Using top-level selectors

On 2008/04/17 16:47 (GMT-0500) Bergamot apparently typed:

Felix Miata wrote:

Links to criticism of http://clagnut.com/blog/348/ need to be more widespread

I gave it a shot:
http://www.bergamotus.ws/rants/how-n...n-the-web.html

Good! There quite a bit I would change though. See it at
http://fm.no-ip.com/auth/tmp/how-not...n-the-web.html
--
"Either the constitution controls the judges, or the
judges rewrite the constitution." Judge Robert Bork

Team OS/2 ** Reg. Linux User #211409

Felix Miata *** http://fm.no-ip.com/

Felix Miata wrote:

On 2008/04/17 16:47 (GMT-0500) Bergamot apparently typed:

Felix Miata wrote:

Links to criticism of http://clagnut.com/blog/348/ need to be more widespread

I gave it a shot:
http://www.bergamotus.ws/rants/how-n...n-the-web.html

Good! There quite a bit I would change though. See it at
http://fm.no-ip.com/auth/tmp/how-not...n-the-web.html

Some of your suggestions are good, but I don't agree with some of what
you said, or the way you said it.

- The factory setting is not 12pt in all browsers, though it may be in
IE. mozilla and Opera both use px, or at least they do on Windows.
- I intentionally didn't want to link to the clagnut page.
- I refrained from using derogatory remarks like deeziner.
- Your synopsis of the method infers the author *recommends* 10px for
the reasons you state. That is incorrect. The clagnut article explicitly
states that value was chosen as an example only because it resulted in
nice round numbers for calculation purposes, and *not* recommended
because it is too small for most people. That, of course, doesn't mean
that those who copy the method think for themselves and use more
reasonable values. They tend not to.
- There's no reason to hail Safari. I seriously doubt many Windows users
would switch to it over this. I wouldn't, for sure.

I'll post a revision later today.

--
Berg

Felix Miata wrote:

On 2008/04/17 16:47 (GMT-0500) Bergamot apparently typed:

Felix Miata wrote:

Links to criticism of http://clagnut.com/blog/348/ need to be more widespread

I gave it a shot:
http://www.bergamotus.ws/rants/how-n...n-the-web.html

Good! There quite a bit I would change though. See it at
http://fm.no-ip.com/auth/tmp/how-not...n-the-web.html

I got an "object not found" error for:

http://fm.no-ip.com/auth/tmp/clagnut.png

--
Ed Mullen
http://edmullen.net
Everybody is somebody else's weirdo.

Bergamot wrote:

Felix Miata wrote:
Links to criticism of http://clagnut.com/blog/348/ need to be more widespread

I gave it a shot:
http://www.bergamotus.ws/rants/how-n...n-the-web.html

I get a 404 on that link.

--
Ed Mullen
http://edmullen.net
Everybody is somebody else's weirdo.

Ed Mullen wrote:

Bergamot wrote:
Felix Miata wrote:
Links to criticism of http://clagnut.com/blog/348/ need to be more widespread

I gave it a shot:
http://www.bergamotus.ws/rants/how-n...n-the-web.html

I get a 404 on that link.

I just moved it. It was renamed to be less combative, and you might have
tried it while I was in the middle of putting in the redirects.

http://www.bergamotus.ws/misc/sensib...xt-sizing.html

--
Berg

On 2008/04/18 08:34 (GMT-0500) Bergamot apparently typed:

Felix Miata wrote:

http://fm.no-ip.com/auth/tmp/how-not...n-the-web.html

Some of your suggestions are good, but I don't agree with some of what
you said, or the way you said it.

That's fine. I had no expectation of you adopting all or even portions of it
without changes. Some portions were put to produce discussion here. :-)

- The factory setting is not 12pt in all browsers, though it may be in
IE. mozilla and Opera both use px, or at least they do on Windows.

