"My appeal is to software and hardware developers to ensure that a 12 point font will be rendered the same size everywhere, regardless of screen size and density."
This is a not very well reasoned appeal. If it were implemented, every Keynote presentation would be unreadably small when projected onto the big screen, literally the same size as the presenter holding up a piece of paper for the audience to read with their binoculars. Likewise, no more than a few characters at a time would be visible on a phone.
There is simply no serious counteragument; nobody can say making things unviewable would be a step forward.
The best you can do is complain that the name of the units is misleading. Fine. So what about introducing a new unit that is truly faithful to a real world measuring stick? You can argue for that all you want, but guess what? Nobody in their right mind would actually use it, outside of exotic scenarios, because generally you want to avoid the above effects. You want your user interface to be visible to users at different screen sizes meant to be viewed from different distances.
Thus, even if we were to introduce a new "TruePoint" or "TrueInch" unit, it would be largely.. and I beg you to excuse the pun here.. pointless.
There are several slight distinctions to be made there. A keynote/powerpoint presentation is typically displayed as though on a piece of paper (A4, say), and thus of course type should be rendered proportionate to the virtual piece of paper, no matter the display. The post is referring mostly to HTML sizes.
The second distinction is that a projector or video monitor should duplicate the master screen, and indeed for a projector you have no idea how big the final screen is, and can change its size by moving the projector around. Video projections, similar to paper, are a proportionate medium.
Not making these distinctions is definitely an oversight of the linked article. His major point is well illustrated in his example picture, where "5 inches" has been displayed with varying degrees of accuracy, but it's clear all the monitors are trying to display something _close_ to 5 inches. Not getting it exact (±½ pixel) is a failure.
I agree about the physical rendering size always being the same for a point, a point should be more general, but I still think there is some utility for having things defined in real-world units showing up as real-world units when measured on glass in the same z. An inch is an inch, and if it's too big for the display it should be clipped and if it's 100 feet away it should be unreadable. It would be useful for classes of devices--instead of special casing for resolutions, ratios, and densities, special case for desktops/laptops, phones, tablets, home projectors, conference projectors, etc.
But what's really missing is the explicit presence of a projection surface and the ability to adjust the factors that go into the calculation when projecting. It would be nice to have a unit that lets you say "I want this item to be as tall as the subjective inch if a person holds a ruler one foot away from them." That way no matter how close or far away I'm looking at the item (assuming the display is large enough--too small and too far will get it clipped, too few pixels will get it blocky), if I hold a ruler one foot away from me the item will be "as tall" as the ruler's inch mark. (Edit: I'm having fun trying to visualize a powerful enough 64x2 display built on a mountain. Where I grew up, there is just a single letter. From afar: http://upload.wikimedia.org/wikipedia/en/d/d9/Little_mountai... And from on top of it: http://2.bp.blogspot.com/-QeeaDcCiD4o/TexykEujKiI/AAAAAAAAHR...)
> This is a not very well reasoned appeal. If it were implemented, every Keynote presentation would be unreadably small when projected onto the big screen
And that's why CSS3 introduces viewport-relative units[0]. No reason why Keynote couldn't have and provide that, instead of removing all meaning from existing ones.
They could also define an arcsecond unit, and each device would carry its typical (configurable and overridable) distance of use to internally convert to pt, and then DPI to rasterize to px.
Comments
"My appeal is to software and hardware developers to ensure that a 12 point font will be rendered the same size everywhere, regardless of screen size and density."
This is a not very well reasoned appeal. If it were implemented, every Keynote presentation would be unreadably small when projected onto the big screen, literally the same size as the presenter holding up a piece of paper for the audience to read with their binoculars. Likewise, no more than a few characters at a time would be visible on a phone.
There is simply no serious counteragument; nobody can say making things unviewable would be a step forward.
The best you can do is complain that the name of the units is misleading. Fine. So what about introducing a new unit that is truly faithful to a real world measuring stick? You can argue for that all you want, but guess what? Nobody in their right mind would actually use it, outside of exotic scenarios, because generally you want to avoid the above effects. You want your user interface to be visible to users at different screen sizes meant to be viewed from different distances.
Thus, even if we were to introduce a new "TruePoint" or "TrueInch" unit, it would be largely.. and I beg you to excuse the pun here.. pointless.
There are several slight distinctions to be made there. A keynote/powerpoint presentation is typically displayed as though on a piece of paper (A4, say), and thus of course type should be rendered proportionate to the virtual piece of paper, no matter the display. The post is referring mostly to HTML sizes.
The second distinction is that a projector or video monitor should duplicate the master screen, and indeed for a projector you have no idea how big the final screen is, and can change its size by moving the projector around. Video projections, similar to paper, are a proportionate medium.
Not making these distinctions is definitely an oversight of the linked article. His major point is well illustrated in his example picture, where "5 inches" has been displayed with varying degrees of accuracy, but it's clear all the monitors are trying to display something _close_ to 5 inches. Not getting it exact (±½ pixel) is a failure.
I agree about the physical rendering size always being the same for a point, a point should be more general, but I still think there is some utility for having things defined in real-world units showing up as real-world units when measured on glass in the same z. An inch is an inch, and if it's too big for the display it should be clipped and if it's 100 feet away it should be unreadable. It would be useful for classes of devices--instead of special casing for resolutions, ratios, and densities, special case for desktops/laptops, phones, tablets, home projectors, conference projectors, etc.
But what's really missing is the explicit presence of a projection surface and the ability to adjust the factors that go into the calculation when projecting. It would be nice to have a unit that lets you say "I want this item to be as tall as the subjective inch if a person holds a ruler one foot away from them." That way no matter how close or far away I'm looking at the item (assuming the display is large enough--too small and too far will get it clipped, too few pixels will get it blocky), if I hold a ruler one foot away from me the item will be "as tall" as the ruler's inch mark. (Edit: I'm having fun trying to visualize a powerful enough 64x2 display built on a mountain. Where I grew up, there is just a single letter. From afar: http://upload.wikimedia.org/wikipedia/en/d/d9/Little_mountai... And from on top of it: http://2.bp.blogspot.com/-QeeaDcCiD4o/TexykEujKiI/AAAAAAAAHR...)
> "I want this item to be as tall as the subjective inch if a person holds a ruler one foot away from them."
Sounds like you're looking for arcseconds.
> This is a not very well reasoned appeal. If it were implemented, every Keynote presentation would be unreadably small when projected onto the big screen
And that's why CSS3 introduces viewport-relative units[0]. No reason why Keynote couldn't have and provide that, instead of removing all meaning from existing ones.
[0] http://www.w3.org/TR/css3-values/#viewport-relative-lengths
They could also define an arcsecond unit, and each device would carry its typical (configurable and overridable) distance of use to internally convert to pt, and then DPI to rasterize to px.
That's a rather interesting idea. You should suggest it to the CSS3 dimensions WG to see what they think of it.