More IE7 Ranting

Posted by

Seems I can’t get enough of the IE7 ranting today!

This time, it’s to do with their new “Page Zoom” feature. Now, this is an awesome feature, and one that I’ve been hoping they’d implement in this release. One of the very cool things about this feature is that when you zoom in on an image, it uses a bicubic filtering to make the scaled image “smoother.”

Now, some people are complaining that they want to be able to switch to “nearest neighbour” scaling for development purposes, and I think that would be a useful feature as well. But that’s not what this rant is about.

You see, I run my laptop at 1920x1200 resolution (which is the screen’s native resolution). To make text readable, I also set my DPI to 144 (which is 150% of normal).

The great thing about CSS is that it was designed with resolution-independence in mind, which is why you can specify units as a percent, in points (pt), in em-units (em) or various other units. These units are great because it means the layout adjusts itself if the system is using a higher dots-per-inch than the “normal” 96DPI.

The problem was that CSS also included the “px” or “pixel” unit, which is directly related to pixels on the screen, and does not scale when the font size is increased. The problem is that millions of web developers running on 96DPI systems would use “px” to layout their pages, thinking that as long as it all looks right on their machine, then it’d be OK for everyone else, to. It was even worse if they mixed pixel-based sizes with point-based sizes.

Fortunately, IE6 came up with a brilliant solution to this problem, the UseHR registry key, which basically fools the website into thinking I’m running at 96DPI and just uniformly scales the whole thing up – images and all. In fact, this is the reason I use IE at home instead of Firefox.

It was awesome. The only problem I had with it, was that it always used the “nearest neighbour” algorithm for scaling images, which meant they looked a little “blocky” – especially since they were scaled 150% on my system, it meant that some pixels were twice as big, some were their original size.

I had high hopes that IE7 would fix that problem (I’d even considered on occasion writing a Browser Helper Object to do it for me (I would probably have had to hook the call to BitBlt and do the scaling myself or something). But it seems it has not. Well, not quite.

The problem, in my opinion is that their new “Page Zoom” feature just makes it a complete joke! Take a look at the following two images. The first was taken at 100% zoom, with UseHR turned on and 144DPI:

100%

And here’s the same setup, but I used the Page Zoom feature to zoom to 99%:

99%

You tell me which one is better... Now, it seems pretty obvious what they’ve done. For performance reasons, they’re calling the regular BitBlt when the zoom is 100% (which is what it would normally be) and only doing the (more expensive) bi-cubic filtering when you zoom.

Let’s hope this is a simple fix that’ll make it into the final release. After all, they should just have to change their condition to:


if (page-is-zoomed || UseHR)
{
    // use bi-cubic filtering
}
else
{
    // use normal BitBlt
}

Fingers crossed!

More ClearType Ranting

Posted by

Just further to my last comments on this subject...

Actually, it’s not just one extra application’s settings I have to fiddle with now to disable ClearType – it’s two! It seems that Outlook Express has a separate setting to turn off its ClearType, under the registry key:

HKCU\Identities\{...}\Software\Microsoft
    \Outlook Express\5.0\Trident\Main\UseClearType

(Note: I haven’t tried it myself, so I can’t confirm whether this actually works or not)

So now there’s actually three different places I have to go to, in addition to the global setting in order to turn off ClearType. Two of those places (Word and Outlook Express 7) are ugly registry hacks!

What annoyed me even more is the response from Peter Gurevich when people complained about it:

Q4: Why is IE not using the system wide setting as default and why is it overwriting it?

A4: We believe strongly that this feature will enhance the user experience of a [sic] many users and believe that the best way to bring it to users is to turn it on by default independent of the system setting.

Basically, he’s saying that since the system-wide setting is too hard to find, they figure it’s better to enable it by default on an application-by-application basis. That is wrong, wrong, wrong! He goes on to say that it’s easier to get users to turn off a feature they don’t like than it is to turn on one they may not know about. Now that may be true, but I hardly see how it’s Internet Explorer’s prerogative to decide which features of Windows are beneficial to me. Maybe if ClearType is such an amazing feature and that nobody knows about, it could be turned on in a service pack of Windows – get people to turn off a feature they don’t like once, rather than doing it for every new Microsoft application that is ever released from now on.

OK, I think I’ve vented enough for today :)

ClearType is NOT King

Posted by

In my previous post, “Disabling ClearType in Reading Layout View,” I went off on a little rant about how applications don’t respect user’s preferences. After all, ClearType is actually a global setting, available from your Display Settings control panel applet, simply click on “Effects…” to get this:

Effects Dialog Box

See? If I wanted ClearType turned on, I’d do it in that menu. I don’t want your application to assume that “ClearType is King™” and turn it on regardless. Now, I’ll admit that option may be a little buried for most users, but it’s there.

But now it seems that Word is not the only Microsoft application that turns on ClearType by default. The new Internet Explorer Beta 2 “preview” is doing the same thing. At least for IE7 the option to turn it off is in the GUI and not buried deep in the registry like with Word.

But now that’s two different applications that I’ll have to configure to turn it off if I want it off. It should just be a simple matter of turning on/off the global setting.

In an ideal world, Windows would detect if I connect an LCD monitor to it, and ask if I want to enable the global ClearType setting. That way, applications wouldn’t have to resort to these hacks to force it on by default for themselves.

New Blog Skin

Posted by

Well, as I mentioned the other day, IE7 was having problems displaying my blog with decent speed. So I thought it was a good excuse to change the skin from one of the supplied defaults to my own.

I made one which is black-on-white, rather than the old white-on-black (or at least white-on-dark blue) ‘cause I think that makes it look a little less amateur-ish.

Anyway, there’s still a few thing which need fixing up (web development is not my strong point, and it’s taken almost all day to write up this skin. I guess with the way the skins on b2evolution are written, it’s basically like writing a complete web-page)

P.S. If you’re an IE developer investigating my bug-report, you can still access the old (very slow) page by changing the skin here

IE7 Beta 2 Preview

Posted by

Apologies to anyone who just downloaded the new IE7 Beta 2 Preview (what’s a beta preview, anyway?) I’ve downloaded it as well, but the codeka.com site runs really slowly for some reason that I cannot fathom.

Anyway, I’ve let the IE folks know (via their feedback email ietell@microsoft.com) so we’ll see if there’s any update.