<small> tags in html5

[Tim W]

<small> used to be a sort of typographical tag which was used to make
text small. It was always (I think) used inside a para or a heading or
some other element.

Now in html5, despite what w3schools says (Differences Between HTML 4.01
and HTML5, NONE.) it has a semantic meaning of something like 'the small
print'.

Does this mean I can use it on its own like:

<p>Some statements about something or other</p>
<small>disclaimers re the above</small>

without wrapping it in a paragraph?

Tim W

 

[Jukka K. Korpela]

<small> used to be a sort of typographical tag which was used to make
text small. It was always (I think) used inside a para or a heading or
some other element.

Yep. Note, however, that there was no definition for how small the text
would be. HTML5 is nominally more explicit: it says that the "suggested
rendering" corresponds to

small { font-size: smaller; }

which really leaves it to browsers decide, whereas CSS 2.1 says, in a
"sample" style sheet for HTML 4:

small { font-size: 0.83em; }

The morale is: set the font size, so that you will not have
*unnecessary* variation across browsers. Usually 0.83em, or maybe a
little larger, is OK, but this somewhat depends on the font you intend
to use.

Now in html5, despite what w3schools says (Differences Between HTML 4.01
and HTML5, NONE.)

Ignore w3schools here, and otherwise; see http://w3fools.com

it has a semantic meaning of something like 'the small
print'.

Which isn't really semantic (i.e., relating to meaning) at all. It says
nothing about the meaning of the content. It just assigns a vague phrase
to the element. The explanation is more obscure than the thing it is
supposed to explain.

The sensible rule is simple: use <small> if you want something to appear
in smaller font than the surrounding text even when CSS is disabled, and
use CSS to set the specific font size proportion. It's really your
business *why* you want reduced font size, and the user won't see the
reason anyway, no matter what markup you use.

You can safely ignore "semantic" babble in this issue, as well as with
<i>, <b>, <u>, and some friends. Unless, of course, your pointy-haired
boss or your rich customer tells you to write "HTML5 conformant" pages.
(And even then, it is a matter of moral and honesty more than anything
else: how much can you lie and deceit and mislead when it really does
not hurt anyone and when trying to be honest at any cost could cost you
your mental health?)

For example, I use

<h1>The pragmatic guide to HTML: Principles<br>
<small>or<br>
The HTML Anarchist’s leaflet</small></h1>

at
http://www.cs.tut.fi/~jkorpela/pragmatic-html.html8
even though such usage might not match HTML5 "semantics" as of now (but
might do so tomorrow, or at some later point).

Does this mean I can use it on its own like:

<p>Some statements about something or other</p>
<small>disclaimers re the above</small>
Yes.

without wrapping it in a paragraph?

Yes. But you might wish to wrap all paragraphs (small print or not) in
<p> elements in order to be able to style them in a uniform manner.

And if the pointy-haired boss or rich customer forces you to be an HTML5
conformist, you can use

<span class=small>disclaimers re the above</span>

with

..small { font-size: 85% }

What the users will then lose is just that if CSS is turned off or
somehow filtered out (e.g., someone is viewing a cached copy in an
archive that has not stored the CSS files), then the disclaimers appear
in normal font size.

 

[Jukka K. Korpela]

If the HTML5 suggestion for [<small>] was that it be used to make side
comments or "small print" then, no matter what we are to think about
the usefulness of semantic markup, it does say something about the
content.

Like what? It's something that someone wants to call "small print".

The expression "small print" is not tautologous here; it is a
meaningful expression to indicate legal text, copyright spiels,
regardless of the size of the text.

Is it? Do you call laws "small print"? If promulgated laws are not legal
text, what is? Is a copyright announcement "small print" if presented in
very large font size to convey an essential message?

How do you *define* "small print" to a person who does not know your
meaning for it (which is, vague as it is, admittedly common in
Anglo-Saxon culture)? That is, what are necessary criteria for
distinguishing small print from other texts?

