&ensp in a monospaced font

On 2008-03-05, Jon Fairbairn <jon.fairbairn (AT) cl (DOT) cam.ac.uk> wrote:

"Jukka K. Korpela" <jkorpela (AT) cs (DOT) tut.fi> writes:
[...]
Generally, use of ligatures (at least for Latin
characters) would be something _unexpected_.

No, I don't think so. I saw it because I was expecting it!
In printed typography, use of ligatures is the norm; the
ligatures are used simply to improve appearance -- eg "fi"
either looks too spaced out or the dot on the top of the "i"
crashes into the kern of the f.

Most of the current generation of browsers (including IE, Firefox 2,
Opera 9) don't even do kerning.

So I think ligature rendering should be off by
default. But this is (at least currently) a quality of
implementation issue, not something required ín HTML specs.

That's what I meant by not mandating that it doesn't
happen. It's certainly the sort of thing that a user might
turn off, but (as long as it was done right), I think the
HTML (or CSS) specs shouldn't say that it mustn't be done.

Definitely. Arabic and Devanagari look pretty bad without ligatures,
although English doesn't really need them.

If you want ligatures, you can explicitly ask for them using
a control character or, more effectively, replacing a
sequence of characters by a compatibility character
representing their ligature, such as &#xfb01;.

The trouble with that is that it turns a presentation issue
into content.

I also think it's unreasonable to expect authors to go that much
trouble.

If my site is called "Waffle shop" I might reasonably hope that (given
a capable output device) the ffl is displayed as nicely as possible,
but I certainly want a user to get 11 characters if they copy and
paste it, and I want it to match those eleven characters in web
searches.

Exactly.

On 2008-03-05, Stan Brown <the_stan_brown (AT) fastmail (DOT) fm> wrote:

Tue, 04 Mar 2008 01:53:34 -0600 from Ben C <spamspam (AT) spam (DOT) eggs>:
On 2008-03-04, Stan Brown <the_stan_brown (AT) fastmail (DOT) fm> wrote:
I've been using Vim for many years, and I can assure you that is not
default behavior. You, or someone on your behalf, may have set up
some sort of mapping.

I've found it in the manual, it's the joinspaces option, and it is on by
default:

'joinspaces' 'js' boolean (default on)
global
{not in Vi}
Insert two spaces after a '.', '?' and '!' with a join command.

WITH A JOIN COMMAND. Not routinely in ordinary typing, which I
thought we were talking about.

I was talking about the gq command which apparently does a join, since
it applies whatever you have set for joinspaces. What do you use to wrap
paragraphs? Perhaps you use auto-formatting.

Ben C wrote:

On 2008-03-05, Jon Fairbairn <jon.fairbairn (AT) cl (DOT) cam.ac.uk> wrote:
"Jukka K. Korpela" <jkorpela (AT) cs (DOT) tut.fi> writes:
[...]
Generally, use of ligatures (at least for Latin
characters) would be something _unexpected_.
No, I don't think so. I saw it because I was expecting it!
In printed typography, use of ligatures is the norm; the
ligatures are used simply to improve appearance -- eg "fi"
either looks too spaced out or the dot on the top of the "i"
crashes into the kern of the f.

Most of the current generation of browsers (including IE, Firefox 2,
Opera 9) don't even do kerning.

So I think ligature rendering should be off by
default. But this is (at least currently) a quality of
implementation issue, not something required Ã*n HTML specs.
That's what I meant by not mandating that it doesn't
happen. It's certainly the sort of thing that a user might
turn off, but (as long as it was done right), I think the
HTML (or CSS) specs shouldn't say that it mustn't be done.

Definitely. Arabic and Devanagari look pretty bad without ligatures,
although English doesn't really need them.
If my site is called "Waffle shop" I might reasonably hope that (given
a capable output device) the ffl is displayed as nicely as possible,
but I certainly want a user to get 11 characters if they copy and
paste it, and I want it to match those eleven characters in web
searches.

Exactly.

I just copied the following three words from an Arabic web page from
both IE7 and Firefox and pasted it into Notepad on Windows Vista:

كت�سير لآيات الكتاب

Then in both copies I inserted a space after every letter. This caused
every letter's appearance to switch to its isolated form AND I was able
to insert a space into the laam-alif ligature at the beginning of the
second word, resulting in the separation of the letters. In fact, I can
do that here in Thunderbird as well:

ك ت � س ي ر ل آ ي ا ت ا ل ك ت ا ب

So it's definitely the content, and not the presentation, that's being
conveyed via the clipboard.

