What do you think of resizing 1em to 10px?

Alan J. Flavell wrote:

On Wed, 9 Jun 2004, Barry Pearson wrote:
[snip]
You've identified a problem (let's call it "P") namely the differences
in perceived size of different fonts at the same nominal size. If
you've been paying any attention, then it's hard to believe that you
wouldn't have already located one existing page, at least, on the
topic, namely http://www.xs4all.nl/~sbpoley/webmatters/verdana.html
Indeed. And I know of others too. I've started a page on the topic, and I will
add to it over time:
http://www.barry.pearson.name/articl...ect_values.htm
But then you bundle it up with other complicating issues (let's call
them "Q") such as specifying size in pixels and/or CSS px units, and
the assumed display settings of your silent "majority" of web users.
I don't "bundle it up". They are a set of issues that need to be addressed.
They may need separate solutions. Some of them may have to be sacrificed for
the sake of others. But they all, at least, need to be on the agenda, so that
positive decisions can be made about priorities, etc.

If they were demonstrably independent issues, then of course we could deal
with them separately! I'm a business analyst - I put lots of energy into
turning complicated problems into separate simple problems, then dealing with
these separately. But that separation hasn't been done here. Instead, I think
what is happening is that non-separate problems are getting ignored or
dismissed. (I often meet this tactic in my work!)
Not surprisingly, people post responses which address the complicating
issues Q, and then you accuse them of not understanding the real
problem P. I can only repeat that problem P is well-known and
documented, but that specifying font size in px units isn't its
solution - and brings its own disadvantages. The solution to P would
be some kind of advanced font-size-adjust specification, but that
isn't currently available: so, web authors have to do what they can
within the parameters that are available.
Indeed they do. And the obvious thing to do is to work out what
font-size-adjust (which is anyway flawed) would say, then make this explicit.
(There are also other approaches). I am coming to the conclusion that many,
perhaps the majority, of authors, are specifying the font sizes that would be
calculated by a proper font-size-adjust property:
http://www.barry.pearson.name/articl...ect_values.htm
http://www.barry.pearson.name/articles/text/trends.htm
I state a problem, and supply some evidence to support my concerns.
You appear to casually dismiss my concerns,

I don't agree. You have identified a known problem, on some scale,
but I'm not the only one who considers that your proposed solution is
like using a sledgehammer to crack a nut - it might do what you want,
for you and your silent "majority", but some of us around here
consider that it has knock-on effects that are out of scale with the
problem it aims to address.

First, please remember this fact: *none* of my web sites specifies either the
typeface or the font-size for the main text, and this has been the case for
quite some time. Unlike some of the people posting here, for example Brian,
who specifies Arial. So don't assume that people responding to me have some
sort of consensus that disagrees with what I do. Some people posting here
probably agree with what I do, and by implication disagree with people who are
disagreeing with me. There is *no* consensus in this newsgroup. There isn't
even a consensus among people responding to me in this thread. There are
incompatible views. Who knows what the majority view is? And how relevant this
is?

Second, I have been examining various approaches to this issue, ("what should
an author specify for the main text size?") I have come the conclusion that
there are 3 main approaches:
http://www.barry.pearson.name/articles/text/

1. "Don't specify either font or font-size for main text". This theory assumes
that the user's own preferences, for example determined by the browser, should
be used for main text.

2. "Specify font, but not font-size, for main text". This is hopefully a
minority theory. The view is that it is OK to override the user's font
preference, but not the user's font-size preference. Its logic is seriously
flawed.

3. "Specify both font and font-size for main text". This appears to be the
adopted theory of the majority of web sites in the world.

I currently practise "1", but I think it has some serious flaws, and needs to
be subjected to scrutiny. I believe most authors in the world practise "3",
and have good reason to do so, because it may be conform to what suits the
majority of *users* need. (Technically-aware users can probably handle
whatever authors do. I'm really concerned with non-technically-aware users). I
believe that "2" has no merit, and sufficient disadvantages for it to be
advised-against in this newsgroup.
But if you want to -rely- on a technique which requires a proportion
of your readers to completely overrule your size specification, then
sure, you'll do it anyway - just like the commercial sites that you
seem to want to praise (most of which are rigid, fragile and
relatively inaccessible in my experience), despite the fact that your
own work shows that you know how to do better.

See above. I believe an author should either not change either (which is what
I currently do - or rather, not do), or should change both. Leave it all to
the user, or take control and make it look right. But don't tweak bits of it
and leave the result what neither author nor user would logically consider.
I've put some comparison material at:
http://www.barry.pearson.name/articl...comparison.htm

I would welcome contributions to the pages above. I would like to cite the
major sources for this topic.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #55

Brian wrote:

Barry Pearson wrote:
Brian wrote:
People who visit my sites have, so far, had no complaints about
font-size.
Then let's start right now.

I object to the font size on the following pages:
http://www.julietremblay.com/brian/

Very good. One person has complained about font size being too large.
That person, however, is a web author with an ax to grind, and is hard
to distinguish from a troll. Since the complaint doesn't seem genuine,
I won't put solving the problem at the top of my to-do list.

I wonder how many people actually know about that web site? What proportion of
them all am I?

I ask you to have a look at the following page, which is intended not only to
give others an insight into how their preferences may be viewed, but also show
what a particular user sees.
http://www.barry.pearson.name/articl...comparison.htm

How about showing us a screen-shot, from your system, of the table at the top,
so that we can see what you see when authors specify typefaces but don't
specify font sizes? (The image at the bottom gives my view). Let's try to
advance our knowledge by sharing our separate views of tricky pages!
Of course, even if it *were* on top of that list, there is no way to
solve the problem for you that would not, at the same time,
inconvenience others.

I accept that there isn't a consensus among users. Therefore there isn't one
single authoring approach that works for all users. We are all just discussing
our opinions of what to do. We are all having to make compromises. We just
disagree about which are the best compromises.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #56

Shawn K. Quinn wrote:

Barry Pearson wrote:
Then let's start right now.
I object to the font size on the following pages:
http://www.julietremblay.com/brian/

Then make it bigger or smaller, IN YOUR BROWSER.

My browser shows the right size, for me, in the following conditions:

1. Any web site that doesn't attempt to specify either the typeface or the
font size for main text. Wonderful! (My own web sites work on that principle).

2. The majority of web sites in the world; they specify both the typeface and
the font size, and choose a font size based on the aspect value of the chosen
typeface compared with my default of 16px "Times New Roman".

I don't know of any method of dealing with a web site that screws up users by
overriding their choice of typeface without making a corresponding adjustment
to the font size. If you know of one, please tell me. My browser is right. So
Brian's pages are just incompatible with my properly-adjusted browser.

Brian appears to have a minority view that an author can/should change the
typeface without making a corresponding adjustment to the font size. How many
others here hold that view view? I've have discussed this at the following
pages:
http://www.barry.pearson.name/articles/text/
http://www.barry.pearson.name/articl...comparison.htm
http://www.barry.pearson.name/articles/text/trends.htm
http://www.barry.pearson.name/articl...ect_values.htm

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #57

Phil Evans wrote:

Barry Pearson wrote:
Which browser would enable me to see 16px Times New Roman when an
author has not specified a main text size, but 13px Arial if the
author has specified "Arial" without a size? I suspect *none*.
None that I know of, but Firefox will allow you to sidestep problems
with individual sites using the URIID extension (see
http://extensionroom.mozdev.org/more-info/uriid)

It allows you to set site-specific CSS, which is *very* useful for
badly-designed sites. It's already overriding many instances of
microfonts on my system.

