- From: Walter Ian Kaye <boo@best.com>
- Date: Wed, 6 Dec 1995 22:44:12 -0800
- To: html-wg@oclc.org
- Cc: www-style@w3.org
At 12:10p 12/06/95, lilley wrote: >Eddie said: > >> In message <9512052314.AA08698@sass027.sandia.gov>, Michael J Hannah writes: > >>> I wish to join the voices raised in opposition to the current mechanism >>> of use of the STYLE attribute on every HTML element as proposed >>> in this draft. >>> [snip] >>> 1. Allow the STYLE attribute on all HTML elements. > >> Am I being really blind or am I right in seeing a contradiction in these >> two sets of statements ? > >The words you are skipping over are "current mechanism of use" > >Michael was proposing a new mechanism of use; each element has a >style attribute, but the value is constrained to be a pointer to >the stylesheet (either in the head or in another document). I had suggested this very thing a short while back, and was pointed toward the CLASS attribute. I felt uncomfortable with that, but was unable to put my finger on exactly why; I can state it better now. The reason is for consistency with the STYLE element in the HEAD. It makes no sense to use CLASS unless you have CLASS in the HEAD. I'm not saying STYLE should replace CLASS. I'm saying the attribute name should match what it references. If you are using CLASS="myname", then you should be defining CLASS names in the HEAD, not STYLE names. To illustrate: <HEAD> <TITLE>The Consistent CLASS Approach</TITLE> <CLASS MIME="application/css"> mainhead {align: center; color: #005C00} H2 {font-family: Helvetica; font-size: 18; font-weight: bold} subhead1 {based-on: H2; font-style: small-caps; color: #000066} subhead2 {based-on: subhead1; font-style: italic; color: #660000} </CLASS> </HEAD> <BODY> <H1 CLASS="mainhead">Walter's Second Take at Style vs Class</H1> <H2 CLASS="subhead1">This is in dark blue</H2> <H2 CLASS="subhead2">This is in dark red, and italic</H2> </BODY> or... <HEAD> <TITLE>The Consistent STYLE Approach</TITLE> <STYLE NOTATION="css"> mainhead {align: center; color: #005C00} H2 {font-family: Helvetica; font-size: 18; font-weight: bold} subhead1 {based-on: H2; font-style: small-caps; color: #000066} subhead2 {based-on: subhead1; font-style: italic; color: #660000} </STYLE> </HEAD> <BODY> <H1 STYLE="mainhead">Walter's Second Take at Style vs Class</H1> <H2 STYLE="subhead1">This is in dark blue</H2> <H2 STYLE="subhead2">This is in dark red, and italic</H2> </BODY> With this, the attribute name is consistent and logical. >This is fine but there have been objections from those who want nameless >formatting changes all over the place. > >Actually, as a technical author, my experience editing the documents produced >by colleagues is that in most cases replacing the on-the fly formatting >with proper named styles resulted in better looking documents. Slapping a >quick bold here and a font change there just ends up looking a mess. And we already have <B>, <I>, <STRONG>, and <EM>; these are sufficient for "direct" formatting occasions. Actually, style sheets render <B> and <I> obsolete, so <STRONG> and <EM> remain. Anything more complex than that is best defined as a style (class, whatever). >Having a good set of per-document-type formatting names actually makes >writing easier (I use the term document type in the sense that lecture >notes and publicity flyers are different document types; not in the >SGML sense). I agree. I do all my web authoring by hand in a drag&drop text editor, using stuff like <font size="+2"><B><I>xxx</I></B></font>. I would GLADLY switch to something like <SPAN STYLE="subhead2">xxx</SPAN> with "subhead2" previously defined, as it makes handwritten code infinitely easier to read and maintain. To summarize my feelings: I am against <XXX STYLE="font: helvetica; size: 14"> I am for a <XXX STYLE="stylename"> I am for a <XXX CLASS="stylename"> I am against CLASS attribute referencing a STYLE element in head I am for a CLASS attribute referencing a CLASS element in head I am for a STYLE attribute referencing a STYLE element in head And I hope I've stated it more clearly this time. -Walter __________________________________________________________________________ Walter Ian Kaye <boo@best.com> | Excel | FoxPro | AppleScript | Mountain View, CA |--------- programmer ---------| http://www.natural-innovations.com/ | Macintosh | Windows |
Received on Thursday, 7 December 1995 01:44:23 UTC