Standards for presentation of numerical data for usability..

Scripsit D.M. Procida:

> I'm trying to find a guide for the formatting and presentation of
> numbers in web pages, which will answer questions along the lines of
> what if anything to use as separators in large figures, how to use
> abbreviations for things like millions of dollars, and so on.

I don't see why the guidelines would be any different from the guidelines
for printed matter, handwritten material, or office documents. The
guidelines vary by language and authority. Since you crossposted to an "uk."
group, I guess a British style for English, like the Oxford Manual of Style,
might be suitable for you.

There is rather little HTML related here, mostly the issue of a no-break
space. If the applicable style rules use a space as a thousands separator,
or between a number and a currency denotation, then you should normally use
the no-break space (character U+00A0, written as such or using the entity
  if desired), e.g. (assuming Finnish standard notation for monetary
quantities)
1 000 000 €
As the example shows, you might use an entity for some currency symbols,
too, if you don't know how to write them otherwise; you would use € for
the euro sign and £ for the pound sign.

Typographic standards may require or suggest that the space used as a
thousands separator be narrower than a normal space, typically the so-called
thin space. This opens a can of worms (the Unicode character THIN SPACE
might have the appropriate width, but many fonts lack it and it's, oddly
enough, defined as breaking instead of nonbreaking), but it's perhaps not
useful to discuss this if it's not relevant to your problem.

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

Jukka K. Korpela <jkorpela.RemoveThis@cs.tut.fi> wrote:

> > I'm trying to find a guide for the formatting and presentation of
> > numbers in web pages, which will answer questions along the lines of
> > what if anything to use as separators in large figures, how to use
> > abbreviations for things like millions of dollars, and so on.
>
> I don't see why the guidelines would be any different from the guidelines
> for printed matter, handwritten material, or office documents. The
> guidelines vary by language and authority. Since you crossposted to an "uk."
> group, I guess a British style for English, like the Oxford Manual of Style,
> might be suitable for you.

I'm wondering how all this will work for users of screen readers and so
on. For example, if using a comma separator in numbers means that
"23,089,241" will be read out as a list of three figures, then that
won't be very good.

Another example, this time for numbers in tables: I might want to make a
table of figures easier to parse by rounding the numbers and using
thousands or millions rather than units. So instead of "$13286723", I
could use "$13.3m", or "13.3m" or "$13.3" or even just "13.3".

But then the currency and/or the multiplier would need to be flagged
somewhere, and I don't know how users of screen readers would be best
served by the placement of that information (in the appropriate <th>?
the summary? the caption? some combination of those?).

And what about the use of conventions like "Annual sales (in $000s)" or
"Annual sales (in $1000s)" or "Annual sales (x $1000)" as heading
information for numerical data? All of these and more regularly appear
in print and are readily recognised, but I don't know how well if at all
these work for users for whom at-a-glance assimilation of such
information is not possible.

Daniele

On Fri, 10 Aug 2007, D.M. Procida wrote:

> Another example, this time for numbers in tables: I might want to make a
> table of figures easier to parse by rounding the numbers and using
> thousands or millions rather than units. So instead of "$13286723", I
> could use "$13.3m", or "13.3m" or "$13.3" or even just "13.3".

Most people would take "13.3m" or "13.3 m" as "13.3 metres".
Further, the symbol for "mega" is "M", meaning one million -
whereas the symbol for "milli" is "m", meaning one thousandth.

(What has this to do with authoring.html?)

Andreas Prilop <Prilop2007.RemoveThis@trashmail.net> wrote:

> > Another example, this time for numbers in tables: I might want to make a
> > table of figures easier to parse by rounding the numbers and using
> > thousands or millions rather than units. So instead of "$13286723", I
> > could use "$13.3m", or "13.3m" or "$13.3" or even just "13.3".
>
> Most people would take "13.3m" or "13.3 m" as "13.3 metres".
> Further, the symbol for "mega" is "M", meaning one million -
> whereas the symbol for "milli" is "m", meaning one thousandth.

Well, exactly. There are a number of conventions surrounding how to
present such figures, and while this generally isn't a problem for users
of printed material, or users who can easily grasp the context and
convention from surrounding material, it still presents accessibility
and usability issues for others in the context of HTML.

> (What has this to do with authoring.html?)

It's a question about HTML authoring.

Daniele

On Fri, 10 Aug 2007, David E. Ross wrote:

> Be sure your HTML indicates language, at the very beginning in your
> !DOCTYPE.

To be specific, this means <html lang="en"> .
http://www.w3.org/TR/html4/struct/dirlang.html#h-8.1

Scripsit D.M. Procida:

> I'm wondering how all this will work for users of screen readers and
> so on.

Yes, I now realize that point, but it too is a general question, not
specific to HTML. It also arises e.g. when plain text is fed into a speech
synthesizer.

> For example, if using a comma separator in numbers means that
> "23,089,241" will be read out as a list of three figures, then that
> won't be very good.