Hm! Thanks for that information. But I would prefer not to deal with cases
like Brian using that method. I would prefer to just see things about the
right size because the author has sorted it out.
To pre-empt one criticism - yes, this is currently far from
user-friendly; it's only for web designers and other people who
understand CSS. However, the mere fact that it exists and is being
used implies demand - can we really be far from a "remember view
settings for this web site" button, or similar?

Jul 20 '05 #58

kchayka wrote:

Barry Pearson wrote:

The snowball effect
has *already* happened. The result is "13px Arial".

Huh? I have no idea how you came to this conclusion, but no matter.
The longer you drag a thread out, the more you sound like either a
troll or a lunatic. I give up.

[snip]

Here are a couple of pages that should show the basis for what I said:
http://www.barry.pearson.name/articles/text/trends.htm
http://www.barry.pearson.name/articl...ect_values.htm

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #59

Darin McGrew

Barry Pearson <ne**@childsupportanalysis.co.uk> wrote:

Indeed they do. And the obvious thing to do is to work out what
font-size-adjust (which is anyway flawed) would say, then make this explicit.

But you don't know "what font-size-adjust would say". And you don't know
what a hypothetical unflawed font-size adjust replacement would say either.

All the author can know is the aspect value of his preferred font. (Well,
the aspect value of one implementation of his preffered font, but why
quibble.) To know what font-size-adjust (or its replacement) would say, you
need more information, such as the aspect value of whatever font is
actually used, and possibly the aspect value of the reader's default font.

There is no way to "make this explicit" if you don't know what "this" is.
--
Darin McGrew, mc****@stanfordalumni.org, http://www.rahul.net/mcgrew/
Web Design Group, da***@htmlhelp.com, http://www.HTMLHelp.com/

"Nice is different than good." - Sondheim (LRRH, "Into the Woods")

Jul 20 '05 #60

Brian

Barry Pearson wrote:

Brian wrote:
Barry Pearson wrote:
I object to the font size on the following pages:
http://www.julietremblay.com/brian/
Very good. One person has complained about font size being too
large. That person, however, is a web author with an ax to grind,
and is hard to distinguish from a troll.

I wonder how many people actually know about that web site?

Probably not many. (Though, strangely enough, it shows up very close
to top in Google for my name.)
What proportion of them all am I?
What difference does that make? You're still indistinguishable from a
troll.
I ask you to have a look at the following page,
I think I'll pass. You try awfully hard to be some sort of web
academic, sure that you're breaking new ground in authoring
philosophy, and apparently unaware that everyone around you has
already covered that ground, decided it wasn't worth many mental
resources, chose a solution, and moved on.
How about showing us a screen-shot, from your system, of the table
at the top, so that we can see what you see when authors specify
typefaces but don't specify font sizes?
I see my chosen size. What else would I see?
Let's try to advance our knowledge by sharing our separate views of
tricky pages!
Oh, let's not.
Therefore there isn't one single authoring approach that works for
all users. We are all just discussing our opinions of what to do.
We are all having to make compromises. We just disagree about which
are the best compromises.

I've heard this before, in your endless -- and pointless -- diatribes
about table layouts. The solution is obvious, except to someone
desperate to demonstrate what a smart and thoughtful guy he is.

--
Brian (remove ".invalid" to email me)
http://www.tsmchughs.com/

Jul 20 '05 #61

JRS: In article <gt**********@newsfe2-gui.server.ntli.net>, seen in
news:comp.infosystems.www.authoring.stylesheets, Barry Pearson
<ne**@childsupportanalysis.co.uk> posted at Fri, 11 Jun 2004 20:13:04 :

Second, I have been examining various approaches to this issue, ("what should
an author specify for the main text size?") I have come the conclusion that
there are 3 main approaches:
http://www.barry.pearson.name/articles/text/

1. "Don't specify either font or font-size for main text". This theory assumes
that the user's own preferences, for example determined by the browser, should
be used for main text.

2. "Specify font, but not font-size, for main text". This is hopefully a
minority theory. The view is that it is OK to override the user's font
preference, but not the user's font-size preference. Its logic is seriously
flawed.

3. "Specify both font and font-size for main text". This appears to be the
adopted theory of the majority of web sites in the world.

Re 3.: That is because authors are a domineering lot, who think they
know better than their readers ...
The choice should be allowed to depend on the nature of the page and of
the bulk of its information content.

A page which is like that of a reputable broadsheet newspaper is a page
predominantly of text, with headings for convenience, pictures for
decoration, and advertisements for ignoring - for that, the ordinary
text should be fonted, face and size, according to the reader's defaults
(if the headers are larger and of a reasonable font-face, they will be
acceptable). That's your Number One, and my policy.

A page which is more or less like that of an advertisement in such a
paper is conceived of as an artistic entity, though when it appears to
the world it may be somewhat changed; it is not designed to be read as
literature, and can rightly be fonted for impact (but any bulk text
within the page should be based on the user's preference).

Your point 2 can perhaps be enlarged, considering only bulk-text parts
and only font faces suited to bulk text :

In a browser set so that default text is, in font face and size, most
convenient to the reader, ISTM that any other bulk-text font face at the
same nominal size will be of tolerable size to the user, so that setting
face but not size is tolerable. Tolerable, however, is a low standard.

ISTM that it may be possible to determine the overall size of an average
character (or of a block of text) in the default font face/size, to
select another font, and to choose, by code, the size of the new font to
match the aforesaid overall size.

You'd then have

2a. Override face, retain point size - tolerable;
2b. Override face, adjust point size to retain physical size - good?.
--
© John Stockton, Surrey, UK. ??*@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
Web <URL:http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.

In MS OE, choose Tools, Options, Send; select Plain Text for News and E-mail.

Jul 20 '05 #62

Brian wrote:

Barry Pearson wrote:

[snip]

What proportion of them all am I?

What difference does that make? You're still indistinguishable from a
troll.

If you believed that, you wouldn't be replying!

I ask you to have a look at the following page,

I think I'll pass. You try awfully hard to be some sort of web
academic, sure that you're breaking new ground in authoring
philosophy, and apparently unaware that everyone around you has
already covered that ground, decided it wasn't worth many mental
resources, chose a solution, and moved on.

You are obviously entitled to choose not to look at new material! But, equally
obviously, you then can't possibly have a clue about whether it was breaking
new ground.

And, of course, the solutions that the majority of authors in the world chose
are very different from what you say! Perhaps 99% table-layout. Mostly
text-sizes for main text based on px values. *That* is the solution they
chose, before they moved on. I am trying to identify what should come next.

How about showing us a screen-shot, from your system, of the table
at the top, so that we can see what you see when authors specify
typefaces but don't specify font sizes?

I see my chosen size. What else would I see?

[snip]

As I said: "How about showing us a screen-shot, from your system, of the table
at the top, so that we can see what you see when authors specify typefaces but
don't specify font sizes?"

