Re: comments on CSS3 Selectors from XSL Working Group

C. M. Sperberg-McQueen wrote:

> We note with great concern your publication of "Selectors"[1].
> You may recall that when you originally proposed to publish this
> document with the title "W3C Selectors" back in 2001, the XSL WG
> had concerns about that title. Our concerns were alleviated by
> your decision to publish the document at that time with the title
> "CSS3 module: W3C Selectors".
> However, your decision to use simply "Selectors" as the title in
> the most recent publication causes us concern once again and we
> wonder why you would think that is is more acceptable now when it
> was not acceptable previously. As you are all aware, the CSS
> selector mechanism is not the only means by which parts of an XML
> document may be selected. XPath and XPointer both offer
> functionality that falls under the general description of
> "selection" as do several functions in the DOM, to name but a few
> of the other selection mechanisms available to users.
> We must therefore once again[2] urgently request that the
> document be titled "CSS Selectors" and not simply "Selectors" or
> the equally misleading "W3C Selectors". It is simply too
> confusing to label any one of the various mechanisms with the
> undistinguished title "Selectors"


1. the spec has been authored without restriction to CSS. It is explicitely
    made as a generic selecting mechanism that is NOT bound to Cascading
    Style Sheets. There is at least another spec called STTS implementing
    Selectors; STTS is itself implemented by a japonese company.
    The spec, as it is today, is standalone and has no generic dependency
    to CSS. CSS is mentioned in its prose only in the case of use inside CSS.

    "Selectors" is a spec authored and released by the CSS WG, that is used
    inside CSS but is explicitely made generic. Just as the HTML WG authors
    and releases generic specs used by XHTML. Did I hear anyone asking to
    rename "HLink" "XHTML HLink" ? Not really.

2. we removed "W3C" from our original title because of Xpath and
    other selecting mechanism existing in W3C, to make it very clear that
    Selectors are not the only selecting mechanism (hear selectos) inside the
    Consortium, to make sure there's no confusion between the specs.

3. Saying "CSS Selectors" implies that "Selectors" are the one and only
    selecting mechanism for CSS styles in the present but also in the future.
    There are some discussions about using XPath in CSS-based style rules;
    there are also discussions about an XMLization of CSS;
    keeping that in mind, it seems to me not desireable to title our spec "CSS
    XPath's real name is "XML Path Language". Is it the only Path Language for
    XML ? No. Selectors are also a path language.

So I disagree with the requested change. As a compromise, I suggest adding the 
following subtitle to our spec:

   "A selection mechanism for CSS 3 and other languages"

or something like that. Is that acceptable from an XSL WG's point of view ?

 > We would also like to draw your attention to the opening
 > paragraph of the abstract in that draft. It says, in part:
 >    Selectors have been optimized for use with HTML and XML, and
 >    are designed to be usable in performance-critical code.
 > We object to the characterization expressed in the above
 > sentence. It could be construed as implying that CSS selectors
 > are somehow superior to all other forms of selection in terms of
 > performance.  There is nothing in the specification to support
 > this claim.  It appears to have been added solely for effect.
 > Please remove.

Remove ? Uhhh ? Something has to be removed when it is false or
incorrect, but it IS true and correct.

This paragraph says _nothing_ about other selection mechanisms and
it's seen to me quite paranoid to see how one could understand from
this paragraph that other W3C selecting mechanism are not suboptimal
in any way.

It only means that THIS spec was designed with browser performance in
mind and that some features were explicitely dropped or designed to
meet the performance required by progressive rendering and provide
the users with a better browsing experience in dynamical environments.
For example, CSS is used to style XUL in Mozilla-based applications.
Doing that for an application's chrome requires resolving tens of
thousands of style rules in a highly dynamical environment.

The phrasing above just says that, and I find it quite concise and
accurate in that description. I fail to see how we could introduce
more political correctness here keeping the wording factual, but
it is just me. David Baron's change proposal seems to me a compromise
but a bad one, reducing readability and simplicity of the original


Received on Friday, 27 January 2006 10:39:30 UTC