@AndrewLeach This is how using a CSS framework like Less can blow your scalability aspirations out of the water.
They are squeezing the balloon in one place and it just keeps expanding in other areas.
Thank you for the clear and insightful response. Also appreciate the contrast and suggestions. Actually, I did have a little issue initially in understanding what the critics mistook however, was quickly able to figure that one out, as mentioned in the analogy in the post. What actually got me confounded was the connotation of the word 'artlessness'. Even the dictionary read: "The quality of innocent naiveté". Synonyms: ingenuousness, innocent. — Vaibhav2 hours ago
Yeah, you can't even give me a single solitary up vote? You're dead to me.
Screw your thanks, I'm in this for the six-digit milestone.
41 more up votes and I can lay my burden down.
And you'd think I could get an Enlightened badge out of this answer. It's so good Kris didn't even down vote it. That's showing remarkable forbearance for him.
I'm sure all the CSS tweaks are annoying. I apologize for the inconvenience. The main benefit to all of this will be our ability to quickly launch new features (like the user profile on Meta Stack Exchange) across the network without having to got through all the sites to adjust the CSS individually.
@tchrist We will be doing this to all sites, but not all at the same time. Most of our sites styles aren't as custom as English, which is part of the reason it hasn't gone as smoothly as we'd hoped. We have style rules that work for 95% of the sites, but not all of them. The new profile and other features will begin to roll out after we've finished the site conversions.
Aw come on, am I going nuts or do you keep going back and forth between vertical and horizontal presentation of votes/answers/views? Also, did the >100k view red color go away and is always orange now?
I agree. I wish font rendering was handled in a standard way across browsers. Webkit browsers have some nice antialiasing options, but they're not available on most PC browsers. Then throw in retina displays and you have even more inconsistencies.
Yes exactly. I absolutely cannot stand to read anything under Windows. I’ve been running a Windows VM inside my Mac at work to test some POS driver issues, and I find the font handling unreadable there.
I haven’t figured out how best to size Stack Exchange sites for optimal legibility.
I normally like most of my screen taken up by the active window, unless I’m hacking and have a bunch of parallel code windows open. But I find that on these big displays, the Stack Exchange font sizes don’t work so well.
Sometimes a zoom-level or two will work ok, but often not.
@KurtisBeavers could these CSS tweaks explain why Super User fonts are much smaller for me? This started happening a while ago and only seems to affect Super User.
@tchrist I agree. Hopefully once we're through with the conversion, we'll be able to test things like font sizes and line lengths, and if we find a more optimal solution administer it across the network easily.
Here’s a very dumb question I know I could find out myself easily enough, but I figure you must know the answer right off the top of your head. When one specifies Georgia as the font, does everyone using it have exactly the same font spec for Georgia, or are there different implementations of it on different systems? Yes, I could just check filesizes or checksums.
(Edited for typos and to hand-tweak the typography. See new appendix at bottom.)
Are there any rules or standards telling me when I should use a ligature and when I shouldn’t?
Yes.
In version 3.2 of Robert Bringhurst’s The Elements of Typographic Style (Hartley & Marks, 2008), the main dis...
@tchrist Because Georgia is a system font on Mac and PC, the browser uses your version of Georgia. There is a webfont version of Georgia called Georgia Pro that can be served from the web server. This fixes a lot of these spacing issues, but there's a tradeoff in page load time. Plus it can be expensive to license custom fonts for all of the network sites given the amount of traffic.
No One Rule Fits All Situations
This is all somewhat complicated, because it ties in with kerning support and font selection, and there is no one-size-fits-all answer that will serve for all situations. In my experience, ligatures are more apt to be needed in a tightly set serif roman or italic...
I lost about six weeks of sleep on my last book because my publisher was trying to use Arno Pro but the people doing their rendering/printing from DocBook had too many bugs in their software to handle its huge kerning rules, so it looked like crap.
I should have thought to have them use Minion Pro, whose kerning tables are done differently.
Web fonts are getting so much better than they were even a few years ago. Screen resolution is going to change things too. Hopefully the gap between print and web fonts will close soon.
@KurtisBeavers It really was. They weren’t matching x-height, and the x-height of Arno Pro is much much smaller than normal fonts usually have. So we blew the point size up from 10.2 to 11.6 or something so that it wouldn’t be too tiny to read, and then when we backed off from Arno Pro, we forgot to fix the point size back to 10.2 so the book is like a Large Print edition. :(
One cannot get an "ebook" version of Bringhurst’s Elements of Typographic Style. I do not know this, but I imagine this is because Bringhurst will not permit his extremely careful work to be destroyed as would happen if it were rendered in anything less than the exact PDF he produces with InDesign. Too many very subtle things he’s trying to show that simply aren’t supported with most ebook formats.
I have eaten of the fruit of the knowledge of good and evil keming, so every single time I read T om, T est, T ye, or NA V AR on websites or slide decks, I cringe. I wish I could unsee those.
Wow, what a crappy rendering of U+202F in fixed fontage here!
Let me try without fixed.
Better.
It is so sad that we have to write about OpenT ype fonts. :)
I told my co-authors that if our publisher did a good job on the typesetting, that I would send our publisher a copy of Bringhurst as a Christmas present — but that if they did a poor job, I would send them two copies. :)
I sent six! :)
They worked very hard at it. It is not their fault.
But the XSLT transforms just didn’t support good enough stuff.
I see, or think I see, that the #chat-body here is "Verdana, Arial, sans-serif".
Hm, let’s see what fixed is.
Boy is that a long font set!
Consolas,Menlo,Monaco,'Lucida Console','Liberation Mono','DejaVu Sans Mono','Bitstream Vera Sans Mono','Courier New',monospace,serif
@Cerberus That still doesn’t matter. If I have 7 combining characters stacked up on a single grapheme base and I press ^H, all 8 go away.
NFC: That these learnèd marks one character be is but a naïve façade.
NFD: That these learnèd marks one character be is but a naïve façade.
NFC: That these learn\N{LATIN SMALL LETTER E WITH GRAVE}d marks one character be is but a na\N{LATIN SMALL LETTER I WITH DIAERESIS}ve fa\N{LATIN SMALL LETTER C WITH CEDILLA}ade.
NFD: That these learne\N{COMBINING GRAVE ACCENT}d marks one character be is but a nai\N{COMBINING DIAERESIS}ve fac\N{COMBINING CEDILLA}ade.
What’s happening is that anything that comes in gets run through NFC by your browser to better find a precomposed code point in Verdana so crap doesn’t fail so badly.
But sometimes it just fails, like with hā̃̈t and hā̃̈t.
If you type the same thing into chat twice in a row using different code point but whose normalized forms are the same, it suppresses the second one as a duplicate.
So I am betting that Stack Exchange is normalizing. Just as well, really.
I had thought it must be happening elsewhere, but I guess not.
@tchrist I have never pressed ^H, but, when I press backspace, the combining characters are removed one by one, separate from the base character, aren't they?
@KurtisBeavers I thought entry in chat (and thus also editing) was handled by Stack Exchange ajax calls, not by the native platform. I must be wrong if we are getting different behaviors between Apple and Microsoft.