What I wrote was "All Windows OEM browsers are factory set “factory set” to
12pt". What I meant to write was "All Windows OEM browsers are factory set to
<em>nominal</em> 12pt". What I meant to write is a fact carved in stone. All
Windows OEM browsers are various versions of Internet Explorer, all of which
always have their medium text size set to nominal 12pt by Microsoft. This is
something vendors never change, so it's just a fact that IE always comes this
way, and only an IE has ever been the Windows OEM browser since back to the
original Windows 98 at least.

Users are free to switch their IE text size to something other than medium,
or goto advanced display settings and choose something other than 96 DPI, or
goto appearance and change the setting for font size to large or extra large
from normal, but this is not how they are ever factory set - with one very
big exception, which is that laptop vendors <em>often</em> change the DPI to 120.

Nevertheless, even when DPI is not 96, every web page that sets sizes in pt
will in IE6 or above in standards mode display 12pt text at the exact same
size that CSS medium is rendered. When the DPI is 120, it will still be 12pt,
even though at that DPI 12pt has become 20px instead of 16px. And this is the
proof that the IE default is 12pt and not 16px. Check with
http://fm.no-ip.com/auth/IE/font-default-IE96med.html and
http://fm.no-ip.com/auth/IE/font-default-IE120med.html or any similar page or
pages you like and you should see that the IE default is 12pt rather than
16px, or whatever px for other than 96 DPI.

Probably half the widescreen laptops for sale today and recently have had the
DPI set to 120. If the vendors weren't doing that, then the models that do it
would be outselling those that don't in a very big way, because on those
higher resolution screens the px are just too small for 96 to produce
legibility for average people.

Safari and Gecko do indeed have their sizes set in px, so users who install
them alongside IE on their 120 DPI laptops and compare pages between IE and
the others will see larger text in IE than the others, as long as they don't
raise their defaults in the others to 20px to match the 20px that IE has
assigned to 12pt.

Note too that the ancient IE versions, those prior to v6 are even worse, and
so too are the newer versions when running in quirks mode. You can see the pt
to px translations according to versions on these pages:

http://fm.no-ip.com/auth/IE/absolute-sizes-IE6.html (applies also to v7)
http://fm.no-ip.com/auth/IE/absolute-sizes-IE5.html

In http://fm.no-ip.com/auth/Moz/absolute-sizes-MvE.html you can see how
closely Gecko matches IE as long as DPI remains 96.

- I intentionally didn't want to link to the clagnut page.

I suspected that, but put it in just in case it wasn't intentional.

- I refrained from using derogatory remarks like deeziner.

Good, but maybe some kind of extra emphasis in certain places would be good too.

- Your synopsis of the method infers the author *recommends* 10px for
the reasons you state. That is incorrect. The clagnut article explicitly
states that value was chosen as an example only because it resulted in
nice round numbers for calculation purposes, and *not* recommended
because it is too small for most people. That, of course, doesn't mean
that those who copy the method think for themselves and use more
reasonable values. They tend not to.

It is true that he does say in bold type "which I’m using purely because it’s
a nice round number for example purposes". The problem is that it isn't
really true within the context of the article, or in Rutter's real intent. If
you read on you should find that the whole strategy is closely tied to using
62.5. Indeed, if you had followed the many email lists discussing this
article, including those on which the author participated in thread
discussion about his page and its method, you would know this. Those several
lists, among them WSG, css-d & WD, which collectively, and some even
individually, commonly have traffic that far outnumbers the traffic on this
newsgroup, have often discussed that it is precisely the 62.5 and its 10px
equivalence that differentiates the Clagnut method from recommendations
elsewhere to set a size less than 100% on body and/or html and em upwards or
downwards accordingly from there. Just about every recommendation elsewhere
calls for more than 62.5, most commonly 76, and none of them do anything for
pixel and layout math but make it more convoluted.

