Word XHTML - Mapping styles to semantics

re: Word XHTML - Mapping styles to semantics

Thu, 08 Jun 2006 16:46:03 GMT

If you have a "strong" or "emphasis" style, then you do clearly have indication of semantics, so or would be appropriate. However, there are many, many ways in which bold and italic styles are used even in English writing practice, let alone other traditions--just note the number of other semantic tags in HTML, that by no means cover the full range of usage. So for text marked as bold or italic without a "strong" or "emphasis" style, it seems to me that and would not be appropriate. If you don't like or , you can always use CSS inline, but and are perfectly reasonable things to use within a blog entry, which is after all usually more a personal essay than a presentation of structured content--most blog structure is outside of the individual entry.

re: Word XHTML - Mapping styles to semantics

bart roozendaal — Thu, 08 Jun 2006 17:31:01 GMT

It may be my knowledge of English, but in this context, to me 'appropriate' means: what does reflect the *intention* of an author best.

And almost all of times an author wants to *emphasis* some piece of text and does not (or should not) care how this task is accomplished. That is (or should be) up to a graphic/visual/publication designer. For one publication presenting the text as bold is best; for others choosing a different color can be better. What's even more: such choices can change over time. (so can the selection of text to emphasis be, but that's another discussion).

If you have a choice: always go for semantics and leave presentation to a later stage of the production of the document. If possible, that should be dealt with at reading time (in this case by applying a CSS). At Sevensteps, we are very, very strict in this from the beginning, and has always helped us accomplish the task.

XHTML output part 3

Joe Friend: Microsoft Office Word — Thu, 08 Jun 2006 18:38:27 GMT

Zeyad has posted more information on the XHTML output from Word's new blog feature: Word XHTML - Mapping...

re: Word XHTML - Mapping styles to semantics

John — Thu, 08 Jun 2006 19:25:10 GMT

Custom Styles would be a great benifit to me and probally many others. That's one of the things that right now will limit my use of Word in this regard. Keep up the good work!

re: Word XHTML - Mapping styles to semantics

Peter Sefton — Thu, 08 Jun 2006 23:36:48 GMT

Zeyad, I'd like to talk about all kinds of lists - and how they might interact with other styles.

To get good list support in XHTML styles would be a good way to go. You can set up styles that work together. Eg how about a set of numbered styles like:

ListNum1 ... ListNum5
ListBull1 ... ListBull5
Quote1 .... Quote5
Preformat1 ... Preformat5
dt1 ... dt5
dd1 ... dd5

This means you can put a Quote2 style after a ListNum1 and the formatter will know that you want to embed the <blockquote> in the <li>. I work on a project where we have built a reasonably complete set of such styles and the formatter to output XHTML. It works well, via a custom nested styles menu. You can see a screenshot in this post of mine: http://ptsefton.com/blog/2006/06/07/word_lists_without_tears

Without the styles I doubt you will be able to build good support for nested lists combined with other semantics like quotes or code sample. With styles it is trivial to code.

If you can't add support for style-based lists as I suggest here then how about adding hooks for a user-defined XSLT stylesheet and/or macros that have access to the underlying Open XML. That way my team can write it for you :-).

re: Word XHTML - Mapping styles to semantics

KJK::Hyperion — Fri, 09 Jun 2006 02:24:03 GMT

Actually, Word (as far back as the venerable 97 edition) already comes with "emphasis" and "strong emphasis" styles (translating of course into normal + italic and normal + bold by default) in its default stylesheet, so Word internally can preserve the semantics. It'd then be just a matter of user interface, hooking up the "bold" and "italics" buttons to the styles (apropos, do styles finally cascade in Word 2007?) rather than creating a new range that subclasses the existing style. You could then define some sort of mechanism to map Word styles into CSS styles and XHTML elements, only resorting to for nameless styles

While I'm here: can lists nest in Word 2007?

re: Word XHTML - Mapping styles to semantics

Kiko — Fri, 09 Jun 2006 04:02:54 GMT

I think it would be better if quotes are mapped to <blockquote></blockquote>, for cases wherein the blog site is going for XHTML Strict. In Strict, blockquotes must only contain block-level elements.

re: Word XHTML - Mapping styles to semantics

Zeyad Rajabi — Mon, 12 Jun 2006 19:03:51 GMT

Y I do agree that there are occasions that having and are what the user really intended, but I think that scenario is less likely than having and . Also, our UI tends to favor the usage of and rather than the use of styles. For those two reasons, we decided to output just styles: and .

