function addStyle(string){
if(/microsoft/i.test(navigator.appName) && !/opera/i.test(navigator.userAgent)){
document.createStyleSheet().cssText=string;
}
else {
var ele=document.createElement(‘link’);
ele.rel=’stylesheet’;
ele.type=’text/css’;
ele.href=’data:text/css;charset=utf-8,’+escape(string);
document.getElementsByTagName(‘head’)[0].appendChild(ele);
}
}

a test link

On another note, I’m not sure if I’ve posted on this software in general. I think it’s a brilliant project both in its usefulness and its level of sophistication. It’s truely kind of you to volunteer it to the community. Thank you.

BTW, pt 2 of my previous comment is "You called setAttribute from a CSSStyleSheet, not a DOM node."

After inspecting the script in comment 14, I found out that the whole script went wrong! (damn)

1. You didn’t append the
   
2. tAttribute from a CSSStyleSheet, not a DOM node.
3. Inserting the rule this way just get Firefox ignoring it (damn again to Firefox this time)

Finally I made the corrections here:
function fontSize(value) {
 if (!(s5ss = document.getElementById('s5ss'))) {
  if (!isIE) {
   document.getElementsByTagName('head')[0].appendChild(s5ss = document.createElement("style"));
   s5ss.setAttribute('media','screen, projection');
   s5ss.setAttribute('id','s5ss');
  } else {
   document.createStyleSheet();
   document.s5ss = document.styleSheets[document.styleSheets.length - 1];
  }
 }

 if (!isIE) {
  while (s5ss.lastChild) s5ss.removeChild(s5ss.lastChild);
  s5ss.appendChild(document.createTextNode('body {font-size: ' + value + ' !important;}'));
 } else {
  document.s5ss.addRule('body','font-size: ' + value + ' !important;');
 }
}

It’s not a huge problem, for if you were using powerpoint you wouldn’t be able to select text, but it might be a nice feature to be able to select and copy text if someone wants to. Just an idea

function fontSize(value) {
if (!document.getElementById('s5ss')) {
if (!isIE) {
document.createElement('style');
s5ss = document.styleSheets[document.styleSheets.length - 1];
} else {
document.createStyleSheet();
document.s5ss = document.styleSheets[document.styleSheets.length - 1];
}
s5ss.setAttribute('media','screen, projection');
s5ss.setAttribute('id','s5ss');
}
if (!isIE) {
s5ss.insertRule('body {font-size: ' + value + ' !important;}',0);
} else {
document.s5ss.addRule('body','font-size: ' + value + ' !important;');
}
}

That gets me the following error in Firefox: “Error: s5ss.setAttribute is not a function”. Which seems weird, because I use it elsewhere in the JS without incident. Anyone know what’s going on and, more importantly, how to fix it?

<meta name="author" content="Eric Meyer, Complex Spiral Consulting" />
<meta name="author" content="Joe Public, ConHugeCo Corp." />

Does that imply the second value will overwrite the first? Are they concatenated instead, and if so, how? Granted, the point of the meta elements is to be human-readable but I’d like to avoid any machine-read problems.

Print size changing– hmmm, no, that wasn’t intentional. That’s happening due to the dynamically added style sheet that sets the font size, since if you toggle to outline view and print, the font size is normal. We need to restrict the created style sheet to media="screen, projection".

Looks like it can be anything.

The interface at this point is that any keyboard or mouse-click advance will trigger incremental display, except clicking on the “next slide” link in the lower right-hand corner. Using that link will jump you to the next slide regardless of the state of incremental display on the current slide. What I’m trying to figure out is if reverse keyboard events (left arrow, page up, etc.) should run the incremental display back a step, or jump straight to the previous slide. At the moment, the JS does the latter, but I’m becoming convinced it should do the former. I’ll have to tackle that in the near future.

At the moment, the testbed’s styles are really only set up to handle incremental bullet points, and only when they’re text to boot. I’m still looking at Henryk’s approach to incremental graphics, and thinking about how I want things to work, and whether or not the two are at all different. They may not be. We’ll see.

Lachlan: so are you saying that commas aren’t a problem in named meta elements, that they don’t imply a concatentation of some kind? So I could have values like "Eric Meyer, Complex Spiral; Joe Public, ConHugeCo Corp." and not run into trouble? That would make me really, really happy.

The best way to do meta data in a standardised way is using the Dublin Core. I haven’t read it all in detail, so I don’t know how it supports multiple authors

Finally, Ugly Virgin (if that is your real name :-), the correct way to encode an ampersand to display, as you tried in comments 1 and 2, is to write it as &.

I know, that was sort of a ‘design decision’ (others would probably call it an ugly hack) because I didn’t want to do bigger modifications to the js. And also because I have no idea how users really would want it to behave. I did it this way because we were presenting with a cordless mouse.

If you give me your opinions about what the interface should be like I would take up the task of modifying it to behave this way. (And clean the code up for inclusion in the latest version.)

–
Henryk Pl