Storing HTML in XML

[bissatch]

Hi,

Is it possible for me to store HTML tags inside XML nodes? I need some
way to share news headlines. Because the headlines differ in their
presentsation, it would be very difficult to store simply the title and
link. If possible, how would I do this?

Burnsy

 

[Joris Gillis]

Is it possible for me to store HTML tags inside XML nodes? I need some
way to share news headlines. Because the headlines differ in their
presentsation, it would be very difficult to store simply the title and
link. If possible, how would I do this?

If the HTML is well-formed, you can treat it as X(HT)ML and at the nodes to your xml document

 

[dingbat]

Is it possible for me to store HTML tags inside XML nodes?

Yes, but it's not pretty.
http://diveintomark.org/archives/2004/02/04/incompatible-rss

I need some way to share news headlines.

Then use RSS 1.0 or Atom 1.0
This is very much a ready-invented wheel.

http://xml.coverpages.org/ni2005-07-15-a.html

 

[dingbat]

If the HTML is well-formed, you can treat it as X(HT)ML
and at the nodes to your xml document

This is problematic (unworkably so, in my enormous experience of doing
it).

- It's probably a fragment, not a whole HTML document.

- If it is a fragment, then it may have multiple root elements, or non
at all. You can manipulate this in XML, but you have to be careful to
use fragment tools on it, not node trees.

- If it's HTML, you just can't guarantee well-formedness. Even quite
well-behaved HTML can omit closing tags, especially if it's an
arbitrary selection from a larger page.

- There's the issue of HTML entities that aren't declared in XML.

- Externally supplied HTML will have garbage in it - one day.

- HTML isn't XML. Applying XML rules to it, such as minimising a
non-empty element with no content (like <script src="foo" ></script> )
can cause no end of trouble downstream.

 

[Nick Kew]

Yes, but it's not pretty.
http://diveintomark.org/archives/2004/02/04/incompatible-rss




Then use RSS 1.0 or Atom 1.0
This is very much a ready-invented wheel.

Hehe. RSS has clearly gone the way of HTML. Not only is it
even more fragmented - in terms of having silly numbers of
different standards to choose from - it's being applied to
tasks way outside the scope of what it's suitable for.

That of course is the consequence of real-world popularity.

 

[Joris Gillis]

Hi Andy,

I stated this wrong. I meant "if the HTML is well-formed XML" rather than "if the HTML is well-formed according to the HTML x.xx recommendation"

This is problematic (unworkably so, in my enormous experience of doing
it).

- It's probably a fragment, not a whole HTML document.

- If it is a fragment, then it may have multiple root elements, or non
at all. You can manipulate this in XML, but you have to be careful to
use fragment tools on it, not node trees.

- If it's HTML, you just can't guarantee well-formedness. Even quite
well-behaved HTML can omit closing tags, especially if it's an
arbitrary selection from a larger page.

- There's the issue of HTML entities that aren't declared in XML.

- Externally supplied HTML will have garbage in it - one day.

- HTML isn't XML. Applying XML rules to it, such as minimising a
non-empty element with no content (like <script src="foo" ></script> )
can cause no end of trouble downstream.

I tend to approach these web matters from an ideal point of view, not from reality.

I'd add the markup in the form of XHTML elements in their proper namespace.
But then again, I'm not a developer, just a hobbyist. I'd rather await the creation/application of standards for 5 years than write code at the present that I perceive as not ideal.

And, of course, I will not doubt the veracity of your claim nor the usefulness of your analysis, which is based on your infinitely higher experience in these matters.

regards,

 

[Peter Flynn]

Hehe. RSS has clearly gone the way of HTML. Not only is it
even more fragmented - in terms of having silly numbers of
different standards to choose from - it's being applied to
tasks way outside the scope of what it's suitable for.

Yes. Trash it and use Atom.

///Peter

 

[Andy Dingley]

Hehe. RSS has clearly gone the way of HTML.

Oh, it's _much_ worse than that!
You know my opinion of Dave Winer - 'nuff said.

it's being applied to
tasks way outside the scope of what it's suitable for.

Not at all. RSS 1.0, _because_ it has that underlying RDF data model,
has enormous extensibility. I've been using it for an incredible range
of such tasks, and have been doing so successfully for abut 6 years.
With RSS 1.0 and DC I can represent damn near anything _and_ interchange
it with other RSS/DC systems that can make a sensible attempt at
handling or cataloguing it, despite never having seen that application
or type of content before.

RSS 2.0 is of course beneath contempt. Jury's still out on Atom, but
the 0.3->1.0 debacle didn't help its case.

 

[Andy Dingley]

Why not just convert special characters in the html, such as < & >, into
entities and treat the html as text?