Peter thanks for the link. I will take a look at your post and see what we can do. In the meanwhile, I should mentioned that we do provide ways to crack open our Open XML file formats. Please reference OpenXMLDeveloper.org. In that way, you guys would be able to write transforms that will convert our new XML files to whatever output you wish. You may also reference Kevin’s blog at http://blogs.msdn.com/kevinboske/.

KJK:Hyperion – Our XHTML output preserves nested lists. However, our output does not match the full power of Word 2007.

Zeyad

re: Word XHTML - Mapping styles to semantics

Peter Sefton — Tue, 13 Jun 2006 06:23:23 GMT

I have written a little more on this, with more concrete examples of list formatting. http://ptsefton.com/blog/2006/06/13/list-samples

Opening up the file and transforming it is fine - but that's not as good as having a built-in mechanism for export. Images are one big issue here. Word has always done a lovely job of exporting images to HTML (since word 97 anyway) whereas if you work only at the file format level you won't get web-ready images.

If you guys don't want to add an XSLT export facility like OpenOffice.org (theirs is broken, doesn't do images at all) then please at least export style names as classes and we can post-process word docs to add correct nested structures.

re: Word XHTML - Mapping styles to semantics

Mr. Dew — Sun, 18 Jun 2006 21:35:35 GMT

I like the custom classes. I hope it works for images too! I have classes like leftalign and rightalign to align my images. :D

Word XHTML - Bullets and Numbering

Brian Jones: Open XML Formats — Wed, 12 Jul 2006 19:39:25 GMT

This is the fourth post by Zeyad Rajabi who owns the XHTML output from Word's new blogging feature. In...

re: Word XHTML - Mapping styles to semantics

Ray — Fri, 14 Jul 2006 22:23:12 GMT

Hi Brian,
I have a question regarding run properties defined in styles (sorry this is a bit off topic, not related to xhtml) :
Let say I have defined two styles, one for a paragraph, the other for a run, as follows :

<style name="P">
<rPr><sz val="56"/></rPr>
</style>

<style name="R">
<rPr><sz val="24"/></rPr>
</style>

now, if I have the follwing content :

<pPr><pStyle="P"/></pPr>
<r>word1</r>
<r><rStyle="R"/>word2</r>


then when opening in Word 2007 Beta, "word1" appears with size 26 (ok) and font weight is bold (ok)
"word2" appears with size 12 (ok) and font weight bold (not ok!!)

now, if I change the run property (in "R" style) and switch it "on" -> "word2" font-size is 12 and font-weight is not bold anymore...

What I don't understand is that defining run properties in a style and applying it to a paragraph seems make all the runs inherit those properties. Except if they override them. It works for <sz> but not for (worst , I have to turn it on!).
I don't understand this behavior, can you put the light on this, please?

thanks.

re: Word XHTML - Mapping styles to semantics

Zeyad Rajabi — Fri, 21 Jul 2006 02:58:28 GMT

Hi Ray,

Tristan has answered your question at:

http://openxmldeveloper.org/forums/thread/371.aspx

Zeyad Rajabi (MS)

re: Word XHTML - Mapping styles to semantics

Kevin Yank — Fri, 21 Jul 2006 04:23:12 GMT

It's great to see you putting so much thought into semantics as you develop the Word blogging feature, Zeyad. On behalf of the web standardistas of the world, we thank you!

I think the nested pre and code elements for the "Code" style make perfect sense as long as what is being presented is a block of code. In many cases, however, an author will want to simply present a single keyword, or short code snippet, and for that I think an "Inline Code" style that simply mapped to a code element would be very useful.

To satisfy those calling for greater control over semantic mappings (such as whether clicking the bold and italic buttons should produce strong and em elements, respectively), you would need to provide a preferences page with a series of mappings that could be switched on and off. Here are some ideas for what such a list might contain:

- Output bold as strong
- Output italic as emphasis
- Output indent as quotation
- Output strikethrough as deleted text
- Output underline as inserted text
- Discard non-semantic formatting (alignment, font, size, color, etc.)

The ability to define custom style mappings to HTML classes is an excellent feature. I very much hope you can get this working!

Word XHTML - Tables

Brian Jones: Open XML Formats — Tue, 15 Aug 2006 05:51:51 GMT

This is the fifth post by Zeyad Rajabi who owns the XHTML output from Word's new blogging feature. In...

Brian Jones Office Extensibility Word XHTML Mapping styles to | bird baths

Brian Jones Office Extensibility Word XHTML Mapping styles to | bird baths — Sun, 14 Jun 2009 18:08:46 GMT

PingBack from http://cutebirdbaths.info/story.php?id=3665