HML5 and micro data validation

[jpmoularde]

Hello!
I'm introducing the "microdata" in my web site, according to http://schema.org/
It consists of adding a few attributes which are interpreted by web search engines, specifically "itemscope", which tells that it's the beginning of a microdata object, "itemtype", which refers to the part of schema.org which defines that object, "itemprop" which gives the name of the corresponding property and "content" which allows to define a standard content for a property, e.g.
<span itemprop="startDate" content="2014-01-01">1st of January 2014</span>

Apparently these attributes are not recognized by the validator used in XML Editor 15.1, build 2013101713

Is there a way to solve that issue?
My html pages are not correctly analysed because these attributes are rejected. Per se it's not a problem, but that behaviour prevents me to see "real" errors, like unclosed tags.

Thanks for your help

[Radu]

Hi Jean-Pierre,

There are two problems with microdata and Oxygen.

1) The microdata specification is mostly for HTML (and not well-formed XHTML) so it has constructs like:

Code: Select all

<div itemscope itemtype ="http://schema.org/Movie">



....

which are not valid XML because the "itemscope" attribute does not have a value.
But this can be fixed:

http://stackoverflow.com/questions/3659 ... -compliant

using the construct like:

Code: Select all

<div itemscope="itemscope" itemtype ="http://schema.org/Movie">



....

2) The schema Oxygen uses to validate XHTML content does not understand the microdata constructs.

I'll add a report on our internal issues list and we'll see if we can do something about this in a future Oxygen version. At least to skip these attributes and validate the rest of the XML content.

Regards,
Radu

[jpmoularde]

Thanks Radu
When I say "HTML5", I mean XHTML, of course. I hardly understand why people continue using non xml compliant html, even when they establish a new standard! Personally I developed a small HTML generator in php, which takes care of unclosed marks and forces attributes to have a value.
So I had the itemscope="itemscope" issue already fixed in my files.
I hope the schema issue will be fixed soon.

[Radu]

Hi Jean-Pierre,

You can edit the following RNC schema used to provide validation for XHTML 5 when working in Oxygen:

OXYGEN_INSTALL_DIR\frameworks\xhtml\xhtml5 (epub3)\mod\html5\html5-attrib-30.rnc

and add at the end of the file additional attribute definitions for the microdata attributes:

Code: Select all

   ##Microdata extension

   html5.global.common.attrs &= micro.scope.attr? & micro.prop.attr? & micro.type.attr?

   ## Microdata item scope. Identifies the purpose of a certain item.

   micro.scope.attr = attribute itemscope { string }

   ## Microdata item type. Identifies the type of a certain item.

   micro.type.attr = attribute itemtype { string }

   ## Microdata item property. Used to add property labels to an item.

   micro.prop.attr = attribute itemprop { string }

We'll probably add these definitions by default in the RNC which comes with Oxygen 15.2.
Are there any additional microdata attributes which need to be handled similarly?

Regards,
Radu

[jpmoularde]

Thanks Radu
The trouble is I don't know how to authorize myself to overwrite that file on my PC (windows 7).
It's the first time in years I have to do that, the way to do has apparently changed from the last time I did it under Windows XP, and the documentation is less than easy to understand.
Regard
JP

[Radu]

Hi Jean-Pierre,

As a workaround you can create a folder named frameworks outside of the Program Files location to a place where you have full access rights.
In that folder you should then copy the entire xhtml framework folder from:

OXYGEN_INSTALL_DIR\frameworks\xhtml

Then in that copied folder modify the RNC file from:

xhtml5 (epub3)\mod\html5\html5-attrib-30.rnc

the way I explained in the previous email.

Then from Oxygen in the Preferences->Document Type Association / Locations page add your custom frameworks folder to the Additional frameworks directories list and restart Oxygen. Your custom document type should be used on the next restart.

Regards,
Radu

[jpmoularde]

Thanks, Radu
I'll try it today and keep you informed

[jpmoularde]

Hello!
I noticed that the latest version of XML Editor includes the "itemprop", "itemscope" and "itemtype" attributes.
But there is still an attribute missing, "content", which allows to give the value of a parameter in a conventional format.
Dates for instance have to be coded in the ISO standard format rather than any of the many human readable formats, as shown below:

Code: Select all

<div class="eventdates">

    From <span itemprop="startDate" content="2014-03-10">10</span> 

    through <span itemprop="endDate" content="2014-03-13">13 March 2014</span>

</div>

[Radu]

Hi,

I'm not sure that the @content attribute should be available on every XHTML 5 element, it seems that it is already present on <meta> tags:

http://html5doctor.com/microdata/#meta

and you can refer to it using @itemref (which we should add to the schema bundled with Oxygen as well along with @itemid).

I cannot find an official specification for Microdata though.

Regards,
Radu

[jpmoularde]

Hi

I cannot find an official specification for Microdata though.

To me the "Bible" is http://schema.org and -- you're right -- there doesn't seem to be another way of expressing a date than including it in a <meta> tag.

[Radu]

Hi,

Thanks for the confirmation.
For the next version of Oxygen I updated the RNC file OXYGEN_INSTALL_DIR\frameworks\xhtml\xhtml5 (epub3)\mod\html5\html5-attrib-30.rnc to recognize the attributes:

Code: Select all

itemscope

itemtype

itemid

itemprop

itemref

If you want to make the changes on your side, the contents of the html5-attrib-30.rnc should be:

Code: Select all

html5.global.attrs &= html5.global.common.attrs & html5.global.xml.attrs & html5.microdata.attrs

html5.i18n.attrs = html5.xml.lang.attr? & html5.lang.attr? & html5.dir.attr?

   

html5.global.xml.attrs &= html5.xml.lang.attr? & html5.xml.base.attr? & html5.xml.space.attr?   

html5.global.common.attrs &= html5.lang.attr? & html5.id.attr? & html5.title.attr? & html5.dir.attr? & html5.class.attr?







html5.lang.attr = attribute lang { datatype.languagecode }

html5.xml.lang.attr = attribute xml:lang { datatype.languagecode }       

html5.dir.attr = attribute dir { 'ltr' | 'rtl' | 'auto' }  

html5.id.attr = attribute id { datatype.ID }   

html5.title.attr = attribute title { string }   

html5.class.attr = attribute class { datatype.html5.space.separated.tokens }   

html5.xml.base.attr = attribute xml:base { datatype.LEIRI }

html5.xml.space.attr = attribute xml:space { 'preserve' | 'default' }



html5.href.attr = attribute href { datatype.html5.URL.spaces }

html5.hreflang.attr = attribute hreflang { datatype.languagecode }      

html5.target.attr = attribute target { datatype.html5.browsing.context.name.or.keyword } 

html5.rel.attr = attribute rel { datatype.html5.space.separated.tokens }  

html5.media.attr = attribute media { datatype.html5.media.query }   



html5.type.mime.attr = attribute type { datatype.mimetype }

html5.type.mime.nocharset.attr = attribute type { datatype.mimetype.nocharset }



html5.charset.attr = attribute charset { datatype.character.encoding }



html5.src.attr = attribute src { datatype.html5.URL.spaces }

html5.usemap.attr = attribute usemap {datatype.html5.name.reference}



html5.cite.attr = attribute cite { datatype.html5.URL.spaces }   

html5.alt.attr = attribute alt { datatype.string }



html5.name.context.attr = attribute name { datatype.html5.browsing.context.name }

html5.name.id.attr = attribute name { datatype.html5.name } 



html5.dimension.attrs &= html5.width.attr? & html5.height.attr?      

html5.width.attr = attribute width { datatype.non-negative.integer }       

html5.height.attr = attribute height { datatype.non-negative.integer }



##EXM-29190 - Microdata extension

html5.microdata.attrs &= micro.scope.attr? & micro.prop.attr? & micro.type.attr? & micro.itemref.attr? & micro.itemid.attr?

## Microdata item scope. Creates the Item and indicates that descendants of this element contain information about it.

micro.scope.attr = attribute itemscope { string }

## Microdata item type. A valid URL of a vocabulary that describes the item and its properties context.

micro.type.attr = attribute itemtype { string }

## Microdata item property. Indicates that its containing tag holds the value of the specified item property. 

micro.prop.attr = attribute itemprop { string }

## Microdata itemref property.  Properties that are not descendants of the element with the itemscope attribute can be associated with the item using this attribute.  

micro.itemref.attr = attribute itemref { string }

## Microdata itemid property. Indicates a unique identifier of the item. 

micro.itemid.attr = attribute itemid { string }

Regards,
Radu

[jpmoularde]

Thanks

By the way, the correct way of coding dates in HTML5 seems to be

Code: Select all

<time datetime="2014-03-13" itemprop="endDate">March 13, 2014</time>

[jpmoularde]

Hello!
It seems that the HTML 5 validator is not in line with specifications on the following items:

• <meta> tags in <body>
• "propref" attribute
• "datetime" attribute in <time>

and maybe some other ones.
Can we expect an update of the validator?

[Radu]

Hi,

<meta> tags in <body>

I don't see anything in the specs saying that <meta> is allowed in <body>, could you give me a link to that particular reference in the specs?

"propref" attribute

The same, I cannot find anything about the "propref" attribute in the W3C specs.