It depends on the screen reader and possibly on its settings. I would use
whatever is common in the language I'm using, since that's the best shot.
After all, the software needs to know the language in some sense in order to
read it properly, and it is rational to expect it to be able to deal with
that language's notations.

It happens that a program reads a sequence of numbers in a wrong way, e.g. a
telephone number as a number and not by digits. But their users are used to
that, and they may have tools (program options) for changing the behavior if
needed.

In a more ideal world, we could use markup that indicates whether a sequence
of digits (possibly with interspersed punctuation) is a mathematical number
or a code-like notation (normally read digit by digit). But HTML has nothing
like that.

> Another example, this time for numbers in tables: I might want to
> make a table of figures easier to parse by rounding the numbers and
> using thousands or millions rather than units.

That might be a very good idea for accessibility reasons. E.g., cognitively
challenged people usually have difficulties with long numbers and can handle
rounded numbers better. Guidelines for "Easy Language" explicitly discuss
such issues. But this depends on the nature of the data and the purposes of
presenting it. Generally, figures should be as accurate as needed, but not
more accurate. In hypertext, you can always have a basic page with round,
easy-to-grasp numbers and with a link to a version with more exact data.

> So instead of
> "$13286723", I could use "$13.3m", or "13.3m" or "$13.3" or even just
> "13.3".

(A notation like "$13286723" is certainly wrong, in anything meant to be
generally readable by human beings. Some grouping and some group separators
are needed.)

> But then the currency and/or the multiplier would need to be flagged
> somewhere, and I don't know how users of screen readers would be best
> served by the placement of that information (in the appropriate <th>?
> the summary? the caption? some combination of those?).

Well, here we an HTML issue, too. Basically, you should explain the
notations used in the table _before_ the table (so that they will be read
before the table data). The <caption> element is for a heading-like text,
but it may contain information like...

> And what about the use of conventions like "Annual sales (in $000s)"
> or "Annual sales (in $1000s)" or "Annual sales (x $1000)" as heading
> information for numerical data?

Here, too, the crucial thing is the conventions of the natural language
used. Use what's best and easiest to understand to users.

The summary="..." attribute is mostly useless, but in special cases, you
_might_ use it for extra explanations that need not be visible but should be
available to people using speech-based browsing. It's debatable, though, and
there is no _guarantee_ that the information in it will be available to and
noticed by people who need it.

The <th> elements should contain understandable column (or row) headers, and
writing them is often challenging. Normally they need to be short, at least
for numeric columns, and then they might not be very informative. In HTML,
you can use the abbr="..." attribute to give a short identification to be
used by speech browsers (when they read data cells, they might say the row
and column headers before the data, so that the user knows where he is in
the table) and the title="..." attribute to specify a longer explanation
that _might_ be presented to the user when reading the column headers
themselves.

So you might have e.g.

<caption>Annual sales (in $1000)</caption>

_and_

<th scope="col" abbr="domestic" title=
"Domestic sales in thousands of dollars">Domestic sales</th>

but such markup is tedious to write (and not always easy to write
_correctly) and of limited usefulness.

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

David E. Ross wrote:

> On 8/10/2007 4:49 AM, D.M. Procida wrote:
>> I'm trying to find a guide for the formatting and presentation of
>> numbers in web pages, which will answer questions along the lines of
>> what if anything to use as separators in large figures, how to use
>> abbreviations for things like millions of dollars, and so on.
>>
>> Thanks,
>>
>> Daniele
>>
>> [Followups set, please excuse the earlier message
>> alt.html.web-accessibility which didn't get cross-posted]
>
> Be sure your HTML indicates language, at the very beginning in your
> !DOCTYPE. This is important because some languages use commas as
> separators and periods as decimal points (e.g., English in the US and
> UK) and some languages reverse this, using periods as separators and
> commas as decimal points.
>
> I would write out the numbers using numerals instead of words. This
> serves two purposes.
>
> First of all, in some countries (e.g., the US) "billion" means a
> thousand million; in others, it means a million million. I think (not
> sure) this UK is in the latter group.

In the UK a million used to be 1,000,000,000,000 but it is now almost
universally accepted to be 1,000,000,000

On 2007-08-10, PDannyD <pdannyd.DeleteThis@yahoo.co.uk> wrote:
> David E. Ross wrote:
[...]
>> First of all, in some countries (e.g., the US) "billion" means a
>> thousand million; in others, it means a million million. I think (not
>> sure) this UK is in the latter group.
>
> In the UK a million used to be 1,000,000,000,000

That was a billion. A million has always been 1e6. David E. Ross was
right.

> but it is now almost universally accepted to be 1,000,000,000

Yes, even in the UK. An old-style billion is now often called a
"trillion".

Scripsit David E. Ross:

> Be sure your HTML indicates language, at the very beginning in your
> !DOCTYPE.

You mean <html lang="...">, as Andreas remarked. Anyway, indicating language
is of little practical value. It's good practice, but it doesn't do
miracles, or much else.

> This is important because some languages use commas as
> separators and periods as decimal points (e.g., English in the US and
> UK) and some languages reverse this, using periods as separators and
> commas as decimal points.

And some languages have varying practices. But does this matter? Almost all
browsers (and search engines) ignore language markup.

> Yes, I do think this was an appropriate question for this newsgroup.

Was there something in the question that was specific to HTML authoring, as
opposite to writing, say, a document in a word processor? I don't think so.
There are some finer points specific to HTML, but most people haven't even
got the big picture: they write numbers and quantities sloppily, no matter
whether it's HTML, e-mail, or something else.

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

In comp.infosystems.www.authoring.html message
<1i2mrhr.17bkhoe1ntjtsN%real-not-anti-spam-address@apple-juice.co.uk>,
Fri, 10 Aug 2007 12:49:25, D.M. Procida <real-not-anti-spam-
address RemoveThis @apple-juice.co.uk> posted:
>I'm trying to find a guide for the formatting and presentation of
>numbers in web pages, which will answer questions along the lines of
>what if anything to use as separators in large figures, how to use
>abbreviations for things like millions of dollars, and so on.

A decimal point should always have a digit on each side, except maybe in
tables where the format makes that unnecessary and impractical. It's
better not to have a number (in digits) at the beginning or end of a
sentence.

There is an international standard for representing currencies by
letters, probably ISO 4217 or successor.

--
(c) John Stockton, Surrey, UK. *@merlyn.demon.co.uk / ??.Stockton@physics.org
Web <URL:http://www.merlyn.demon.co.uk/> - FAQish topics, acronyms, & links.
Correct <= 4-line sig. separator as above, a line precisely "-- " (SoRFC1036)
Do not Mail News to me. Before a reply, quote with ">" or "> " (SoRFC1036)

Scripsit Dr J R Stockton:

> A decimal point should always have a digit on each side, except maybe
> in tables where the format makes that unnecessary and impractical.
> It's better not to have a number (in digits) at the beginning or end
> of a sentence.

There are many opinions, practices, and recommendations on such matters,
depending on language, authority, context, etc. And there is nothing
particularly HTML-related here.

> There is an international standard for representing currencies by
> letters, probably ISO 4217 or successor.

Or by numbers. Such codes have been defined for the purposes of
international banking business, internal data base representation, etc., not
for normal text. And there is nothing particularly HTML-related here.

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

In comp.infosystems.www.authoring.html message <Xepvi.205746$396.144606@
reader1.news.saunalahti.fi>, Sat, 11 Aug 2007 23:46:47, Jukka K. Korpela
<jkorpela RemoveThis @cs.tut.fi> posted:
>Scripsit Dr J R Stockton:
>
>> A decimal point should always have a digit on each side, except maybe
>> in tables where the format makes that unnecessary and impractical.
>> It's better not to have a number (in digits) at the beginning or end
>> of a sentence.
>
>There are many opinions, practices, and recommendations on such
>matters, depending on language, authority, context, etc. And there is
>nothing particularly HTML-related here.

No; my intent was to be helpful to the OP, rather than to be
unnecessarily offensive. My first sentence there re-expresses something
which may, IIRC, be found in IUPAP/SUNAMCO, which is sufficiently
authoritative where it formally applies and an excellent guide wherever
the same reasoning applies. My second has no particular authority other
than my own; but IMHO will be preferred by any well-educated person.

>> There is an international standard for representing currencies by
>> letters, probably ISO 4217 or successor.
>
>Or by numbers. Such codes have been defined for the purposes of
>international banking business, internal data base representation,
>etc., not for normal text.

The alphabetical code, while inappropriate for text intended to have
great literary appeal, can reasonably be used in paragraphs of a more
technical nature, and is certainly appropriate in relevant HTML tables.

--
(c) John Stockton, Surrey, UK. ?@merlyn.demon.co.uk Turnpike v6.05 IE 6
news:comp.lang.javascript FAQ <URL:http://www.jibbering.com/faq/index.html>.
<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.

PDannyD wrote:

> David E. Ross wrote:
>
>> On 8/10/2007 4:49 AM, D.M. Procida wrote:
>>> I'm trying to find a guide for the formatting and presentation of
>>> numbers in web pages, which will answer questions along the lines of
>>> what if anything to use as separators in large figures, how to use
>>> abbreviations for things like millions of dollars, and so on.
>>>
>>> Thanks,
>>>
>>> Daniele
>>>
>>> [Followups set, please excuse the earlier message
>>> alt.html.web-accessibility which didn't get cross-posted]
>>
>> Be sure your HTML indicates language, at the very beginning in your
>> !DOCTYPE. This is important because some languages use commas as
>> separators and periods as decimal points (e.g., English in the US and
>> UK) and some languages reverse this, using periods as separators and
>> commas as decimal points.
>>
>> I would write out the numbers using numerals instead of words. This
>> serves two purposes.
>>
>> First of all, in some countries (e.g., the US) "billion" means a
>> thousand million; in others, it means a million million. I think (not
>> sure) this UK is in the latter group.
>
> In the UK a million used to be 1,000,000,000,000 but it is now almost
> universally accepted to be 1,000,000,000

Doh! I meant a billion, not a million.
Well spotted Ben C.