On Mon, 03 Mar 2008 22:12:37 -0500, Ed Mullen <ed (AT) edmullen (DOT) net> wrote:

Stan Brown wrote:
Mon, 3 Mar 2008 14:07:38 -0000 from mrcakey <nospam (AT) spamispoo (DOT) com>:
Just out of interest, what is the logic behind collapsing spaces in text?
I'd love to be able to display a document I'd written with the conventional
2 spaces after a full stop.

Two spaces is *not* conventional after a full stop, any more than
"between he and I" is conventional in grammar. It is a common error,
and that's all.


Not sure what you mean by "conventional." Whose convention? Specified
when? In what context? Defined by whom for what?

It sure was how I was taught touch typing in 1963. Period, space space.
I might even have an old style manual around here somewhere. If I do
I'll look it up. But, my memory is firm on that, and so is the memory
built into my touch-typing finger tips.

I agree, the Web is different. Not sure why the decision was made on
collapsing spaces was made. Not sure it makes any sense to me at all.
But, ok, that's the way it is, fine, I'll get with the program when
working on the Web. No problem. I'm still curious how and why that
decision was made to abandon an long standard practice.

It is typically a US habit. Some editors even ADD a second space after
a period automatically when paragraphs are reformatted. I think the
practice dates from the monospace typewriter era. -- Viajero 17:42, 25
Oct 2003 (UTC)

Here in Europe we always use 1 space.

But may be the world changes ...
"AP Writing Style" is one space after a period. 19 December 2006

Go to http://www.apstylebook.com/ask_editor.php
and.... search "spaces after a period"

In article <47cfa2eb.4563984 (AT) news (DOT) individual.net>,
Raymond.Schmit (AT) pircarre (DOT) be (Raymond SCHMIT) wrote:

On Mon, 03 Mar 2008 22:12:37 -0500, Ed Mullen <ed (AT) edmullen (DOT) net> wrote:

Stan Brown wrote:
Mon, 3 Mar 2008 14:07:38 -0000 from mrcakey <nospam (AT) spamispoo (DOT) com>:
Just out of interest, what is the logic behind collapsing spaces in text?
I'd love to be able to display a document I'd written with the
conventional
2 spaces after a full stop.

Two spaces is *not* conventional after a full stop, any more than
"between he and I" is conventional in grammar. It is a common error,
and that's all.


Not sure what you mean by "conventional." Whose convention? Specified
when? In what context? Defined by whom for what?

It sure was how I was taught touch typing in 1963. Period, space space.
I might even have an old style manual around here somewhere. If I do
I'll look it up. But, my memory is firm on that, and so is the memory
built into my touch-typing finger tips.
[snip]

It is typically a US habit. Some editors even ADD a second space after
a period automatically when paragraphs are reformatted. I think the
practice dates from the monospace typewriter era. -- Viajero 17:42, 25
Oct 2003 (UTC)

Here in Europe we always use 1 space.

Not in the UK when I was there. However, I think it was originally
a typewriter thing. Publishers tend (or at least, tended) to use a
long space after a period, rather than a regular single space. The
only way to do the same thing on a monospaced typewriter was to hit
the space bar twice.

As a point of reference, the writer's guides for the two publishers
I've been involved with have specified that the submitted electronic
version of the manuscript should only have a single space after the
period; their computerized typesetter automatically converts ". "
into whatever is required by the publishers specific design guidelines.

But may be the world changes ...
"AP Writing Style" is one space after a period. 19 December 2006

Go to http://www.apstylebook.com/ask_editor.php
and.... search "spaces after a period"

Scripsit Ben C:

Arabic and Devanagari look pretty bad without ligatures,
although English doesn't really need them.

A good-quality browser displays different scripts differently. It is
reasonable to expect that a browser (or a word processor) e.g. selects
suitable contextual glyphs for an Arabic letter, since that's normal for
the Arabic script, whereas it should not convert combinations of Latin
letters into ligatures, for a comparable reasons: that's not normally
done (though typesetting programs may do that, and they normally give
the user control over that). It would be unexpected and may violate
authors' reasonable expectations.

If ligatures of Latin letters are desired on web pages, then the
practical approach is to use the compatibility characters. It's actually
not difficult. You might teach your authoring software generate them
(from the input of certain character strings), or you might define
keyboard shortcuts, or you might do some postprocessing. Unfortunately,
search engines will be confused, and so will many spelling checkers and
Find functions of browsers, so there's a cost.

Some day in the distant future, CSS might have a property for enabling
ligature binding, hopefully with the reasonable initial value (for the
Latin script) of no ligation.

--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/