So, that 62.5 = 10px is the answer to the question you beg "If designers
insist on setting type size smaller than the default, why do it this
roundabout way instead of leaving body text at 100%, and simply setting the
content block to 75% in the first place?".

- There's no reason to hail Safari. I seriously doubt many Windows users
would switch to it over this. I wouldn't, for sure.

My point controverted your original inference that Windows users were
helpless to do anything about the end of cascade effect in Gecko and Opera,
when the truth is that those with control over their puters do have the
option to use Safari instead.

Now as to new version http://www.bergamotus.ws/misc/sensib...xt-sizing.html

Webkit and KHTML are not exactly the same thing. If you look at a Safari UA
string, you'll see a Safari version # and a Webkit version #, but no KHTML
version #. Because Webkit is a fork of KHTML, it is not the same thing, and
therefore it is technically incorrect to refer to KHTML in place of Webkit
when discussing Safari. Your opinion that Safari won't take significant share
on Windows ignores the share on Mac. Mac users browse the web too. Their
number is not insignificant, and their OS comes standard with Safari only,
just as doz comes with IE only. Safari appears to have roughly 4 times the
market share of Opera: http://www.thecounter.com/stats/2008/April/browser.php
Note too that a not insignificant part of Firefox share comes from Linux,
while none of Safari's comes from Linux.

IMO, the word "monitor" when referring to a computer display has become an
anachronism.

I've never seen a page with font-size: 8px set on body or html. 8px cannot
produce fully formed characters throughout all standard character sets, as
that size character box simply has too few px to do it. Virtually all px
precisionist idiots are aware of this. I'm sure some of them would go even
smaller if they could. :-p

Most users of minimum size seem to set the default at their comfort level,
and the minimum as something less, which means text, while always legible,
will nevertheless often be too small for <em>comfort</em> while minimum size
is in effect.
--
"Either the constitution controls the judges, or the
judges rewrite the constitution." Judge Robert Bork

Team OS/2 ** Reg. Linux User #211409

Felix Miata *** http://fm.no-ip.com/

Scripsit Felix Miata:

What I meant to write was "All Windows OEM browsers
are factory set to <em>nominal</em> 12pt".

"Nominal" is a bit obscure word. The essential point (no pun intended)
is that "pt" is not really implemented as defined in CSS specifications
and typographic practice. Rather, it is a "reference point" (again, no
pun intended) so that for the factory settings of the display unit, "pt"
roughly equals the defined unit. But when the settings are different,
either due to different DPI setting or - more importantly IMHO - due to
different resolution setting (e.g., 800×600 as opposite to 1152×864),
then the "pt" unit changes its meaning. The change is roughly
proportional to the proportion of pixel sizes in the different settings.

This also extends to "physical" units like "mm".

Thus, anything you set in "pt", "mm", "in", etc., varies in meaning.
What you set as 12pt might show up as about 15 typographic points (a
unit defined in terms of metric units), for example - or as about 9
typographic points. Usually the differences are smaller, but they can be
this large, or larger.

Of course, setting anything in "px" is absolute in terms of number of
pixels and therefore relative in terms of physical dimensions.

I've never seen a page with font-size: 8px set on body or html.

Neither have I. But I have seen 9px, which isn't much better.

8px
cannot produce fully formed characters throughout all standard
character sets, as
that size character box simply has too few px to do it.

Michael Everson says about his Everson Mono Unicode:
"I have found it quite legible at sizes as small as 4 points."
( http://www.evertype.com/emono/ )
Legibility is relative, and so is a point in terms of pixels, but
Everson Mono Unicode has "fully formed" characters even in 8px size.
Legibility depends on pixel size, among other things. And I don't think
anyone, including Everson, regards that font as particularly pretty; but
it's useful when you need a monospace font with a wide (though not
exhaustive) Unicode coverage.

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

On 2008/04/19 09:10 (GMT+0300) Jukka K. Korpela apparently typed:

Scripsit Felix Miata:

What I meant to write was "All Windows OEM browsers
are factory set to <em>nominal</em> 12pt".