Then *we* can see.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #63

Darin McGrew wrote:

Barry Pearson <ne**@childsupportanalysis.co.uk> wrote:
Indeed they do. And the obvious thing to do is to work out what
font-size-adjust (which is anyway flawed) would say, then make this
explicit.
But you don't know "what font-size-adjust would say". And you don't
know what a hypothetical unflawed font-size adjust replacement would
say either.

These pages gives some clues about "what font-size-adjust would say". I
suspect they provide some information that hasn't been published on the web
before.
http://www.barry.pearson.name/articl...ect_values.htm
http://www.barry.pearson.name/articl...comparison.htm

I am currently investigating how to correct the more obvious problems with
"font-size-adjust". It appears that width needs to be taken into account, as
well as height.
All the author can know is the aspect value of his preferred font.
(Well, the aspect value of one implementation of his preffered font,
but why quibble.) To know what font-size-adjust (or its replacement)
would say, you need more information, such as the aspect value of
whatever font is actually used, and possibly the aspect value of the
reader's default font.

There is no way to "make this explicit" if you don't know what "this"
is.

See the above pages. For example, see how the application of
"font-size-adjust" to "16px Times New Roman" (a typical browser default)
suggests values such as "14px Arial" and "13px Verdana". Which is close to
what many authors specify.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #64

Alan J. Flavell

On Sat, 12 Jun 2004, Barry Pearson wrote:

As I said: "How about showing us a screen-shot, [...] Then *we* can see.

That was a bit disingenuous of you, you know. Does the term "petitio
principii" ring any bells?

If you take a pixel-based screen-shot from /your/ browser, at the ppi
setting that /you/ use, taking into account /your/ eyesight whatever
it may be; and we view that image on our displays, pixel for pixel, at
the ppi settings which we use, and with the eyesight that we have,
then there's no guarantee at all that "we will see" whatever fine
distinction it was that you saw.

And that's a substantive part of this whole discussion. I've had cases
of folks complaining "far too big" about something to that, to me, was
uncomfortably small - they were using less than 72ppi while I was
using (on my desktop) 135ppi. That was because its size had been
specified in pixels - or rather, in CSS px units, but neither of our
browsers actually implement CSS px units, they simply treat them as
physical pixels. Verstehe?

Jul 20 '05 #65

Alan J. Flavell wrote:

On Sat, 12 Jun 2004, Barry Pearson wrote:
As I said: "How about showing us a screen-shot, [...]
Then *we* can see.

That was a bit disingenuous of you, you know. Does the term "petitio
principii" ring any bells?

Of course it does. And this doesn't resemble that! There are a variety of
different outcomes that could be revealed here. Evidence is needed.

For example, suppose Brian's screen-shot of my table showed that "100% Arial"
was exactly the same as his default. (Remember - his own pages are,
effectively, that specification). This would show that, while he has reason to
believe that his specification was OK, he was perhaps being led to that
conclusion by his own set-up.

Obviously, I am being led to my own conclusions about my web sites by my own
parameters. I recently asked for a site check on a web site of mine that
looked good to me, and found problems in other people's systems. That was
vitally important information. I get a lot of useful feedback from
screen-shots from other systems. I regularly use the iCapture site. I have got
useful information from screen-shots taken from IE/Mac systems.

Anyone wondering why I specify that my button-links are "80% Arial (etc)"
would get an insight from my screen-shot. I want to know why other people do
things that, to me, are clearly flawed, such as specifying a main-text
font-family without an accompanying font-size. We can all gain a lot by
sharing such screen-shots. I went first!
If you take a pixel-based screen-shot from /your/ browser, at the ppi
setting that /you/ use, taking into account /your/ eyesight whatever
it may be; and we view that image on our displays, pixel for pixel, at
the ppi settings which we use, and with the eyesight that we have,
then there's no guarantee at all that "we will see" whatever fine
distinction it was that you saw.
Some things will be obvious! That is why I put the background image in place,
with the measuring lines at 10 pixel intervals, and the bluish lines every 50
pixels. I expect that the results won't be subtle - they will be clearly
visible. I'm an engineer- I want to be able to measure things objectively.

If you haven't looked at the page, you may not realise its nature. Have a
look:
http://www.barry.pearson.name/articl...comparison.htm
And that's a substantive part of this whole discussion. I've had cases
of folks complaining "far too big" about something to that, to me, was
uncomfortably small - they were using less than 72ppi while I was
using (on my desktop) 135ppi. That was because its size had been
specified in pixels - or rather, in CSS px units, but neither of our
browsers actually implement CSS px units, they simply treat them as
physical pixels. Verstehe?

I am trying to measure things objectively, to cut through (and out) such
concepts as "too big". 1% too big? 20% too big? So what?

By policy, I run my browsers at their default settings when examining my own
pages for the first time. I assume that most of my audience will be doing
this. I grew out of my "blame the user" attitude a long time ago.

The key point for any such investigation is "what are the facts?" Numbers (I
like numbers!), references, quotes, evidence. Things that we can all agree
upon without argument, so that we can build on it. Our disagreements should be
about matters of opinion & aims, not about the basic facts!

I would welcome screen-shots from you too.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #66

Shawn K. Quinn wrote:

Barry Pearson wrote:

[snip]

I suspect I am in the majority. And that is why the vast majority of
authors on this planet cater for my requirements, and those of
people like me.

Your requirements are different from mine, and one of the ideas of
the World Wide Web is that the user can configure the browser to suit
his/her own needs. Platform- and device-independence, and all that.

[snip]

True. But we should surely not insist that users do this. Many are not really
capable. Some of my audience use PCs in public libraries, or schools, or other
people's PCs.

I suspect that browsers come with a fairly predictable default configuration.
Eg. 16px Times New Roman for the main text. The trick is to make our web sites
work for such users without significant effort, then enable them to get
something better if they have the knowledge & technology & circumstance to do
so.

What CSS rules will be suitable for a majority of users using their browsers
(indeed, their systems) at their default settings?

Then, how should we enable *aware* users to achieve better results?

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #67

Alan J. Flavell wrote:

On Fri, 4 Jun 2004, Barry Pearson wrote:

[snip]

Not only should %s and font-families be set together, but whenever
text rules change, the new % and the new font-family should have a
known relationship with the original. (Otherwise, what does %
actually mean?)

Which is why we need some enhanced version of font-size-adjust.
Meantime, we need to make the best use of what we have.

I believe we are not about to get such an enhanced version. But it would be
interesting to know what it would look like. I'll review what I understand
about it, in case there is something better in the pipe-line.

W3C:

This appeared in CSS2. It uses the ratio of the "x" height and the font-size
to derive an "aspect value". The author can specify the aspect value of the
first family-name in the font-family list, and the browser is supposed to
ensure that if it uses a later font instead, it adjusts its size accordingly.
There are some (acknowledged) errors in the CSS2 recommendation, including the
statement that the specified aspect value should be that of the first
family-name in the list. There may not be such a family-name. There may only
be the generic-family.

The current draft of CSS2.1 withdraws font-size-adjust. The current draft of
the CSS3 Font Module corrects the acknowledged errors in the CSS2
recommendation. This is more logical, but it is still a flawed concept.

Browsers:

IE & Opera 7.23 don't support it. Firefox 0.8 & Netscape 7.1 support it. They
use the CSS3 version, and apply it even to the first family-name.

Typographical credibility:

I'm not convinced that "aspect value" is a recognised typographical concept. I
suspect that it was invented for the final CSS2 recommendation. "Aspect value"
didn't exist before the final version - earlier versions used the inverse
(called "z value").