The real meaning of "small print" is "text presented in small font". The
only thing that "small print" adds here is an unpleasant connotation.

The expression gets its name from
the fact that it is customary to put such content in harder to read
small text.

The connotation is that "small print" is text made intentionally harder
to read, so that nobody is really expected to read it, but people can
still be sued and condemned on the basis that they should have read it.

 

[Tim Streater]

If the HTML5 suggestion for [<small>] was that it be used to make side
comments or "small print" then, no matter what we are to think about
the usefulness of semantic markup, it does say something about the
content.

Like what? It's something that someone wants to call "small print".

The expression "small print" is not tautologous here; it is a
meaningful expression to indicate legal text, copyright spiels,
regardless of the size of the text.

Is it? Do you call laws "small print"?

In English we do, yes. The expression "Did you read the small print?"
means, did you read any warnings, legal notices, terms and conditions,
etc etc, that might be relevant to the point being discussed, whatever
that might be.

 

[Jukka K. Korpela]

In English we do, yes. The expression "Did you read the small print?"
means, did you read any warnings, legal notices, terms and conditions,
etc etc, that might be relevant to the point being discussed, whatever
that might be.

The question was whether you call laws "small print". Laws are documents
issued by authorities, such as parliaments and presidents, expressing
their ruling power to dictate what shall be done or shall not be done.
Such as the Constitution of the United States, or a Copyright Act, or a
Penalty Code.

Now, if "small print" means "legal text" among other things, then surely
such laws are "small print". Or? If a book is a compilation of laws, is
it entirely "small print"? If it additionally contains annotations to
laws, printed in small font size, would you still say that the laws are
"small print" and the annotations are something else?

What you seem to be saying is that some texts that *cite* laws may be
"small print".

But again, what would then be a *definition* of "small print"? Would
*all* warnings be "small print"? All terms and conditions? Or are you
actually referring to a presentation style where some bulk of text is
printed in small font size for some reason?

 

[Ben C]

Which isn't really semantic (i.e., relating to meaning) at all. It says
nothing about the meaning of the content.

If the HTML5 suggestion for it was that it be used to make side
comments or "small print" then, no matter what we are to think about
the usefulness of semantic markup, it does say something about the
content.

The expression "small print" is not tautologous here; it is a
meaningful expression to indicate legal text, copyright spiels,
regardless of the size of the text. The expression gets its name from
the fact that it is customary to put such content in harder to read
small text. A lot of money has been made by a lot of people from this
practice. <g>[/QUOTE]

So probably browsers should display the contents of <small> elements in
a nice clear easy-to-read font that's a bit larger than usual.

 

[dorayme]

If the HTML5 suggestion for [<small>] was that it be used to make side
comments or "small print" then, no matter what we are to think about
the usefulness of semantic markup, it does say something about the
content.

Like what? It's something that someone wants to call "small print".

No, it's not just someone calling it that.

Is it? Do you call laws "small print"?
No.

Is a copyright announcement "small print" if presented in
very large font size to convey an essential message?

No.

How do you *define* "small print" to a person who does not know your
meaning for it (which is, vague as it is, admittedly common in
Anglo-Saxon culture)? That is, what are necessary criteria for
distinguishing small print from other texts?

It's a meaning that must be understood from a family of contexts. You
cannot always define an expression so that people unfamiliar with it
will thereby be conversant with it. Dictionaries can only take you so
far.

As for criteria, one thing Wittgenstein did get right, was that many
concepts are family terms, not able to be cashed in terms of necessary
and sufficient conditions.

The real meaning of "small print" is "text presented in small font". The
only thing that "small print" adds here is an unpleasant connotation.

No, again. Small print is not just small text.

The connotation is that "small print" is text made intentionally harder
to read, so that nobody is really expected to read it, but people can
still be sued and condemned on the basis that they should have read it.

