- From: Bailey, Bruce <Bruce_Bailey@ed.gov>
- Date: Thu, 21 Dec 2000 13:31:04 -0500
- To: "'w3c-wai-gl@w3.org'" <w3c-wai-gl@w3.org>
- Cc: "'Kynn Bartlett'" <kynn-edapta@idyllmtn.com>, "'Leonard R. Kasday'" <kasday@acm.org>
Dear Group, Like a few others in this group, as much as anything I use this list as an extension of my brain -- sort of a public thinking out loud kind of technique. My apologies to anyone who is offended by this, but I will probably keep it up so long as I don't get flamed too badly. Let me run this by you... Has anybody experimented much using relative units for graphics? Like most on this list, I assumed that Checkpoint 3.4 applied only to text-oriented elements -- and that specifying exact IMG height and width size in pixels was of course not a problem. From my read of the HTML 4.01 specs, specifying height and width in em units (for example) is NOT allowed and that XHTML 1.0 doesn't change this. Am I correct on this point? I experimented with a navigation button bar (individual graphics, not an image map). The only relative unit I got to work at all was when I specified something like width="10%" (with the height attribute omitted). This, of course, was not of much utility in stretching the graphics (since the window wouldn't get any bigger than the full size of my screen). I could use this to shrink the images, but how helpful is that! I could not get an acceptable rendering with the 4x browsers using height="5%" (with or without a width attribute) nor using height="1*" width="4*" nor using height="2em width="12em". My testing was hardly comprehensive, but I think I hit a dead end. The point of all this was, of course, if any of these techniques worked -- then SVG would have distinctive advantages over bitmaps in terms of how good the image looked when scaled -- and dynamic scaling by changing the default font size would be possible. Len (and anyone else interested), I was hoping you might experiment some with these ideas and, if you have any success, add the results to your most excellent resources at URL: <http://astro.temple.edu/~kasday/wai/> (Isn't that nice of me to try and give an assignment to someone who doesn't work for me? Especially around the holidays! Smile!) Kynn (and anyone else who can help me with my understanding), I am reluctant to mention it, but I do have a bone to throw you with regard to your apparent contempt for SVG. I understood that one of the potential promises for SVG was its accessible to screen reader and text-only browsers. I don't see that in practice. Textual information is, of course, embedded in the SVG file. This is, of course, in theory much better than relying on the good will of the author to provide a separate and stand-alone textual equivalent. As it stands now, Lynx users just get <q>[EMBED]</q>. I could not get JAWS to do anything with an SVG file from within IE 5. I did have one partial success. I renamed an SVG file (from Adobe) from clock.svg to clock.htm. Opening the local file in IE revealed text! Opening the local file in Lynx revealed text! This was quite neat I thought. Can any SVG gurus on this list speak to how this aspect is suppose to work in theory and practice? Cheers, Bruce
Received on Thursday, 21 December 2000 13:31:26 UTC