The CSS2 "casually" quotes aspect values for various fonts. For example, 0.58
for Verdana. These values appear a lot on the web, but I believe they are not
inherent properties of the design of these fonts, but instead were calculated
for W3C's CSS2, then copied a lot without checks. For example, on my PC the
value is 0.55 for Verdana. My guess is that browsers don't extract it as a
property of the font, but evaluate it on demand, and perhaps then cache it.

Effectiveness:

"Better than nothing", but certainly not right. Fonts such as Georgia, and
especially Verdana, have a width that is not reflected in the aspect value.
And, even more, so do fixed-pitch fonts. "Courier New" gets *worse* if you
apply font-size-adjust! In any complete scheme, width *must* be taken into
account. And the width of "x" is not a good indicator of the width of the font
as a whole.

I've tried to identify a better value. For example, should it be the "volume"
of an "o" including its character-spacing? (Or, rather, the square-root of the
volume). That makes things a little better, but still doesn't give good
line-lengths.

Use for main text sizes:

I can't see how to use font-size-adjust to solve the problems being discussed
here. How would it be used to ensure that the text in button-links matched the
main text? Perhaps I don't know enough about how CSS font property inheritance
works to solve this.

It may be more useful in browsers. Perhaps I would be able to say "my default
is 16px "Times New Roman" with an aspect value of 0.45". Then a web site that
specified "Arial" (or "100% Arial") would have its font-size adjusted
accordingly, to about 13px or 14px Arial. But there are lot of unanswered
questions there.

I have written a page on this topic:
http://www.barry.pearson.name/articl...ect_values.htm

Views?

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #68

Alan J. Flavell wrote:

On Sat, 5 Jun 2004, Barry Pearson wrote:
Darin McGrew wrote:
>
> If users complain that their own browsers' font is too small, then
> your best bet may be to refer them to a page like this one:
> http://www.xs4all.nl/~sbpoley/main/adjust.html
The complaint *wasn't* that their own browser's font was too small.

Not in so many words, no: but that was the gist of it, nevertheless.
It was that *my* main text size was too small.

And since you know that you didn't size that text, you have to
conclude that it was displayed at the user's configured preference
size, right?

Yes. But perception is important. It isn't good to be technically right but to
be perceived by the target audience to be wrong. And it doesn't help to tell
the user that they need to correct their system! If someone links to one of my
sites from another one, then links to yet more sites from mine, I don't want
to be the odd one out. History shows plenty of examples of the best
(technically) product being the loser. What is the merit in that?

I seek technically-good pages that also work well in the user's context. (And
I'm sure you recognise that this is what I am trying to do). But, frankly, if
forced to make a decision, I would go for the pages that were most immediately
acceptable to the audience, rather than those that were technically best. I'm
trying to communicate with my audience, not win prizes in the "mark-up of the
year" award.

I don't want users looking at the site and believing that there is
something wrong with it.

There will always be someone to find fault with it. Any web author
has got to live with that.

The trick is to choose who finds fault. I can live with participants of this
newsgroup finding fault! But I don't want to piss-off my *real* audience.
Now that you've got some body text to work with, I feel more confident
to offer a report, and my view is that on my Mozilla the body text is
fine, while on IE6 by default it's too large. But that's OK because I
normally have IE6 cranked-down at least one notch (to "smaller") if
not two ("smallest") to compensate for its unnecessarily large default
text size. Of course when idiot web authors come along and crank it
down for me, assuming that I would be too dim to be able to do that
for myself, then we get microfonts. But you didn't do that.
Thank you for that valuable response. I feel that I am not doing something
dramatically wrong, but some users will still be critical.

[snip] But it's not solved by changing to units which (a) don't scale in the
way that you need and (b) aren't implemented to specification and (c)
if they were implemented to specification they still wouldn't be what
users in general need. Those units might (thanks to a combination of
bugs) seem to produce the end result you wanted, on the displays you
use, and with your particular eyesight capability, but they still
aren't the right choice for WWW use, and it's not just me that's
saying so.
I see 2 webs. One is described & desired by experts. The other exists in the
real world, published as millions of web sites, and accessed by 100s of
millions of users, or more.

"The right choice" may be a compromise driven by those users & publishers, not
an unrealisable ideal driven by those experts.
Sure, those with special needs can override them, but to my mind
there's something comfortable about choosing units which are *meant*
to cascade with the users own needs, rather than forcing users into a
position where they have to fight back by overruling the author
completely. I apply that verdict not only to pt units, but also, for
the most part, to px units, when we're in the context of WWW display.

[snip]

I suspect that many, perhaps most, users in the world do not want to fight
back. I suspect that most users are happy with what publishers do. For most
users, "13p Arial" may be a good main text size.

Things may change in future. But at the moment I suspect that it is the
experts, not the majority of users, who have the complaints. Perhaps the
majority of users simply adapt. And then future publishers simply conform to
what the users have adapted to. And browser developers adapt to both of these,
by default.

See:
http://www.barry.pearson.name/articles/text/trends.htm

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #69

In article <nF**************@newsfe3-win.server.ntli.net>,
ne**@childsupportanalysis.co.uk says...

BBC: 13px Verdana

Perhaps someone can answer a quesion about this.

We've always been told that declaring web font size in pixels or points is
poor practice on the screen, because it prevents the user from resizing it.
Screens and windows do vary in size, so dont mess with the users font size,
which sounds right to me. So I went to sleep and havent thought about this in
a long time.

But in fact, the BBC 13px text can be resized (at least in IE6 and NS7). So
in this case there seems less objection to the "fixed" size, since it isnt
actually fixed (although their default is too small for me, I wish they had
left it alone).

But I am curious about how they are are even able to do this? Is their
JavaScript able to react to browser font size changes? (which if possible,
sure seems the hard way).

Or is it a change in how current browsers work, so that px is no longer fixed
size? (NS4 does remain fixed size there, and even smaller text yet for me).
If so, this is a welcome improvement in browsers.

Jul 20 '05 #70

Darin McGrew

Wayne Fulton <no****@invalid.com> wrote:

But in fact, the BBC 13px text can be resized (at least in IE6 and NS7). So
in this case there seems less objection to the "fixed" size, since it isnt
actually fixed (although their default is too small for me, I wish they had
left it alone).

AFAICS, the BBC site doesn't specify fonts in px. Rather, it uses HTML:
 and .

So it doesn't specify fixed-size fonts. It merely specifies fonts that are
one or two sizes smaller than the reader's normal font.
--
Darin McGrew, da***@TheRallyeClub.org, http://www.TheRallyeClub.org/
A gimmick car rallye is not a race, but a fun puzzle testing your
ability to follow instructions. Upcoming gimmick car rallye in
Silicon Valley: The Suppranos (Saturday, July 3)

Jul 20 '05 #71

In article <ca**********@blue.rahul.net>, mc****@stanfordalumni.org says...

AFAICS, the BBC site doesn't specify fonts in px. Rather, it uses HTML:
 and .

So it doesn't specify fixed-size fonts. It merely specifies fonts that are
one or two sizes smaller than the reader's normal font.