"Nominal" is a bit obscure word. The essential point (no pun intended)
is that "pt" is not really implemented as defined in CSS specifications
and typographic practice. Rather, it is a "reference point" (again, no
pun intended) so that for the factory settings of the display unit, "pt"
roughly equals the defined unit. But when the settings are different,
either due to different DPI setting or - more importantly IMHO - due to
different resolution setting (e.g., 800×600 as opposite to 1152×864),
then the "pt" unit changes its meaning. The change is roughly
proportional to the proportion of pixel sizes in the different settings.

This also extends to "physical" units like "mm".

Thus, anything you set in "pt", "mm", "in", etc., varies in meaning.
What you set as 12pt might show up as about 15 typographic points (a
unit defined in terms of metric units), for example - or as about 9
typographic points. Usually the differences are smaller, but they can be
this large, or larger.

I'm not sure what the point of your exposition is, but apparently mine wasn't
clear either. By "nominal", I meant that the IE default is equal to CSS
'font-size: 12pt', regardless how big in physical size that may be on any
particular display, and that at all of the non-default (non-96; e.g. 120) DPI
settings selectable in the Windows DPI select list, its default will not be
equal to 'font-size: 16px'.

FWIW, it is quite common that 'font-size: 12pt' on Linux is in fact
physically 12pt in size. Long gone are the days when X assumes an arbitrary
DPI by default as Microsoft does. Xorg has gotten pretty good at using its knowledge
of screen size acquired as DDC and/or EDID display feedback in setting a
working X desktop DPI. If you install a current top ten distro on a 17" WUXGA
laptop, even though the screen has 133 PPI, you'll most likely be seeing 12pt
sized text any time 'font-size: 12pt' is applied on a page you view. e.g.
http://fm.no-ip.com/auth/Font/fonts-ptdemo.html

8px
cannot produce fully formed characters throughout all standard
character sets, as
that size character box simply has too few px to do it.

Michael Everson says about his Everson Mono Unicode:
"I have found it quite legible at sizes as small as 4 points."
( http://www.evertype.com/emono/ )
Legibility is relative, and so is a point in terms of pixels, but

I should have written "intelligible under typical conditions". :-p

Everson Mono Unicode has "fully formed" characters even in 8px size.
Legibility depends on pixel size, among other things. And I don't think
anyone, including Everson, regards that font as particularly pretty; but
it's useful when you need a monospace font with a wide (though not
exhaustive) Unicode coverage.

I don't see Everson anything on
http://www.codestyle.org/css/font-fa...ultsFull.shtml so
I think it probably safe to say most people don't know it exists, must less
has those described qualities that only a pixel perfectionist moron could
love. Note too I wrote "throughout all", about which Everson Mono Unicode,
not being exhaustive, apparently is not.
--
"Either the constitution controls the judges, or the
judges rewrite the constitution." Judge Robert Bork

Team OS/2 ** Reg. Linux User #211409

Felix Miata *** http://fm.no-ip.com/

Scripsit Felix Miata:

I'm not sure what the point of your exposition is, but apparently
mine wasn't clear either.

My point is that in practice, "pt" does not mean the typographic point,
except by accident.

By "nominal", I meant that the IE default
is equal to CSS 'font-size: 12pt',

I see.

regardless how big in physical
size that may be on any particular display,

Then it's 12pt only in the distorted IE sense. If if were 12pt, it would
remain constant, irrespectively of DPI setting, phase of the moon, and
other factors. The typographic unit point, as referred to in CSS
specifications, is as absolute as the physical unit meter; in fact, the
point is defined in terms of the meter.

It may of course occupy different numbers of pixels. And since
presentation on screen is discrete (even when subpixels are used),
anything set to 12pt can only be 12pt up to some accuracy. But that's a
different issue.