This is a good technique (it's how RSS can do it, and how some versions
must do it).

One caveat is that you must _always_ do this. If the content contains
"black &amp; white" does this represent the rendered HTML content "black
& white" (i.e. it has been encoded), or is it really "black &amp;
white", such as might appear in a HTML tutorial ? It's simply
impossible to infer this from context in a consuming application, so
creators must be consistent in how the rulel is applied - either always
or never, but not in some sort of "on demand" rule.

Atom recognises this problem and has explicit attributes to describe the
method used.

 

[Malcolm Dew-Jones]

(e-mail address removed) wrote:
: Hi,

: Is it possible for me to store HTML tags inside XML nodes? I need some
: way to share news headlines. Because the headlines differ in their
: presentsation, it would be very difficult to store simply the title and
: link. If possible, how would I do this?

Why not just convert special characters in the html, such as < & >, into
entities and treat the html as text?

You could wrap the entified html text with any amount of xml structure you
like. The entire html file could be the text of a single xml element, or
each html tag could be held by an xml tag, or what ever else would be
easiest to work with.

<the-entire-html-file>
&gt;html&lt; &gt;head ... etc ...
</the-entire-html-file>

<a-tag original="&gt;html&lt;" />
<a-tag original="&gt;head&lt;" />
<a-tag original="&gt;title&lt;" />This is the original text
<an-end-tag original="&gt;/title&lt;" />
<an-end-tag original="&gt;/head&lt;" />
<a-tag original="&gt;body&lt;" />welcome to my web site
<an-end-tag original="&gt;/body&lt;" />
<an-end-tag original="&gt;/html&lt;" />

or what ever

$0.10

 

[Nick Kew]

(e-mail address removed) wrote:
: Hi,

: Is it possible for me to store HTML tags inside XML nodes? I need some
: way to share news headlines. Because the headlines differ in their
: presentsation, it would be very difficult to store simply the title and
: link. If possible, how would I do this?

Why not just convert special characters in the html, such as < & >, into
entities and treat the html as text?

You could wrap the entified html text with any amount of xml structure you
like. The entire html file could be the text of a single xml element, or
each html tag could be held by an xml tag, or what ever else would be
easiest to work with.

<the-entire-html-file>
&gt;html&lt; &gt;head ... etc ...
</the-entire-html-file>

<a-tag original="&gt;html&lt;" />
<a-tag original="&gt;head&lt;" />
<a-tag original="&gt;title&lt;" />This is the original text
<an-end-tag original="&gt;/title&lt;" />
<an-end-tag original="&gt;/head&lt;" />
<a-tag original="&gt;body&lt;" />welcome to my web site
<an-end-tag original="&gt;/body&lt;" />
<an-end-tag original="&gt;/html&lt;" />

or what ever

That would be
<html:element name="html" id="elt0">
<html:element name="head" id="elt1">
<html:element name="title" id="elt2">
<html:text id="text0">This is the original text</html:text>
</html:element>
.... etc
And for those entities:
<html:entity type="alpha" name="amp" id="ent0"/>

Works very well, and of course is easy either to
manipulate or to reconstruct the original from.
All it needs is an HTML parser to construct -
well-formedness of the original HTML is not a requirement.

$0.10

Inflation?

 

[Shmuel (Seymour J.) Metz]

Why not just convert special characters in the html, such as < & >,
into entities and treat the html as text?

That wouldn't have the same semantics. If the OP wants to eventually
render the text properly, then he must eventually serve, <b> as <b>,
not as &lt;b*gt;.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to (e-mail address removed)

 

[Keith Davies]

That wouldn't have the same semantics. If the OP wants to eventually
render the text properly, then he must eventually serve, <b> as <b>,
not as &lt;b*gt;.

It should be fine, as long as it's applied correctly.

Consider: You're treating the inbound HTML as plain text. Plain text
must be correctly escaped. Thus, do (in Perl)

$html =~ s/\&/\&amp;/g;
$html =~ s/</\&lt;/g;
$html =~ s/>/\&gt;/g;

This will correctly handle all conversioned necessary for these
characters ("&amp;" becomes "&amp;amp;", etc.). On extracting from the
XML container, do

$html =~ s/\&gt;/>/g;
$html =~ s/\&lt;/</g;
$html =~ s/\&amp;/\&/g;

(to be honest, I don't remember if you have to escape the & in the
first part, but it harms nothing)

This will correctly and adequately handle the escaping. Now, if you put
broken HTML in, it'll still be broken coming out... but you'll get back
what you put in, at least. Assuming nothing goofy like whitespace
removal happens, of course.


Keith