Sandfeld Jensen
2018-02-26 23:39:14 UTC
Permalink
Post by Ambrose LI
Post by OwN-3m-All
I'm glad you are all experts on underlining (Is there even such a
thing? If there is, that is pathetic)! Links have ALWAYS been
underlined as best practice for web development. [stuff deleted]
This is white supremacy. In some languages underlines are punctuation,
and underlining links is,technically speaking, grammatically wrong.
??
What the...

We are speaking about latin like text. For other types of text other ways of
indicating links, but many parts of the CSS standards relate to latin-like
text specifically and makes little or no sense in other scripts: Italic, sans-
serif vs serif, upper-case/lower-case, floating ::first-letter, ligatures,
kerning, and I could keep going. In fact most of is *cough* "white supremecy",
or perhaps better put, the parts dealing with latin-like text rendering are
mainly dealing with latin-like text rendering.

'Allan
Loading...