When did IE stop recognizing entity references without ";"?

[Jukka K. Korpela]

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.

I'm pretty sure that previous versions of IE rendered them by the
specifications. I first thought this has something to do with XML (i.e.
maybe IE pretends to play a little bit of XML game when processing HTML,
even when I serve it as text/html and use an HTML 4.01 DOCTYPE). But on
closer look, I notice that
&auml &eacute
work correctly.

It seems that ISO Latin 1 entities work without a semicolon, newer
entities don't. Am I mistaken, i.e. was this always the case on IE?

(Of course, it has always been recommendable to use the terminating
semicolon in entity references, but it is not required by SGML, hence by
classic HTML rules, when the reference is followed by a character that
is not a name character, e.g. by a space. There must be _lots_ of legacy
documents that play fast with entity references without semicolons.)

[Leif K-Brooks]

Jukka K. Korpela wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.

Internet Explorer 5.5 seems to behave the same way.

[Els]

Leif K-Brooks wrote:

Jukka K. Korpela wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.

Internet Explorer 5.5 seems to behave the same way.

Mine too, as does my IE4.0. But I reckon it maybe due to those being
installed on WinXP with SP2. For a real test, you should look at an
installation on Win98 or Win95 I think.

--
Els http://locusmeus.com/
Sonhos vem. Sonhos vão. O resto é imperfeito.
- Renato Russo -

[Spartanicus]

"Jukka K. Korpela" <jk******@cs.tut.fi> wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.

I'm pretty sure that previous versions of IE rendered them by the
specifications. I first thought this has something to do with XML (i.e.
maybe IE pretends to play a little bit of XML game when processing HTML,
even when I serve it as text/html and use an HTML 4.01 DOCTYPE). But on
closer look, I notice that
&auml &eacute
work correctly.

I get the same rendering on IE6 6.0.2800.1106, the about dialog says
that SP1 is installed.

--
Spartanicus

[Andreas Prilop]

On Tue, 11 Oct 2005, Jukka K. Korpela wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.
Google does the same.
http://www.google.com/search?q=%26om...te:helsinki.fi
There must be _lots_ of legacy
documents that play fast with entity references without semicolons.)

They deserve it to be punished. ;-)
Actually, I doubt that there a lots of such documents.

[Jukka K. Korpela]

Andreas Prilop wrote:

On Tue, 11 Oct 2005, Jukka K. Korpela wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.
Google does the same.
http://www.google.com/search?q=%26om...te:helsinki.fi

Most interesting. Do you know whether Google treats any other
non-letters than "&" as something that it looks for, instead of treating
it just separators between "words"?

There must be _lots_ of legacy
documents that play fast with entity references without semicolons.)

They deserve it to be punished. ;-)

Maybe. The world is cruel but unjust. I still wonder why Microsoft
thinks they can afford that - and do it for non-Latin-1 characters only.

(I checked on Windows 98 with IE 6.0, and noted that it too displays
&euro literally. Maybe my memory does not serve me right, or maybe I'm
just getting mad. Maybe I just imagined that some IE got this by the specs.)
Actually, I doubt that there a lots of such documents.

You're right in the sense that most documents that sloppily (or, in rare
cases, intentionally) omit the semicolon usually do that for ISO Latin 1
characters only. But search for &omega without site restriction gives
43,500 hits, and although many of them actually discuss the &omega;
or &Omega; entity reference, there are lots of documents (beyond the
pages of my alma mater too) that really try to present the Greek letter
in a manner that fully conforms to HTML specifications, though not to
good style.

With &euro, I get 786,000 hits, and nausea. (It's not _that_ bad, since
most people who know what "euro" means probably figure out what
"42 &euro" is supposed to mean. I'm not so sure about &harr for example.
Not to mention &lang, which is mostly just a syntax error.)

[David Ross]

"Jukka K. Korpela" wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.

I'm pretty sure that previous versions of IE rendered them by the
specifications. I first thought this has something to do with XML (i.e.
maybe IE pretends to play a little bit of XML game when processing HTML,
even when I serve it as text/html and use an HTML 4.01 DOCTYPE). But on
closer look, I notice that
&auml &eacute
work correctly.

It seems that ISO Latin 1 entities work without a semicolon, newer
entities don't. Am I mistaken, i.e. was this always the case on IE?

(Of course, it has always been recommendable to use the terminating
semicolon in entity references, but it is not required by SGML, hence by
classic HTML rules, when the reference is followed by a character that
is not a name character, e.g. by a space. There must be _lots_ of legacy
documents that play fast with entity references without semicolons.)

Section 5.3 of the HTML 4.01 specification says:
"Note. In SGML, it is possible to eliminate the final ";" after a
character reference in some cases (e.g., at a line break or
immediately before a tag). In other circumstances it may not be
eliminated (e.g., in the middle of a word). We strongly suggest
using the ";" in all cases to avoid problems with user agents that
require this character to be present."

Thus, the absence of a semi-colon is not necessarily wrong even if
it is contrary to a strong suggestion to always have the
semi-colon.

--

David E. Ross
<URL:http://www.rossde.com/>

I use Mozilla as my Web browser because I want a browser that
complies with Web standards. See <URL:http://www.mozilla.org/>.

[Andreas Prilop]

On Tue, 11 Oct 2005, Jukka K. Korpela wrote:

Do you know whether Google treats any other
non-letters than "&" as something that it looks for, instead of treating
it just separators between "words"?

Period and comma inside numbers. Others I don't know.

[Guy Macon]

Andreas Prilop wrote:

Jukka K. Korpela wrote:

Do you know whether Google treats any other
non-letters than "&" as something that it looks for, instead
of treating it just separators between "words"?

Period and comma inside numbers. Others I don't know.

It just so happens that I do know... <grin> It is something that
a lot of folks like to reference, so I will post it under a more
descriptive subject line.

[Lachlan Hunt]

Jukka K. Korpela wrote:

I noticed that Internet Explorer (6.0, on Win XP SP 2, all fixes
installed) incorrectly renders e.g.
&harr &euro &Omega
literally and not as characters denoted by the entities, but if a
semicolon is appended to each of the entity references, they work.
...
It seems that ISO Latin 1 entities work without a semicolon, newer
entities don't. Am I mistaken, i.e. was this always the case on IE?

Support in IE actually looks a little worse than that. It is, indeed,
as you mentioned, IE only supports ISO-8859-1 entity references without
the semi-colon, plus &amp, &lt, &gt and &quot from the markup
significant category. All others are only supported with the semi-colon.

However, it also fails to support any hex character reference without
the semi-colon, but does support numeric (decimal) character references
just fine with or without.

See the tests:
http://lachy.id.au/dev/markup/tests/...charref/syntax

--
Lachlan Hunt
http://lachy.id.au/
http://GetFirefox.com/ Rediscover the Web
http://GetThunderbird.com/ Reclaim your Inbox