The pixel grid vs. the font designer

by Michael S. Kaplan, published on 2007/06/15 03:01 -04:00, original URI:

Just a few hours ago I said a few words about Safari on Windows in the post entitled Let's go Uniscribe / ClearType has a real good vibe / So please use them in Safari, Steve [Jobs]!.

Now admittedly I did not talk about some of the bigger issue, and essentially thereby posted a bit of flame bait. :-)

Though the points I raised, I believe -- and in my opinion it would be a better result.

But the bigger issue was already covered by the likes of Joel Spolsky in Font smoothing, anti-aliasing, and sub-pixel rendering and Jeff Atwood in Font Rendering: Respecting The Pixel Grid. Which in the end says a lot about why Uniscribe is unlikely be used here and why Safari is unlikely to be altered to do "better" here for the subpixel piece of the font story.

Kind of the side effect of having no taste, I suppose....

But I have to wonder what the goal of Safari is, if so strong of a philosophical stance would be taken that is foreign to the users who are "native" to the platform. Is it to bring more Apple people to the PC once apple has injected more "cool" into it? Or to try to convert the heathen PC folk to the church if superior design? Either way it seems a bit of a miscalculation to me....

Now, not being a type designer or a typographer, a reader here might find it pretty easy to dismiss my words, and I think that's fine. But it may be worthwhile to spend some time trying to read the text in both browsers (after they improve their security story, I mean) and decide which side of the "Pixel grid vs. Font designer" Jeff pointed out deserves to sway your opinions.

Maybe we will all learn something about taste. :-)


This post brought to you by(U+17c5, KHMER VOWEL SIGN AU)

Random Reader on 15 Jun 2007 3:43 AM:

The Safari beta was updated to fix security issues:

As for why do this, a point I've seen several people make is that Safari is effectively the SDK for the iPhone.  I imagine it's important to them to allow Windows-using developers to design and test apps for the iPhone using an environment that renders just like it will.

Of course, I've also found it futile to try to assign one single reason for Apple doing anything :)

Michael S. Kaplan on 15 Jun 2007 3:51 AM:

Of course, as was reported, some of the vulnerabilities were also in the shipping version of Safari, not just the beta one on Windows. Kind of sad, though the sadder part is that they will get away with it. :-(

Zhasper on 15 Jun 2007 5:18 AM:

What Random Reader said :)

Doug on 15 Jun 2007 6:10 AM:

One side-effect of ClearType/TrueType is that since it snaps to the pixel grid, line lengths are dependent on the output device.  This makes it difficult to make page breaks on the fly and have the print preview match the printed output.

Mihai on 15 Jun 2007 1:18 PM:

"difficult to ... have the print preview match the printed output."

100% agree.

But let's not forget we are talking about a browser here, not some kind of advanced DTP application.

The main-main role of a browser is to put on screen crappy designed HTML.

HTML is not accurate, does not look the same in 4 different browsers on the same OS. It is not realistic to expect it to look the same on screen and on paper. It was not designed for that, it is not good at it, and that is ok with me.

I want 100% accuracy, then I use a PDF with embedded everything.

We have to select the right tool for the job.

Ben Cooke on 15 Jun 2007 2:07 PM:

Whether or not Apple's algorithm is better, I'm glad that Safari on Windows is the same as Safari on Mac. My primary use for Safari on Windows is to test how sites I'm developing will look on a Mac, so the more similar they are the better.

Michael S. Kaplan on 15 Jun 2007 2:12 PM:

Hey Ben -- I hope you recognize that as an argument it is not very sustainable when compared to the best result for customers overall, including yours. Right? :-)

Mikkin on 15 Jun 2007 8:33 PM:

+1 what Mihai said, only more so.

Remember when purists felt HTML tags should not be used to control rendering at all? "Only semantic HTML tags are supported." That debate was settled long ago, and this situation is a logical consequence. But I know what fonts look good to me on my display device, the author does not.

Rosyna on 17 Jun 2007 4:56 AM:

"Of course, as was reported, some of the vulnerabilities were also in the shipping version of Safari, not just the beta one on Windows."

Uhm, says who? Someone that has proof and doesn't just want their blog post repeated on different sites?

Michael S. Kaplan on 17 Jun 2007 10:26 AM:

I was referring to this site.

Rosyna on 17 Jun 2007 12:21 PM:

Yeah, don't refer to that site. He has a history of making up vulnerabilities to get attention. There's no evidence he found something and his one screenshot has a significant typo in it.

Rosyna on 17 Jun 2007 12:22 PM:

Bah, I meant to say there is also more fun in these comments

Michael S. Kaplan on 17 Jun 2007 12:57 PM:

I personally don't care much for the people who release exploit code, as it really serves no one other than the egos of the ones who post it and the criminals who use it.

But I recognize that across the industry opinions vary on this subject...

Dean Harding on 17 Jun 2007 7:54 PM:

I don't get that Joel article. Is he arguing that Apple's rendering is actually better? He seems to think that someone like me will only prefer the ClearType solution because that's what I'm used to...

But in what situation is blurry text *ever* better? When you're comparing printed text to on-screen text? How often do you do that?

Please consider a donation to keep this archive running, maintained and free of advertising.
Donate €20 or more to receive an offline copy of the whole archive including all images.

referenced by

2011/01/04 Short-sighted text processing #4: Squeezing every bit of text you possibly can out of MacOffice 2011

2007/06/15 +1 on what he said (and some girl's ones are bigger than other girl's ones)

go to newer or older post, or back to index or month or day