You could be right, but I am confused by it and not sure.
I was looking at the first random convenient news story page:

http://news.bbc.co.uk/1/hi/uk_politics/3803505.stm

and yes, there are some font size="1" in the html for some elements,
like the top header line: Home TV Radio Talk, Where, A-Z

But its css file http://news.bbc.co.uk/nol/shared/css/news.css says:

body{background-color:#FFFFFF;border:0px;margin:0px;font-family:Verdana,Aria
l,Helvetica,sans-serif,"MS sans
serif";font-size:13px;font-weight:normal;color:#000000;}

Also 13px there for table, and tr, etc, and some 10px, but it seems like
everything is px in the css file.

Oddly, the columns of overhead and menus dont change size, but the news text
does. (the top header line is "1", and it changes size too). It seems too
disorganized for me to figure out what applies to what. I suppose that news
text is in some element I have not found yet.

Jul 20 '05 #72

Wayne Fulton wrote:

In article <nF**************@newsfe3-win.server.ntli.net>,
ne**@childsupportanalysis.co.uk says...
BBC: 13px Verdana
Perhaps someone can answer a quesion about this.

[snip] But in fact, the BBC 13px text can be resized (at least in IE6 and
NS7). So in this case there seems less objection to the "fixed"
size, since it isnt actually fixed (although their default is too
small for me, I wish they had left it alone).
The main text looks about the same size as my browser-default, which is 16px
"Times New Roman". That suits me. A lot depends on how browsers are set up.
Someone with a browser-default of (say) 16px Arial would see the main text too
small.

[snip] Or is it a change in how current browsers work, so that px is no
longer fixed size? (NS4 does remain fixed size there, and even
smaller text yet for me). If so, this is a welcome improvement in
browsers.

In response to you & Darin. What appears to happen here is that the BBC
specifies:

body{
background-color:#FFFFFF;
border:0px;
margin:0px;
font-family:Verdana,Arial,Helvetica,sans-serif,"MS sans serif";
font-size:13px;
font-weight:normal;
color:#000000;
}

For some pages, such as the initial news page, this has the result that it is
tricky to get IE to resize the text. Then, in *addition*, on (some? all?)
article pages the main text is surrounded by:

 ... 

This appears to leave the text sized the same in my browsers by default, but
enables IE to resize via the View > Text Size menu. Stripped to its simplest,
the same effect appears with the following. The "Second paragraph" can easily
be made larger/smaller in IE, while the "First paragraph" is tricky:

HTML:

First paragraph 
Second paragraph

CSS:

p {
font: 13px Arial;
}

A rather curious way of using to achieve something useful! (That is new
to me - I hadn't realised those consequences of mixing CSS & HTML
presentation. Hm!)

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #73

In article <A8ezc.179$lC3.9@newsfe3-gui>, ne**@childsupportanalysis.co.uk
says...

For some pages, such as the initial news page, this has the result that it istricky to get IE to resize the text. Then, in *addition*, on (some? all?)
article pages the main text is surrounded by:

 ...

Ah so, thanks Barry. That must be what Darin meant, but I had missed that.
It seems clear now, at least the what if not the how or why (or why the px at
all). It sure does seem the hard way, but it obviously works much better than
px alone. It must have been a quick patch to recover from the px being
unacceptable. I didnt realize that html font overrides the css px.
Interestingly, this case doesnt override in NS4.

That also explains why the table font is not affected by the resize. In the
days before css, I intentionally always used just one global font face at top
of body (in a similar way), with the effect that the one declaration applied
everywhere (until changed, then the change continued to apply everywhere).
Except unfortunately, table cells required font declarations in every cell.
That global concept was considered wrong practice, wasnt supposed to apply
globally and validators fussed, but it must be like the browser default is
global too. It always worked great (except tables) and I considered it quite
elegant. <g> Anyway, that explains the tables not resizing in this case.
This BBC page has substantial validation problems too.

Jul 20 '05 #74

Wayne Fulton wrote:

In article <A8ezc.179$lC3.9@newsfe3-gui>,
ne**@childsupportanalysis.co.uk says...

For some pages, such as the initial news page, this has the result
that it is tricky to get IE to resize the text. Then, in *addition*,
on (some? all?) article pages the main text is surrounded by:

 ... 

Ah so, thanks Barry. That must be what Darin meant, but I had missed
that. It seems clear now, at least the what if not the how or why (or
why the px at all). It sure does seem the hard way, but it obviously
works much better than px alone. It must have been a quick patch to
recover from the px being unacceptable. I didnt realize that html
font overrides the css px. Interestingly, this case doesnt override
in NS4.

[snip]

Yes, I wondered why they did it. I suspect you are right about it being a
quick patch. But using to override CSS is surely very unusual!

Why 13px to start with? We appear to have arrived at a state where many,
perhaps most, web sites on the planet use about this size. I think there is
some sort of reinforcement among users, authors, and browsers. I'm not sure
what came first, but I suspect something like this:

If most users are using browsers that default to 16px Times New Roman, what
should an author use who wants to use a common sans-serif font? The following
look about the same practical size as 16px Times New Roman: 13px Arial; 80%
Arial; 0.8em Arial; 12px Verdana. They are all typical web site
specifications.
http://www.barry.pearson.name/articl...ect_values.htm
http://www.barry.pearson.name/articl...comparison.htm
http://www.barry.pearson.name/articles/text/trends.htm

If most web sites specify such a range of sizes, what should browsers have as
their factory-settings? The only one that now makes sense appears to be: 16px
Times New Roman! (Although 15px Georgia looks about the same size as 16px
Times New Roman, it might make 80% Arial or 0.8em Arial look too small, more
like 12px Arial). If browsers used 16px Arial, (or especially 13px Arial), and
users did whatever was necessary to make this size comfortable for them, then
the following (common) web site specifications will look too small: 80% Arial;
0.8em Arial. So they are probably not good factory-settings for browsers.

So, now, what motivation is there for users to change their browser settings?
Most probably don't know how to anyway. But they now mostly see sans-serif
text that looks about the same size as their browser settings. For sites,
(like mine), that doesn't specify font parameters at all for main text, the
user will see the 16px Times New Roman, which is about the same practical
size. If this is too small for them, perhaps they have to increase the size
some other way. For example, I believe many users use their systems configured
with fewer, hence larger, pixels on the CRT monitors. Perhaps they run the
monitor at 1024x768 instead of 1280x1024.

This is like an "Evolutionary Stable Strategy". Any divergence by users,
authors, or browsers is likely to be less than ideal relative to most of the
others. A web site that specifies 100% Arial has text that is too big for most
users. Users who change their browser settings are likely to find that either
13px Arial or 80% Arial looks wrong, perhaps too small. And so it goes!

Any theory of what, if anything, to specify for a web site needs to take into
account what most other web sites do, what most users do, and what most
browsers do. Otherwise, it may look poor, in spite of appear to be an
impecable theory.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #75

In article <h4**************@newsfe1-gui.server.ntli.net>,
ne**@childsupportanalysis.co.uk says...

Any theory of what, if anything, to specify for a web site needs to take into
account what most other web sites do, what most users do, and what most
browsers do. Otherwise, it may look poor, in spite of appear to be an
impecable theory.

Font face size differences are not much issue with me, so I'm not sure poor is
always the right word. It seems more a choice, and choices are offered. For
example, sites much like the BBC and busier need the small side columns and
menus because they feel the need to pack the page so tightly, kitchen sink
design. But others that simply present volumes of text in a plain fashion
probably do want a larger font face and lots of margin, for readability. These
are different needs, and choice is good, to a degree. I assume this is the
purpose of those larger fonts, to be larger. Not all 12 point fonts print
exactly 6 lines per inch on paper either.

But use of fixed dimensions is a different issue than face size. Web site
designers ought to learn that what they see is not what others see.

Jul 20 '05 #76

JRS: In article <3I2zc.440$Ee2.196@newsfe4-gui>, seen in news:comp.info
systems.www.authoring.stylesheets, Barry Pearson <news@childsupportanaly
sis.co.uk> posted at Sun, 13 Jun 2004 20:16:22 :

I seek technically-good pages that also work well in the user's context
... The trick is to choose who finds fault. I can live with participants of this
newsgroup finding fault! But I don't want to piss-off my *real* audience.

http://www.barry.pearson.name/articles/text/ has that effect on me.

In it, text font face & size is invariably as I wish (as you know).

But, in MSIE4 with a window 640 px wide :

The first ~170 px of each line are wasted, once below the box containing
(c) Copyright. That column could instead be a top row.

In the menu, top left, the last character of each entry needs to be
followed by a visual space of about an em or en; you could shift a
couple of   in each case. I don't see the point of slapping a menu
over a pleasant picture.

The picture of two birds is too small; at first, I thought it was after
something painted by Goya.

In the top left of the main text area, there is a pink area about
170*400 px, serving no apparent purpose.

In the box containing "ALA Topics", and three others, there is no right
padding.

The first such box overlaps "What is this topic about?", which has one
word per line.

Source is easier to read (and to post) if it has a nominal 72-character
right margin - granted, that may not be a target.

As regards the textual content : again, ISTM that there are different
classed of Web page, and that needs to be recognised in giving such
advice.

The third "have" should be "has".

--
© John Stockton, Surrey, UK. ?@merlyn.demon.co.uk Turnpike v4.00 IE 4 ©
<URL:http://jibbering.com/faq/> JL / RC : FAQ for news:comp.lang.javascript
<URL:http://www.merlyn.demon.co.uk/js-index.htm> jscr maths, dates, sources.
<URL:http://www.merlyn.demon.co.uk/> TP/BP/Delphi/jscr/&c, FAQ items, links.

Jul 20 '05 #77

Dr John Stockton wrote:

JRS: In article <3I2zc.440$Ee2.196@newsfe4-gui>, seen in
news:comp.info systems.www.authoring.stylesheets, Barry Pearson
<news@childsupportanaly sis.co.uk> posted at Sun, 13 Jun 2004
20:16:22 :

I seek technically-good pages that also work well in the user's
context ...
The trick is to choose who finds fault. I can live with participants
of this newsgroup finding fault! But I don't want to piss-off my
*real* audience.

http://www.barry.pearson.name/articles/text/ has that effect on me.

Thank you for the following constructive comments.
In it, text font face & size is invariably as I wish (as you know).

But, in MSIE4 with a window 640 px wide :

The first ~170 px of each line are wasted, once below the box
containing (c) Copyright. That column could instead be a top row.
I wrestle with this issue daily. Should the site navigation go on the side, or
at the top, (or elsewhere)? In another site, I have enabled the choice even to
be on a page-by-page basis. Have a look at these 2 pages, which are identical
except for the ID in the <body>, (which decides the CSS rules that apply). In
this case, the idea is that the "top 9" index pages (those with buttons),
have the site navigation on the left, and all the "bulk content" pages have it
at the top.
http://www.kingsnorton.info/
http://www.kingsnorton.info/galleries/elsewhere.htm

I have lots of pages that work one way, and lots the other way. Please just
accept that I am sympathetic to what you say, and haven't ruled out the
possibility of moving that way. I think there is a lot of merit in having some
"outer" pages that work one way, before getting in to weighty pages that use
different principles.
In the menu, top left, the last character of each entry needs to be
followed by a visual space of about an em or en; you could shift a
couple of   in each case. I don't see the point of slapping a
menu over a pleasant picture.
That suggests that you are viewing under considerably different conditions
from me. I hadn't realised there was such a problem. (I already use some
 s. I will consider using more).
The picture of two birds is too small; at first, I thought it was
after something painted by Goya.
It is simply the thumbnail (125 pixels wide) that I use for that picture. It
is a link rather than a picture in its own right. Isn't its nature as a link
clear on your system?
In the top left of the main text area, there is a pink area about
170*400 px, serving no apparent purpose.
See above.
In the box containing "ALA Topics", and three others, there is no
right padding.
Hm! The CSS rule for that box is as follows. It should have a padding of 1em
all round! At the moment, I don't accept responsibility for your lack of
padding on the right!

div.sidenote {
float: right;
width: 30%;
margin: 0.5em 0 0.5em 1em;
padding: 1em;
border-style: solid;
border-width: 1px;
}
The first such box overlaps "What is this topic about?", which has one
word per line.
Wow! This is telling me that you are viewing under extreme conditions. There
should be no overlap. The box is floated to the right. If the browser can't
put the text to the left of the box, surely it should drop it down below the
box? I know of no way to stop an errant browser overlapping elements.
Source is easier to read (and to post) if it has a nominal
72-character right margin - granted, that may not be a target.
Sorry, I don't understand that statement.
As regards the textual content : again, ISTM that there are different
classed of Web page, and that needs to be recognised in giving such
advice.
Suggestions? I thought of the higher-level index pages, (those identified by
the buttons of the site navigation), and others. (I also have different CSSs
for academic papers).
The third "have" should be "has".

Thanks. Corrected.

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #78

Shawn K. Quinn

Barry Pearson wrote:

Alan J. Flavell wrote:
On Sat, 5 Jun 2004, Barry Pearson wrote:
It was that *my* main text size was too small.
And since you know that you didn't size that text, you have to
conclude that it was displayed at the user's configured preference
size, right?

Yes. But perception is important. It isn't good to be technically right
but to be perceived by the target audience to be wrong. And it doesn't
help to tell the user that they need to correct their system!

In a lot of cases, such as the idiotic "best viewed with NetXploder" or
"best viewed at [some arbitrary screen resolution]" buttons, this is a
valid point. In the case of text sized at the user's default font and size,
it shouldn't need any correcting, but if you want to teach the user how to
use their respective browsers, you can do so (but please DO NOT write the
instructions in a manner that assumes *everyone* is using Internet Explorer
on Microsoft Windows $CURRENT_VERSION_THING (where the value of
CURRENT_VERSION_THING is currently "XP"), as I have seen way too many sites
do).
If someone links to one of my sites from another one, then links to yet
more sites from mine, I don't want to be the odd one out.

Not everyone goes to McDonald's for a hamburger, either.

--
Shawn K. Quinn

Jul 20 '05 #79

Shawn K. Quinn wrote:

Barry Pearson wrote:

[snip]

Yes. But perception is important. It isn't good to be technically
right but to be perceived by the target audience to be wrong. And it
doesn't help to tell the user that they need to correct their system!

In a lot of cases, such as the idiotic "best viewed with NetXploder"
or "best viewed at [some arbitrary screen resolution]" buttons, this
is a valid point. In the case of text sized at the user's default
font and size, it shouldn't need any correcting, but if you want to
teach the user how to use their respective browsers, you can do so
(but please DO NOT write the instructions in a manner that assumes
*everyone* is using Internet Explorer on Microsoft Windows
$CURRENT_VERSION_THING (where the value of CURRENT_VERSION_THING is
currently "XP"), as I have seen way too many sites do).

[snip]

I think Stephen Poley has set a standard that there is now no excuse for
falling below:
http://www.xs4all.nl/%7Esbpoley/main/adjust.html

But this still makes any such web site look like a do-it-yourself kit. (All my
web sites work like that, so I am criticising them too).

Some of my audience don't even have their own computers. They use other
people's, or systems in public libraries. (I don't get any income from those
people, but they *are* my audience!) I would like to know just what proportion
of users feel comfortable & willing to make such changes. And why?

People posting to *this* newsgroup must surely be among the least
representative users on the planet! Where is the research on this topic from
proper researchers?

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #80

JRS: In article <brmzc.725$fE6.430@newsfe6-win>, seen in news:comp.info
systems.www.authoring.stylesheets, Barry Pearson <news@childsupportanaly
sis.co.uk> posted at Mon, 14 Jun 2004 19:44:16 :

In the menu, top left, the last character of each entry needs to be
followed by a visual space of about an em or en; you could shift a
couple of   in each case. I don't see the point of slapping a
menu over a pleasant picture.

That suggests that you are viewing under considerably different conditions
from me. I hadn't realised there was such a problem. (I already use some
 s. I will consider using more).

Well, it is MS IE 4. The window is near enough half the width of a
1280*1024 screen, i.e. about 640 px wide.

The picture of two birds is too small; at first, I thought it was
after something painted by Goya.

It is simply the thumbnail (125 pixels wide) that I use for that picture. It
is a link rather than a picture in its own right. Isn't its nature as a link
clear on your system?

Not until I happen to move the mouse over it (when it moves a few px to
the right, and the cursor changes).

In the top left of the main text area, there is a pink area about
170*400 px, serving no apparent purpose.

See above.

? <URL:http://www.merlyn.demon.co.uk/barry-p.htm> is a temporary page
bearing a GIF produced by Paint from the Alt-PtrScr capture of the
display of your page. It is a fair representation except that your pink
has mostly changed to light purple, but the background in the small box
has gone to white - artefacts of GIFing.

In the box containing "ALA Topics", and three others, there is no
right padding.

Hm! The CSS rule for that box is as follows. It should have a padding of 1em
all round! At the moment, I don't accept responsibility for your lack of
padding on the right!

But perhaps you can for giving me a clue as to why I don't get right
padding in some boxes attempted on my site.

The first such box overlaps "What is this topic about?", which has one
word per line.

Wow! This is telling me that you are viewing under extreme conditions.

IMHO, 640px wide is reasonable; I generally edit my pages with editor &
browser side by side. I wonder whether I see a coloured block where you
do not?

Source is easier to read (and to post) if it has a nominal
72-character right margin - granted, that may not be a target.

Sorry, I don't understand that statement.

Using View Source, hence Notepad, I see each text paragraph of yours is
a single long line, wrapped by Notepad, and so obscuring the
indentation; if I put Notepad to 80 characters width, the same applies
to list elements. Perfectly legal, of course. And obviously
satisfactory to you when using your chosen editing tool(s).

But less convenient if you ever want to post it to News, and less
convenient for others with other tools to read the source. You are
perfectly entitled to be unconcerned.

I've found it convenient to work with a source right margin of 72
characters, except for where lines must be longer :
"http://www.
llanfairpwllgwyngyllgogerychwyrndrobwllllantysilio gogogoch.co.uk/"
needs to be on one line.

Most DOS tools are visually less satisfactory with lines over 80,
MiniTrue for example - and EDIT.

As regards the textual content : again, ISTM that there are different
classed of Web page, and that needs to be recognised in giving such
advice.

Suggestions? I thought of the higher-level index pages, (those identified by
the buttons of the site navigation), and others. (I also have different CSSs
for academic papers).

(a) Pages which are substantially ordinary text, though possibly with
illustrations; material such as is commonly found in books for the
ordinary reader.

(a') Pages such as might be published in learned journals; the audience
is specialised, and can be assumed - or required - to be intelligent and
industrious.

(b) Pages which are substantially pictorial, such as many
advertisements. Blocks of text should be treated as in (a); but it
might be safer from the art-designer point of view to render them in
graphics. A GIF of text is reliable, and may be no bigger than the
HTML/CSS needed to describe its effect. The aim of the coder is to
satisfy a designer, and perhaps also the readers.

(c) Pages which are auxiliaries; not useful in themselves, only as part
of the site. For example, pages wholly navigational.
Your page does repeat the term "main text", of course; but it might help
to insert in the first paragraph something specifically excluding pages
without a significant textual component.

--
© John Stockton, Surrey, UK. ?@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
Web <URL:http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
I find MiniTrue useful for viewing/searching/altering files, at a DOS prompt;
free, DOS/Win/UNIX, <URL:http://www.idiotsdelight.net/minitrue/> Update hope?

Jul 20 '05 #81

Dr John Stockton wrote:

JRS: In article <brmzc.725$fE6.430@newsfe6-win>, seen in
news:comp.info systems.www.authoring.stylesheets, Barry Pearson
<news@childsupportanaly sis.co.uk> posted at Mon, 14 Jun 2004 [snip]
That suggests that you are viewing under considerably different
conditions from me. I hadn't realised there was such a problem. (I
already use some  s. I will consider using more).

Well, it is MS IE 4. The window is near enough half the width of a
1280*1024 screen, i.e. about 640 px wide.

Thank you. Gosh - I never tested in IE 4. It is nothing like IE 5, for that
page!

The picture of the girl on the left is only revealed by "hover". The block of
pink/purple between the site navigation/admin and the main content doesn't
exist. There is no text overlap. I must be using CSS features that IE 4 simply
can't handle. (I had tested it with IE 5, IE 6, Netscape 7.1, Firefox 0.8, and
Opera 7.23. It looked as I wanted on those. It validates).

For interest, I just tested it with iCapture to check what it looked like with
Safari. Apart from Mac's anti-aliasing, it looks like it does on W2000. This
is the result (but only for 24 hours):
http://capture.danvine.com/results/52591.png

The picture of two birds is too small; at first, I thought it was
after something painted by Goya.
Your screen-shot doesn't show all the colours. Is that because your screen is
set for less than 24-bit colour (eg. 16-bit colour), or is it a result of the
GIF 256-colour pallette?

[snip] In the top left of the main text area, there is a pink area about
170*400 px, serving no apparent purpose.
See above - it doesn't appear on IE 5 or later. The text is adjacent to the
vertical line. I have no idea what I am doing that IE 4 can't handle. But ...
IE 6 has something called the "3 pixel jog" to the right of a floated element.
You appear to be getting the equivalent of a 170-pixel jog! Beats me.

[snip] ? <URL:http://www.merlyn.demon.co.uk/barry-p.htm> is a temporary
page bearing a GIF produced by Paint from the Alt-PtrScr capture of
the display of your page. It is a fair representation except that
your pink has mostly changed to light purple, but the background in
the small box has gone to white - artefacts of GIFing.
Very illuminating! I know that IE/Mac struggles with some of my pages. (I
don't have a Mac to test with). I have never tested any of them with IE 4. (I
test them all with IE 5).