I don't see Everson anything on
http://www.codestyle.org/css/font-fa...ultsFull.shtml
so I think it probably safe to say most people don't know it exists,

Of course. My point was about the general remark on fonts that you made.
Special fonts may, however, be used on the web for special purposes.
People who read pages with specialized content, making heavy use of
special characters, can be expected to take some extra trouble in
downloading and installing fonts like Everson Mono Unicode or Code2000.

Note too I wrote "throughout all",
about which Everson Mono Unicode, not being exhaustive, apparently is
not.

There is _no_ font that covers all of Unicode. Everson Mono Unicode is
probably among the 10 if not 5 most exhaustive (by some metrics).

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

Felix Miata wrote:

On 2008/04/18 08:34 (GMT-0500) Bergamot apparently typed:

- The factory setting is not 12pt in all browsers, though it may be in
IE. mozilla and Opera both use px, or at least they do on Windows.

What I meant to write was "All Windows OEM browsers are factory set to
em>nominal</em> 12pt". What I meant to write is a fact carved in stone. All
Windows OEM browsers are various versions of Internet Explorer

The article may be focusing on Windows users, but not IE. So OEM or 12pt
isn't relevant, especially considering that IE doesn't support a minimum
font size.

Users are free to switch their IE text size to something other than medium,
or goto advanced display settings and choose something other than 96 DPI, or
...

For my purposes it is sufficient just to state that users can and will
change their default as they see fit. How and to what is immaterial. And
again, IE is not the focus of the article.

- There's no reason to hail Safari. I seriously doubt many Windows users
would switch to it over this. I wouldn't, for sure.

My point controverted your original inference that Windows users were
helpless to do anything about the end of cascade effect in Gecko and Opera,
when the truth is that those with control over their puters do have the
option to use Safari instead.

You made it sound like an advert in favor of switching to Safari (are
you a Mac lover, perchance?). I am vehemently opposed to forcing a
browser choice on anyone, so making such a suggestion is just plain
wrong, IMO. Besides, it would be foolish to give up a browser you are
otherwise happy with over this.

I reworded it to say that zooming text out does nothing, though forgot
about page zoom in Opera. Now that I think about that, though, it still
defeats the purpose of minimum font size, but making the user zoom (up
*or* down) to make the page comfortably readable.

Now as to new version http://www.bergamotus.ws/misc/sensib...xt-sizing.html

Webkit and KHTML are not exactly the same thing.

The differences are irrelevant, as far as the article is concerned.
Safari is KHTML-based, and the only such browser available on Windows,
at least at this time. Konqueror behaves the same as Safari as far as
minimum font size goes. Other differences between the 2 don't matter in
the scheme of things.

Your opinion that Safari won't take significant share
on Windows ignores the share on Mac.

What does the Mac share have to do with Safari on Windows? Nothing that
I can discern. Regardless, the focus of the article is Windows users.
What happens on Mac (or Linux) isn't significant in the scheme of things.

IMO, the word "monitor" when referring to a computer display has become an
anachronism.

So call me a dinosaur.

I've never seen a page with font-size: 8px set on body or html.

I didn't say that setting was on body, but I've seen attempts at font
settings as small as 6pt. They're not common, but do exist.

8px cannot
produce fully formed characters throughout all standard character sets

That doesn't stop anyone from setting a stupidly small font-size. The
browser will interpret that as it sees fit.

Most users of minimum size seem to set the default at their comfort level,
and the minimum as something less, which means text, while always legible,
will nevertheless often be too small for <em>comfort</em> while minimum size
is in effect.

I'm not sure what your point there is, but I set my default at what I
consider optimum. The minimum is the smallest I find tolerable for
reading any quantity of text. If that were too small for me I would have
bumped it up long ago. Besides, if the minimum and default were the
same, there'd be no variation in type size other than headings.
Typographically speaking, that's not a good thing, either.

--
Berg