The connotation is not that quite. It may very well be intended, and
as I said, people have made a lot of money from such such an unsavoury
aspect. An honest author, advertiser, publisher might well draw
attention to the fine print, small print, in large print. So it is not
a direct connotation. It may just be a convenience to everyone
considering a lot of legal blah is very similar to have it not too
prominent.

 

[dorayme]

So probably browsers should display the contents of <small> elements in
a nice clear easy-to-read font that's a bit larger than usual.

Perhaps in a juster world! <g>

But no, when I said "regardless of the size of the text", I was
meaning that the absolute size of the text was not relevant. It
sometimes if not often denotes material the authors consider to be
standard between it and its competitors, that it risks the death from
boredom of potential customers. The urge for small or fine print
sections is sometimes honourable, to void confusion.

Sydney has just introduced the Opal card, you tap it on a machine at
the beginning and end of ferry, bus and train journeys. There are
constant reminders to do this as it is so new. On the ferries (my main
experience since I live on the harbour) there are loud announcements
at every stop! When everyone gets used to this card, I am hoping the
operational advice will get said more quickly in a quieter voice or
abandoned altogether. In other words, I want it in the fine print of
life.

When people want to read a timetable, it is not their primary interest
to be told about things other than times, so... other interesting but
not relevant to the main purpose things can be left to the fine or
small print...

 

[Ben Bacarisse]

The connotation is that "small print" is text made intentionally
harder to read, so that nobody is really expected to read it, but
people can still be sued and condemned on the basis that they should
have read it.

I don't think it need always be so nefarious. The key feature, I feel,
is that it's extra detail -- more detail than more people will want on
the first reading.

For example, on a food product page, I might put the ingredients in
<small>...</small>, and I might use it for those attribution/credit
blocks that so many site designs seem to need at the bottom. Neither is
being made hard to read for any underhand reason, but rather it's being
downplayed as being of secondary interest. Small print may be vital,
but it's considered to be detail that's less interesting than the rest
of the text.

 

[Jukka K. Korpela]

It's a meaning that must be understood from a family of contexts.

So you can't give a definition, can you?

HTML specifications have some tradition of using vague and idiomatic
expressions that are understandable only to the small fraction of
world's population who speak some forms of upper-class English and even
to them in conflicting ways. Don't make me started on <cite> and
<acronym>, for example.

But <small> used to be a nice little tag before some people made two
wrong decisions: to expel all "presentational markup" from HTML As We
Accept It, and to save <small>, due to its widespread use, my
transmogrifying into something nominally "semantic". They are doing the

You
cannot always define an expression so that people unfamiliar with it
will thereby be conversant with it. Dictionaries can only take you so
far.

In specifications, such expressions are to be avoided. It is now
impossible to say whether a page using <small> is conforming or not.
Everyone and his brother has a different idea of the "meaning" of

As for criteria, one thing Wittgenstein did get right, was that many
concepts are family terms, not able to be cashed in terms of necessary
and sufficient conditions.

Anyone tempted to be Wittgensteinian that way should refrain from trying
to write, edit, or read specifications in that mood. Just as you should
not give public speech when in Solipsistic mood, or perform surgery when
in Nihilistic mood.

No, again. Small print is not just small text.

So what *is* it?

Considering a hypothetical web author living outside your linguistic
community, how do you expect him to use a markup element properly if it
is defined as denoting "small text" and dictionaries don't tell what it
means? Should he call you, or Hixie, or the W3C director whenever he is
considering the use of <small>?

 

[Tim Streater]

The question was whether you call laws "small print". Laws are documents
issued by authorities, such as parliaments and presidents, expressing
their ruling power to dictate what shall be done or shall not be done.
Such as the Constitution of the United States, or a Copyright Act, or a
Penalty Code.

Now, if "small print" means "legal text" among other things, then surely
such laws are "small print". Or? If a book is a compilation of laws, is
it entirely "small print"? If it additionally contains annotations to
laws, printed in small font size, would you still say that the laws are
"small print" and the annotations are something else?

What you seem to be saying is that some texts that *cite* laws may be
"small print".

But again, what would then be a *definition* of "small print"? Would
*all* warnings be "small print"? All terms and conditions? Or are you
actually referring to a presentation style where some bulk of text is
printed in small font size for some reason?

The phrase "small print", is a colloquial one, and renders expressions
such as "He didn't read the small print" colloquial, the meaning being
that there was legal [1] text that this person didn't read, but should
have done for his own protection. Whether the text in question is
actually presented in small font size does not matter, it would still
be colloquially referred to as "small print".

[1] Using the word "legal" here in a general sense to include warnings,
Ts & Cs, contract text, and the like.

 

[Tim Streater]

2013-11-05 22:39, dorayme wrote:

If the HTML5 suggestion for [<small>] was that it be used to make side
comments or "small print" then, no matter what we are to think about
the usefulness of semantic markup, it does say something about the
content.

Like what? It's something that someone wants to call "small print".

The expression "small print" is not tautologous here; it is a
meaningful expression to indicate legal text, copyright spiels,
regardless of the size of the text.

Is it? Do you call laws "small print"?

In English we do, yes. The expression "Did you read the small print?"
means, did you read any warnings, legal notices, terms and conditions,
etc etc, that might be relevant to the point being discussed, whatever
that might be.

No, we don't use "small print" to refer to laws. "Small print" is a
colloquiallism referring to legalese in a document, NOT a law.

A law is a statute enacted by a state or the federal congress. It is
NOT fine print.[/QUOTE]

Stop behaving like PointyHead. And our laws are enacted by Parliament.

 

[Jukka K. Korpela]

Sorry, not getting what the issue is.

Suppose that your boss, your client, or your professor tells you to use
HTML that conforms to HTML5. Never mind the absurdity of conforming to
mutable drafts that have not been really identified (HTML5 CR? HTML 5.1
WD; WHATWG Living HTML? as of now, or as of yesterday?). Just consider
what HTML5 (currently all those mutable documents) says about <small>.
How dare you use it for *anything*? (Well, anything but legalese babble
that you were forced, e.g. using a gun, to include into your page,
despite having no useful purpose there.)

 

[dorayme]

So you can't give a definition, can you?

So what? There are many terms that are understood, in all languages,
that have no definition and certainly no particular explanations that
are immune to misunderstanding.

....

In specifications, such expressions are to be avoided.

Then you better object to all the colour words because they too are
largely undefinable except ostensibly.

It is now
impossible to say whether a page using <small> is conforming or not.

Impossible is too strong. The idea of side comments, things that would
distract from the central message being conveyed, common legal
disclaimers and such are often easy to understand as conforming. Not
that I particularly want to defend the idea. It is not an impossible
idea, it may be not a very useful one, there may be better ideas for
this element, but it is not impossible.

Anyone tempted to be Wittgensteinian that way should refrain from trying
to write, edit, or read specifications in that mood. Just as you should
not give public speech when in Solipsistic mood, or perform surgery when
in Nihilistic mood.

It is not any kind of mood to be aware that some concepts do not have
neat definitions or necessary and sufficient conditions attached to
them. It is simply being realistic.

So what *is* it?

In the thread I have made some gestures towards explaining it, Tim
Streater has too and also Ben Bacarisse. I had no difficulty
understanding their remarks.

Considering a hypothetical web author living outside your linguistic
community, how do you expect him to use a markup element properly if it
is defined as denoting "small text" and dictionaries don't tell what it
means? Should he call you, or Hixie, or the W3C director whenever he is
considering the use of <small>?

He should call anyone who knows the expression "the fine print", "the
small text" in its uses in ordinary conversation - and if he still
finds a use for semantic markup, I am sure he will find it easy enough
to judge for himself.

I have no axes to grind on this one. If it is simpler to think the
element is for whenever you want smaller text than your main body
text, that is fine by me. I quite like the more complicated one
because it is not just about style, it is about a purpose which many
authors might well find useful.

 

[Jukka K. Korpela]

Then you better object to all the colour words because they too are
largely undefinable except ostensibly.

In HTML and CSS specifications, all color words are defined with
specific definitions in terms of RGB color values. ”Purple” means
exactly #800080.

Impossible is too strong.

Then tell us how to decide whether <h1>Foolish ideas <br>
<small>are often called semantics</small></h1> is conforming or not. If
this is not objectively decidable, conformance concept is useless -
there is no point in requiring conformance (in instructions, contracts,
teaching) if there is no way to make two rational, educated people agree
on whether something if conforming or not.

The idea of side comments, things that would
distract from the central message being conveyed, common legal
disclaimers and such are often easy to understand as conforming.

It is not any kind of mood to be aware that some concepts do not have
neat definitions or necessary and sufficient conditions attached to
them. It is simply being realistic.

Indeed so. And the rational conclusion would be that specifications must
not use such concepts, at least not in normative parts.

In the thread I have made some gestures towards explaining it, Tim
Streater has too and also Ben Bacarisse. I had no difficulty
understanding their remarks.

I have, but more importantly, the statements have been rather different.

If you cannot explain what a phrase-level element means, in one
sentence, or at least while standing on one foot, then it is guaranteed
that the element will be used in so many incompatible ways that it loses
all meaning in practice. You cannot realistically draw any conclusions
from it. It’s useless markup, except that it triggers font size
reduction by default.

He should call anyone who knows the expression "the fine print", "the
small text" in its uses in ordinary conversation - and if he still
finds a use for semantic markup, I am sure he will find it easy enough
to judge for himself.

If it is simpler to think the
element is for whenever you want smaller text than your main body
text, that is fine by me. I quite like the more complicated one
because it is not just about style, it is about a purpose which many
authors might well find useful.

No two persons will really agree on what the purpose *is*, except by
accident. And no software will pay the least attention to the "semantic"
definitions. Users will not see the markup - they will not see whether
<small> or just font-size was used. It would have been better to
introduce a new element, like <fineprint>, which would have no default
impact on rendering, or anything. You could define it as you like, and
use it as you like - pure write-only markup.

In conclusion, all this "semantics" is really "semantics" in the worst
sense of the word. What I have been trying to do is to save HTML authors
from it - in vain, it seems. (I'm not referring to this thread; rather,
discussions in mailing lists about HTML5.)

 

[Tim Streater]

2013-11-06 12:48, dorayme wrote:

Really? Side comments were something that <aside> was for, when I last
checked (that is, yesterday).

An aside is something else again - nothing to do with small print.

Asides are what you get when for instance a committee chairman doesn't
do a good job of keeping people focussed on the matter at hand. In the
theatre, an actor saying something directly to the audience would also
be an aside (the convention there being that the other actors do not
hear what this actor is saying. The actor might whisper loudly to the
audience to emphasise this pretence).

You might get asides in stories or web pages, where it might be
annoying for the reader because the asides distract from the main
thread.

 

[dorayme]

In HTML and CSS specifications, all color words are defined with
specific definitions in terms of RGB color values. ²Purple² means
exactly #800080.

OK, but the point is how is an author to know when to use #80080? When
he wants purple. But he needs to understand what purple is. Surely? In
the case of "fine print" or "small print", it means exactly (according
to one interpretation of the HTML5 recommendation) <small></small> for
authoring purposes. I am simply playing devil's advocate for this
notion, not saying it has to be this way. Just as different people
might have different ideas as to what purple is, so too there may be
some vagaries about fine print (as explained by various folk in the
thread and as widely understood). The lack of sharp boundaries to a
concept does not seem to me to make it useless for authoring

Then tell us how to decide whether <h1>Foolish ideas <br>
<small>are often called semantics</small></h1> is conforming or not.

I would say clearly not from your example.

If
this is not objectively decidable, conformance concept is useless -
there is no point in requiring conformance (in instructions, contracts,
teaching) if there is no way to make two rational, educated people agree
on whether something if conforming or not.

