- From: Henri Sivonen <hsivonen@iki.fi>
- Date: Wed, 21 Mar 2007 11:37:05 +0200
On Mar 3, 2007, at 21:58, Ian Hickson wrote: > The question isn't whether or not you should have the ability to scale > images; it's clear that this is desirable. The question is whether it > makes sense to put this in HTML as opposed to CSS. Why would HTML > be the > place to put this? Because the dimensions vary from image to image, putting the dimensions in an external style sheet would mean moving the dimensions even further away from the images they pertain to. Generic reusable styles make sense in an external sheet. ID selectors specific to particular image files don't. OTOH, moving the dimensions from attributes to style='' or <style> within the HTML file is totally pointless from any *practical* point of view and would make it harder to implement structural HTML editors that don't tamper with styles. As for requiring pixel dimensions to be "correct": No, it shouldn't be required, because for backwards compat the <img> width and height are in CSS pixels and image dimensions are in real pixels. With emerging high-res displays it may soon make sense to have bitmaps whose bitmap pixels are smaller than CSS pixels. As for allowing percentages: Yes, they should be allowed. Percentages have to be implemented for backwards compat anyway, so pretending that they aren't there doesn't have a practical advantage when (considering the above) the attributes themselves should stay. > If we put this in HTML, how can we still drop <font>, <table > border>, <td width>, etc? Those aren't generally tightly coupled with something like the shape of a particular image file. (OTOH, the left/right alignment of table cells *is* often tightly coupled with the cell data, which suggests that the cell alignment attributes should not be dropped.) -- Henri Sivonen hsivonen at iki.fi http://hsivonen.iki.fi/
Received on Wednesday, 21 March 2007 02:37:05 UTC