[snip] The first such box overlaps "What is this topic about?", which has
one word per line.

Wow! This is telling me that you are viewing under extreme conditions.

IMHO, 640px wide is reasonable; I generally edit my pages with editor
& browser side by side. I wonder whether I see a coloured block
where you do not?

Yes. It looks as though IE 4 is beyond my limits. (I know that NN4 can't
handle some of my stuff either). I found it hard enough to get it to look
reasonable in IE 5, IE 6, Mozilla, & Opera. I doubt if I could sort out IE 4
as well! (Unless I used a layout table - I could probably make that more
cross-browser compatible).

Source is easier to read (and to post) if it has a nominal
72-character right margin - granted, that may not be a target.

Sorry, I don't understand that statement.

Using View Source, hence Notepad, I see each text paragraph of yours
is a single long line, wrapped by Notepad, and so obscuring the
indentation; if I put Notepad to 80 characters width, the same applies
to list elements. Perfectly legal, of course. And obviously
satisfactory to you when using your chosen editing tool(s).

OK, I understand. You are right - Dreamweaver does line-wrap at the boundary
of the editting window, so that isn't a problem for me.

[snip] I've found it convenient to work with a source right margin of 72
characters, except for where lines must be longer :
"http://www.
llanfairpwllgwyngyllgogerychwyrndrobwllllantysilio gogogoch.co.uk/"
needs to be on one line.

[snip]

I used to do that. But with a line-wrapping editor, any hard-new-line within
an element is redundant, and may interfere with my editting view. Yet another
web compromise.

Thanks for the input. I suspect the formula for supporting older browsers may
be: use layout tables, test with older browsers, and don't worry about
validation. With this page, I failed on all three!

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #82

JRS: In article <j7Mzc.210$%A3.86@newsfe3-gui>, seen in news:comp.infos
ystems.www.authoring.stylesheets, Barry Pearson <news@childsupportanalys
is.co.uk> posted at Tue, 15 Jun 2004 23:57:15 :

For interest, I just tested it with iCapture to check what it looked like with
Safari. Apart from Mac's anti-aliasing, it looks like it does on W2000. This
is the result (but only for 24 hours):
http://capture.danvine.com/results/52591.png
By chance, I had just seen iCapture cited in the June PCW magazine. It
kills my system. ISTM that those who know enough about HTML, CSS, etc.,
to write a public tester generally get carried away by their power and
use the latest IE4-crashing features in their testers (which, of course,
they do not need). This makes use difficult.

The picture of two birds is too small; at first, I thought it was
after something painted by Goya.

Your screen-shot doesn't show all the colours. Is that because your screen is
set for less than 24-bit colour (eg. 16-bit colour), or is it a result of the
GIF 256-colour pallette?

That will be the GIFing.

Yes. It looks as though IE 4 is beyond my limits.

Your page passes the most important tests.
It does not kill the browser
It is of reasonable size
The text can be read easily
The controls (IIRC) work.

Here's a trick which may or may not be obvious : my browser is set to a
local index page, containing links both local and remote. I've now
changed its <BODY> to <BODY onLoad="window.resizeTo(640, 960)">.

That means that not only does it open at the sort of width it was
developed for, but all pages opened directly or indirectly from it
inherit that size unless there is an intentional change; and any other
page (unless it sets its own size) can be re-sized by pressing the
browser's Home, Back. Pages open at a size compatible with A4.

By minor modification, other standard sizes could be selected.

--
© John Stockton, Surrey, UK. ?@merlyn.demon.co.uk DOS 3.3, 6.20; Win98. ©
Web <URL:http://www.merlyn.demon.co.uk/> - FAQqish topics, acronyms & links.
PAS EXE TXT ZIP via <URL:http://www.merlyn.demon.co.uk/programs/00index.htm>
My DOS <URL:http://www.merlyn.demon.co.uk/batfiles.htm> - also batprogs.htm.

Jul 20 '05 #83

Beauregard T. Shagnasty

Quoth the raven Dr John Stockton:

Here's a trick which may or may not be obvious : my browser is set
to a local index page, containing links both local and remote.
I've now changed its <BODY> to <BODY onLoad="window.resizeTo(640,
960)">.

I use this JavaScript in my local index page, rather than the <body> code:

<a href="javascript:resizeTo(240,400)">240x320</a> /
<a href="javascript:resizeTo(640,480)">640x480</a> /
<a href="javascript:resizeTo(800,600)">800x600</a> /
<a href="javascript:resizeTo(1004,728)">1024x768</a>

Adjust numbers accordingly. The first gives an approximation of a
mobile phone, taking my toolbars into consideration.

--
-bts
-This space intentionally left blank.

Jul 20 '05 #84

Dr John Stockton wrote:

JRS: In article <j7Mzc.210$%A3.86@newsfe3-gui>, seen in
news:comp.infos ystems.www.authoring.stylesheets, Barry Pearson
<news@childsupportanalys is.co.uk> posted at Tue, 15 Jun 2004 [snip]
Yes. It looks as though IE 4 is beyond my limits.

Your page passes the most important tests.
It does not kill the browser
It is of reasonable size
The text can be read easily
The controls (IIRC) work.

I specifically tried to ensure that, even if it couldn't be rendered right, it
would at least degrade to something readable. (If only by disabling CSS!)

I had assumed that IE 4 was pretty well superceded.
Here's a trick which may or may not be obvious : my browser is set to
a local index page, containing links both local and remote. I've now
changed its <BODY> to <BODY onLoad="window.resizeTo(640, 960)">.

That means that not only does it open at the sort of width it was
developed for, but all pages opened directly or indirectly from it
inherit that size unless there is an intentional change; and any other
page (unless it sets its own size) can be re-sized by pressing the
browser's Home, Back. Pages open at a size compatible with A4.

[snip]

Is this specifically so that you can easily test pages at different sizes? I
use Firefox with the Web Developer's Toolbar, and it is easy enough to
simulate one of the typical sizes. But I also try aggressively squeezing the
viewport up to see what happens. (I don't expect to solve all the problems.
I'm normally happy enough if my pages work well with a viewport of at least
600 pixels. And 700 for photograph pages).

--
Barry Pearson
http://www.Barry.Pearson.name/photography/
http://www.BirdsAndAnimals.info/
http://www.ChildSupportAnalysis.co.uk/

Jul 20 '05 #85

JRS: In article <8K********************@twister.nyroc.rr.com>, seen in
news:comp.infosystems.www.authoring.stylesheets, Beauregard T. Shagnasty
<a.*********@example.invalid> posted at Wed, 16 Jun 2004 17:48:52 :

I use this JavaScript in my local index page, rather than the <body> code:

<a href="javascript:resizeTo(240,400)">240x320</a> /
<a href="javascript:resizeTo(640,480)">640x480</a> /
<a href="javascript:resizeTo(800,600)">800x600</a> /
<a href="javascript:resizeTo(1004,728)">1024x768</a>

So do I, now, with adjusted numbers - thanks.

--
© John Stockton, Surrey, UK. ?@merlyn.demon.co.uk Turnpike v4.00 MIME. ©
Web <URL:http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
Proper <= 4-line sig. separator as above, a line exactly "-- " (SonOfRFC1036)
Do not Mail News to me. Before a reply, quote with ">" or "> " (SonOfRFC1036)

Jul 20 '05 #86