All a browser need do is make the print smaller (they are not
intelligent language users). The human author can use it as his
personal tactic to put in legalese, copyright notices etc. And never
use it *just* to make text smaller, using CSS instead.
....

Indeed so. And the rational conclusion would be that specifications must
not use such concepts, at least not in normative parts.

I don't think the idea is so unclear as to be useless in such specs
just because it cannot be dictionary defined.

I have, but more importantly, the statements have been rather different.

If you cannot explain what a phrase-level element means, in one
sentence, or at least while standing on one foot, then it is guaranteed
that the element will be used in so many incompatible ways that it loses
all meaning in practice. You cannot realistically draw any conclusions
from it. It¹s useless markup, except that it triggers font size
reduction by default.

If all of those who understand the phrase from being natives of the
language or are as conversant as the natives, use it just for
legalese, some copyright notices, side remarks, it could be a handy
common understanding between authors. For merely making print small,
just use CSS.

That did not happen with <cite>, <acronym>, and <abbr>. People won¹t
call anyone. If they did, they did get different, incompatible answers.

Oh well, if it causes confusion to too many people, then the idea will
wither away and die. Let me try it out for a year or two, as a
personal practice. I'll let you know how it feels. But I get the
feeling you are trying to strangle the poor thing before it can get

No two persons will really agree on what the purpose *is*, except by
accident. And no software will pay the least attention to the "semantic"
definitions.

Yes, OK, but that is really a different question, a broader one. I was
assuming there is a reasonable point to semantic markup. If it is
really true (and you have argued strongly for this in recent times)
that the promise of semantic markup having a proper resonance in
browsers is now doomed, then I invite you all to think of it as I do:
to go on marking up as semantically as possible to inculcate a
repeatable orderly habit in oneself, to use CSS for styling as much as
possible, to not start every project with "whatever is practical,
whatever looks ok in browsers", this latter being too undirected.

 

[Ben Bacarisse]

But is that not also true for, say, <em> and <strong>? Not all of the
mark-up follows inevitably from the content. Sometimes the markup is
there to impart a meaning what is otherwise not there. Given:

<em>Not all the contributing authors share this view</em>

is very different to

<small>Not all the contributing authors share this view</small>

Neither is right nor wrong -- they are intended to say different things.

If you cannot explain what a phrase-level element means, in one
sentence,

The OED uses only one:

"The detailed information or conditions qualifying the principal text
of a document, typically printed in a smaller type."

or at least while standing on one foot, then it is
guaranteed that the element will be used in so many incompatible ways
that it loses all meaning in practice. You cannot realistically draw
any conclusions from it. It’s useless markup, except that it triggers
font size reduction by default.

That may turn out to be true in this case. This very thread suggests
that the meaning is not universal whereas, previously, I thought it
was. Too much ambiguity will eventually render the element meaningless.

I don't think the OED does a bad job, but it's quite possible that other
dictionaries disagree in some significant way. The OED also lists a
literal meaning ("printed in small type") but I think the HTML5 document
says enough to make it clear that this is not was is intended.

<snip>

 

[Osmo Saarikumpu]

5.11.2013 14:19, Tim W kirjoitti:

... what w3schools says (Differences Between HTML 4.01
and HTML5, NONE.)...

Wow, they got that one right

 

[Tim W]

I don't think it need always be so nefarious. The key feature, I feel,
is that it's extra detail -- more detail than more people will want on
the first reading.

For example, on a food product page, I might put the ingredients in
<small>...</small>, and I might use it for those attribution/credit
blocks that so many site designs seem to need at the bottom. Neither is
being made hard to read for any underhand reason, but rather it's being
downplayed as being of secondary interest. Small print may be vital,
but it's considered to be detail that's less interesting than the rest
of the text.

I am thinking that I might be careful not to put anything in <small>
that I want google to pick up and give weight to. The implication being
that <small> is detail you may not want to even read.

Tim w