I list below all topic groups, which I have done according to subjects, which they handle. You can return to this topic group by using this menu and the link Table of topic groups on the top of the each page.
| ||||||||||||||||||||
![]() | Table of topic groups > Front page of CSS-guide > Proposals > Proposals for CSS3 |
|---|
I have found some problems in CSS or I have desired of some new properties. I put in this file my proposals into the CSS3.
I have found that it is not always very handy to use CSS. Using preceding elements in tag sequences like this:
<blockquote><i>...</i></blockquote>
I can't get a matching pattern based rule to the block quote,
if I need a special rule to the element BLOCKQUOTE.
I can give it only to the element I. But this kind
of rule could do it:
blockquote < i {} /* the rule concerns
the parent element blockquote
, not the element i
; This is
different as blockquote > i {}
, which concerns the element i
*/
In order to avoid difficulties, this might need to limit only
in one level. Just to the parent element, not to all possible
ancestor elements (using in the rule for example
<<) like style="color:#660033;
background-color:yellow;". Browsers might not be able to
handle them (and those cause extra formatting work). This is much
easier afterward to exchange as these kinds rules:
<blockquote class="some_class"><i>
This might have following rule:
.some_class i {color:blue}
If you have not originally marked the class, you must search
all possible places, where you have used these elements. But this
rule blockquote < i {color:blue} you could add
afterward without touching to the document!
CSS Pattern matching would work more powerful with these new pseudo-classes. I have found also a new pseudo element useful, when I have read magazines.
:second-child pseudo-class, :third-child
pseudo-class and so on.:second-sibling pseudo-class,
:third-sibling pseudo-class and so on.:paired-paragraph pseudo-class;
an-paired-paragraph pseudo-classMy proposal to a new pseudo-element is
p:first-word - the first-word pseudo-element. In
Finland "Tekniikan maailma" use this kind of system in
paragraphs:
THE first word of each of paragraph is in upper case letters.
Look at also the next section, where I handle ::viewport.
Opera Software and Microsoft had made different conclusions of
CSS1 border display. To avoid this problem, I have a proposal. In
future CSS-definitions would be added
border-attachment -value, which could work according
to the same principles as background-attachment and
does have values scroll and fixed.
Fixed would go according to the
viewport - like MS 5.0. Scroll means
the system, which Mozilla and Opera have just now - and very
often there is no border on the top or bottom. That could be
expressed like this way:
If borders ofBODYorHTMLcontainer element is definedborder-attachment:fixed, they are fixed at the respect of the viewport and borders are placed according to the viewport instead of outside of the element's padding area.
Alternatively the same result could be achiever if the CSS3 could have the pseudo-elements ::viewport and ::canvas and borders have been set for those pseudo-elements. ::viewport could represent a single window. ::canvas represent a set of windows in the case of the FRAMESET elements has been used and a sigle window, if framesets has not been used. In the latter case ::canvas == ::viewport.
Background image size properties; image-height
and image-width. This was left open in CSS2, but it
is useful especially making different style sheets using the
media rule - small devices needs smaller images.
Indeed this can be fixed in principle by using frame sets or
with the overflow property.
CSS doesn't cover <iframe frameborder="1" style="border:1px solid black">. Now the element has double borders. CSS3 should have the frameborder property. In addition of frames, it could be useful with form control elements, which have natural borders. The frameborder attribute could be understood as inner borders.
I have also some additional comments to my proposals for CSS3.
These proposals concerns also other than CSS, but all of the proposals has something to do with CSS: