• 200618 Nov

    Recently I had to code a box with rounded corners for a client, and I came up with a solution which works with 99% of the browser market and uses the CSS3 border-radius declaration for browsers which support it.

    The simplified HTML code is:

    <div class="curves top">
    <img src="http://www.css3.info/wp-content/uploads/2009/09/topl.png" width="10"
    height="10" alt=""/>
    <div class="box"></div>
    <div class="curves bottom">
    <img src="http://www.css3.info/wp-content/uploads/2009/09/botl.png" width="10"
    height="10" alt=""/>

    The divs .top and .bottom will be hidden from browsers which support border-radius (or a proprietary implementation of it).

    The basic CSS is:

    div {
    margin: 0;
    padding: 0;
    width: 100px;
    background-color: #ff0000;
    .box {
    min-height: 100px;
    -webkit-border-radius: 10px;
    -moz-border-radius: 10px;
    border-radius: 10px;
    .curves {
    display: none;

    This will display a box 100×100 pixels with curved corners in Firefox and Safari, hiding the .top and .bottom divs and implementing border-radius.

    To make it display in IE, we add some conditional comments with extra CSS which displays the .top and .bottom divs with four rounded corner images:

    <!--[if IE]>
    <style type="text/css">
    .box { min-height: 80px; }
    .curves {
    display: block;
    height: 10px;
    .top {
    background: #ff0000 url('images/topr.png')
    no-repeat right top;
    .bottom {
    background: #ff0000 url('images/botr.png')
    no-repeat right top;
    <!--[if lt IE 7]>
    <style type="text/css">
    .box {
    height: 80px;

    The main drawbacks with this method are that the CSS won’t validate and it only works with fixed width boxes; however, a little recoding will fix this second problem.

    Users of Opera and other browsers will see a square box; as my client is corporate and has never received a hit from an Opera browser, I was willing to make this sacrifice. If they ever implement -opera-border-radius I could add this in.

    Of course there are other solutions for this, but I thought this way was quite simple.

  • 200610 Nov

    It’s all gone a bit quiet round here… sorry about that, but there’s not a lot of movement on CSS3, and I know the authors here are quite busy personally.

    Anyway, potentially big news is that the WHATWG are asking for developer feedback on HTML5.

    I’m still wading through the document, but of what I’ve read so far, of most interest to CSS fanatics will be a group of new, semantic tags; for example


    Of course these aren’t directly related to CSS3, but it should help save a lot of classes, ids and tag soup if implemented – and that’s the key. Even if it turns into a recommendation, how long until HTML5 is supported in IE – if at all. Many questions, many variables; but stay positive, and give your opinion.

  • 200616 Oct

    Last week we unveiled the CSS selector test over at css3.info. The test consists of over several hundred separate test cases, each designed to test a certain aspect of the compatibility of your browser with the CSS selector standards. Today we are going to expand the number of test cases to 578.

    The new tests have a large impact of the results. There are quite a few browsers that used to pass with the old test cases, but fail with the new test cases. Generally we added test cases for the following situations:

    • Dynamic updates: If you update the DOM using Javascript this should also affect the CSS selectors. Take for example the :empty selector. It only matches elements without any children. If you dynamically add a new child it should not longer match that particular element.
    • White space in attribute selectors: There are four different ways an attribute selector could be written: [attribute=value], [attribute =value], [attribute= value] and [attribute = value]. We now test if your browser supports all of these variants.
    • Case sensitivity of the value in attribute selectors: The previous version of the test contained a test case for determining if the value of an attribute selector was compared in a case-insensitive way. However this was not complete. We only tested the align attribute – which should be treated in a case-insensitive way. Only Konqueror failed this test. But there are also a lot of other attributes which should be tested in a case-sensitive way. Now almost every browser fails this test. More information about this case-sensitivity can be found on rakaz.nl: CSS selector bugs: Case sensitivity

    We also changed the way results are reported. Instead of just showing whether a selector failed or passed, we now detect if the selector is fully supported, buggy, or not supported at all. This should give all of us a better idea about the state of the compatibility of each browser.

    Browser Version Supported Buggy Unsupported No. tests passed
    Internet Explorer 6 10 1 32 276
    Internet Explorer 7 RC 1 13 4 26 330
    Opera 8.5.4 18 3 22 317
    Safari 2.0.4 21 7 15 336
    Firefox 1.0.8 24 9 10 352
    Opera 9.0.2 25 3 15 346
    Safari r16925 25 9 9 355
    Firefox 26 10 7 357
    Konqueror 3.5.4 37 6 0 570
  • 200611 Oct

    KHTML, the Konqueror rendering engine, has received another upgrade – and with it, more implementation of the still-officially unannounced CSS3!

    According to the release notes, KHTML 3.5.5 now has support for the HSV/HSVA color values. I must confess to being a little baffled by this; the CSS3 color module names HSL/HSLA values (of which HSL is supported), but not HSV/HSVA. Perhaps someone more au fait with colours could help me out on this one.

    Also now supported, apparently, is the outline-offset property – which does exactly what it says; offsets the outline around a page element.

    Please bear in mind I haven’t tested these, yet.

Hosting by: