Re: Making pt a non-physical unit

On Thu, Jan 7, 2010 at 8:34 PM, Felix Miata <mrmazda@earthlink.net> wrote:
> Why not the same for extra large devices? If either or both should, why not
> all? Really, there's no good reason why every user shouldn't be able to scale
> her device's output to her comfort level. Yet, AFAIK, no OS X user yet can.

/Applications/Performance Tools/Quartz Debug.app

It's actually fun (I used 3.0 for a while to watch things break), but
it causes some interesting pain. One thing it does let me do is get
decent sizes for my icons in the dock (i can't otherwise).

> Windows users can only at significant usability expense. That leaves mainly
> only the more astute Linux users (via DPI fudging) or special device users
> with that highly desirable ability.

Given that the Linux apps I look at can't even get their UIs to fit
properly at default resolutions, I think we'll let Linux pass.

-------

FWIW, I think this entire thing is a total mess.

I work on a high DPI system, one of the very few that actually ships
Gecko. -- The n900 is 267 ppi (according to Wikipedia).

At some point Boris noted that he knows of very few browsers which do
true zoom w/o reflow. MicroB (on the n900) is one of these browsers.
By doing true zoom without reflow, we get a performance "win", and a
usability loss (try visiting google.com in the n900 and zooming, the
text area very quickly stretches offscreen on both sides).

It's a compromise, and often it's a really bad compromise. But we're
mostly ****d

And roughly, since I have no sympathy for standards authors (I work on
a standard I'm not going to be proud of, so now I resemble what I
resent).

CSS screwed up by allowing authors access to physical dimensions
<period>. There was plenty of evidence when CSS was being developed
(let's call this 1994-1996) that authors didn't get this stuff right.
We're talking about the time period when Windows 95 was introduced,
there was a way to tweak DPI and it didn't work, so it was rarely
used. MS also was forced to use the wrong number even when it had
better numbers.

To assume that web authors would do the right thing was incredibly
naive. The vast majority of web authors have and will always do the
wrong thing, that which is quick, and dirty, but works for them. And
they will not be reachable, they might die (and may the few who were
good rest in peace, the rest should just be reviled).

The gateway to the web is the web browser. When a web site doesn't
work (e.g. OWA's right click menu which only works in IE), the user
will blame the browser (bug number available if you care). The user is
not capable of complaining to the vendor (Microsoft, I'm looking at
you), and the vendor will not fix it (and yes, I've tried complaining
to Microsoft that OWA did not work in my mobile phone's browser -- it
almost worked, except I couldn't compose a message because of the way
they sized the page -- and, while I work on browsers, and know that I
should blame microsoft, and I tried to get them to do something, in
the end, as a user, I blame my cell phone which had a webkit based
browser for not *fixing* the *broken* page).

It's charitable to claim that 1/10 of 1% of all web designers will get
any of this stuff right. But that still leaves the rest of the world
with pages which break on any slightly different browser or device.

Standards need to be written as Jails, we need to stick the content
authors into untrusted jails, allowing them sufficient flexibility to
do stuff, but not as much flexibility as we've allowed to date,
because what they've been able to construct is sufficient that it
hurts my customers, my friends, my family, and the world at large.

Received on Friday, 8 January 2010 00:34:00 UTC