"datetime" attribute in <time>

A construct like:

Code: Select all

<time datetime="2014-03-13" itemprop="endDate">March 13, 2014</time>

should be valid in Oxygen from what I tested if the XHTML document is interpreted as XHTML 5. Does your document has a DOCTYPE declaration?



Regards,
Radu

[jpmoularde]

As said above, all items I'm considering in that post are related to microdata

<meta> tags in <body>
I don't see anything in the specs saying that <meta> is allowed in <body>, could you give me a link to that particular reference in the specs?

see http://www.whatwg.org/specs/web-apps/cu ... ta-element
It says that <meta> tags containing an "itemprop" attribute are allowed in "flow content" and "phrasing content", i.e. anywhere in the <body>

"propref" attribute
The same, I cannot find anything about the "propref" attribute in the W3C specs.

see http://www.w3.org/TR/2011/WD-microdata-20110525/
That attribute is used to refer to a part of the HTML file which is not included in the current parameter set, for instance:

Code: Select all

<div id="Monteverdi" itemprop="author" itemscope="itemscope" itemtype="http //schema.org/Person">...</div>

...

...

<div itemscope="itemscope" itemtype="http://schema.org/CreativeWork" itemref="Monteverdi">...</div>

which means that the <div> with 'id="Monteverdi"' refers to the author of the work described later.

"datetime" attribute in <time>
A construct like:

Code: Select all

<time datetime="2014-03-13" itemprop="endDate">March 13, 2014</time>

should be valid in Oxygen from what I tested if the XHTML document is interpreted as XHTML 5. Does your document has a DOCTYPE declaration?

see http://dev.w3.org/html5/markup/time.html
Yes, my document contains a DOCTYPE declaration, but if

Code: Select all

<time datetime="2014-03-13" itemprop="endDate">March 13, 2014</time>

is accepted, another expression like

Code: Select all

<time datetime="2014-03" itemprop="datePubllished">March 2014</time>

is not accepted.
I know that the standard has been modified several times; <time> has even been deleted from the documents for some time, but -- as much as I know -- it has been reintroduced about 2 years ago with a lot of extensions.
see: http://www.webmonkey.com/2012/02/the-ht ... than-ever/

Regards
JP

[Radu]

Hi,

1) According to the specs the <meta> element is not accepted as flow content:

http://www.w3.org/html/wg/drafts/html/m ... -content-1

2) There is no propref attribute, I think you mean itemref which will be recognized as an attribute in Oxygen 16. I already gave you in a reply above what changes to make to the html5-attrib-30.rnc in order for itemref to be recognized.

3) About the time datetime allowed patterns, I think you are right, it should also allow patterns like "2014-03" the specs:

http://www.w3.org/html/wg/drafts/html/m ... e-datetime

We'll see if we can rectify this on our side in the XHTML 5 schemas.

Regards,
Radu

[jpmoularde]

1) According to the specs the <meta> element is not accepted as flow content:
http://www.w3.org/html/wg/drafts/html/m ... -content-1

It's not what I read in that very document:

3.2.4.1.2 Flow content

Most elements that are used in the body of documents and applications are categorized as flow content.

a
abbr
.......
link (if the itemprop attribute is present)
.......
meta (if the itemprop attribute is present)

So if there is an "itemprop" attribute, a <meta> tag can be included in the <body>.

2) There is no propref attribute, I think you mean itemref which will be recognized as an attribute in Oxygen 16. I already gave you in a reply above what changes to make to the html5-attrib-30.rnc in order for itemref to be recognized.

You're right, my mistake!

3) About the time datetime allowed patterns, I think you are right, it should also allow patterns like "2014-03" the specs:

http://www.w3.org/html/wg/drafts/html/m ... e-datetime

We'll see if we can rectify this on our side in the XHTML 5 schemas.

Thanks

[Radu]

Hi,

You are right about the remark:

meta (if the itemprop attribute is present)

We'll see if we can make some changes to the schemas used for HTML 5 validation to accommodate this as well.

Regards,
Radu

[jpmoularde]

Thanks
It's not very important, as most of my web pages are generated by a program in php which of course does not submit the result. But for the few handwritten html files I have, I prefer checking their validity first. As the validator stops at the firs "error" it encounters", the less "false errors" there is, the best it is.
Regards
JP

[Radu]

Hi,

Probably depends on the case but usually validation should not stop at the first reported problem (unless it is considered to be fatal in some way).
If you have a precise case in which validation stops at the first problem, just give us a small sample and we'll see if we can improve this behavior.

Regards,
Radu

[jpmoularde]

Hi Radu!

It does not really stop, you're right: there is that series if red bars in the rightmost part of the window that shows lines where errors have been found, so there can be a large number of errors found in one shot.
Thanks for drawing my attention on that point, it will save me time.
Regards
JP