W3C home > Mailing lists > Public > www-html@w3.org > September 2001

Authoring tool specific attributes in XHTML 1.0+

From: Nicholas Atkinson <nik@casawana.com>
Date: Tue, 11 Sep 2001 12:16:39 +0100
Message-ID: <005601c13ab3$5a661100$0300000a@natkinwkstn>
To: <www-html@w3.org>

I would like to propose that an additional "special" attribute (such as
"xml:app") be available for use on all elements in XHTML, which can be used
for "authoring tool specific purposes", whilst still allowing the document
to be validated.

Here's a real-world example:

When a web designer uses a tool such as Adobe GoLive to insert an image into
an HTML page, the tool allows the designer to select the source "asset" and
target file and optimization settings.  In practice, this is really powerful
since the source assets are high resolution, may be shared between projects
and are unoptimized (and may consist of several layers if they are Photoshop
files, for example), but the target files are in a "flattened" web browser
format (gif/jpeg/png) and are "optimized for the web".  Indeed, if you
resize an image on the page, the tool regenerates the target image.

But to do this the tool has to store additional information in the image
element.  Which unfortunately means that validating the document will fail.

It would be ideal if this tool-specific information could be put in the
value of a special attribute which UAs will ignore (and maybe even other
authoring tools if some convention were agreed based on the first few
characters of the value, or something).  But, crucially, the document would
still be valid XHTML.

There are lots of other examples where an additional attribute would be
useful.  Especially with media objects that have a "source project"
associated with them.  And also in other cases, such as where pages are
constructed in the authoring tool from a number of source page "components".

If the DTD could be extended to allow the storage of an authoring tool
specific attribute, as I propose, then the creators of authoring tools would
not need to make up proprietary attributes and tags.

I suppose I could always tweak the DTD I use for validation.  But does it
make sense to validate, if I am validating against a different DTD from
everybody else?!



----- Original Message -----
From: "Christian Wolfgang Hujer" <Christian.Hujer@itcqis.com>
To: "Matt Brooks" <matt@mbjlp.com>; <www-html@w3.org>
Sent: 05 September 2001 22:15
Subject: RE: "alt" attribute required by XHTML 1.0

> Hello Matt,
> you were ignoring the following facts (most of this was already said, but
> wanted to write a full list that I can recycle... ;) (It is great of you
> that at least you tried to validate, many so-called "web designers" or
> programmers" (ouch!) even do not know that something like validation
> at all...):
> - requirement of the alt attribute for the <img/>-element is not new. For
> HTML 4.01, HTML 4, HTML 3.2, HTML 3.0 and HTML 2.0 I am sure they also
> required the alt attribute for valid documents. For HTML+ and HTML I am
> sure, neither I am for ISO-HTML, but since ISO-HTML is the most
> HTML I've ever seen (you may not use H3 after H1, and you may not use H2
> you didn't use H1 before...), I believe it also requires the
> - that spacer-Images are *not* what HTML-Cracks use for positioning since
> this won't work as intended, anyway.
> - HTML is a language for description of document semantics, for layout use
> - that people are using text browsers if they are freaks using lynx or
> small devices that cannot display fancy graphics
> - that there are blind people surfing the web using speech browsers
> - that validation generally is a good idea
> - that if you do not write valid HTML, you are not writing HTML at all,
> are just writing some SGML or XML nonsense which the web browsers do their
> best to display
> - that the main attribute for Tooltip display is the title attribute, so
> could try to override the alt="" empty tooltip with title="" for no
> in some browsers
> - that alt="" won't display a tooltip in most browsers anyway
> There is not alternative to valid documents except for the following:
> - your XSLT transformation that generates your HTML documents is not fully
> XSLT 1.1 compliant, so no namespace cleanup is performed and you may get
> additional namespace attributes
> - you are using a language mix of HTML and another namespace like SVG,
> or MathML
> Neither is the case with your documents.
> For your <a name="label"/>-problem:
> - The correct writing is <a id="label"/>
> - It is no problem to include <a id="label"/> in block elements only, try
> using <div><a id="label"/></div> if nothing else helps, but usually <a
> id="label"/> comes in context of a document structure, so a good position
> would be within a <hn/> or <dt/> element.
> You should include div {margin:0;padding:0;} in your stylesheet to avoid
> problems with older browsers using <div/> like <p/>.
> Believe me, it is perfectly possible to write documents that
> - use "latest" features like ECMAScript ("JavaScript"), Java Applets, SVG
> Graphics and CSS Level 2 Layout
> - are fully backwards compatible to nearly every browserš
> - display great on new browsers and old browsers
> - display great on text browsers like Lynx or w3m
> - look as if they had frames in IE5, 5.5, 6 and Netscape6 / Mozilla with
> use of the CSS Level 2 overflow property though they do not use the
> document type, so they are viewable with browsers that aren't capable of
> displaying frames
> - are valid XHTML Basic 1.0, XHTML 1.0 Strict or XHTML 1.1.
> If you have questions or problems regarding this, feel free and drop a
> to me.
> I think that frames will die anyway, XHTML 1.1 does only have one document
> type, not three, that corresponds to strict; Transitional and Frameset are
> likely to die. XHTML Basic doesn't support Frames, too, of course.
> šNetscape 4 and elder require empty element tags for empty elements to use
> whitespace before the /, so write <br /> instead of <br/>;
> javax.swing.text.JEditorPane based browsers like HotJava or self-made
> Java-written browsers have a bug and will display the / in empty element
> tags. Always use ASCII only as encoding, this is UTF-8 compatible since it
> only uses characters with numbers 0-127, and omit the XML declaration,
> only is allowed if you use UTF-8 for encoding, which is the case if you
> ASCII. Encode all characters that are not ASCII-characters (US-ASCII-7)
> Umlauts with their corresponding entites.
> Greetings
> Christian
> -----Original Message-----
> From: www-html-request@w3.org [mailto:www-html-request@w3.org]On Behalf Of
> Matt Brooks
> Sent: Wednesday, September 05, 2001 6:43 PM
> To: www-html@w3.org
> Subject: "alt" attribute required by XHTML 1.0
> I was surprised to see that the ALT attribrute is required by the XHTML
> Transitional DTD. I was going to use XHTML 1.0 Transitional in a web
> development project, but have now changed my mind because the ALT
> are not needed on every image.
>  - Matt
> -----Another Original Message-----
> No, because "" produces an empty (but displayed) tooltip in some browsers.
> This is unacceptable.
Received on Tuesday, 11 September 2001 07:22:25 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:58 UTC