On Wed, 26 Sep 2012 10:37:10 -0700 "H. S. Teoh" <hst...@quickfur.ath.cx> wrote: > > I have one objection to your list though: although _for the most part_ > AA's can work with any kind of key, there are a lot of bugs in that > area. The language itself, of course, in theory supports any kind of > key, but the current implementation is honestly a mess. I've tried > fixing things but one thing leads to another and nothing short of a > total overhaul will completely address all of the problems. >
There is that, however, it beats the hell out of what I've seen in other languages, like Haxe, where the *value* is generic or templated, but the key is a string. Period. (And then I think Haxe also has a IntHash type now, too, but that's...a signal of not quite *getting* generic code.) > Can you believe that prior to C++11, true AA's weren't > even a part of the standard library? Yes, I can believe that very easily. :/ > > Well, there's a GUI front-end for it (LyX), I don't know if that > handles things like native UTF support. I was thinking more of a 21st > century rewrite of LaTeX that has modern support like native UTF, > revamped syntax to replace anachronisms, etc., but adhering to the > original design principles. Sorta like what Knuth & Lamport would've > come up with, if they had developed TeX/LaTeX in 2012. > Right. I mean like what CoffeeScript does for JavaScript. But then I don't know if you'd would be able to solve all of latex's issues that way, mostly just syntactic ones. > Sites like Wikipedia use LaTeX to generate math formula images by > passing embedded <math> tags through LaTeX for formatting. :) > Seriously, that's what makes math even remotely tolerable to write in > Wikipedia. It's imperfect, though, 'cos the baseline of the formatted > text in the image often doesn't line up with the baseline of the > surrounding HTML text. And font sizes don't always match up. But it's > better than the horror of attempting to write math in HTML. > Interesting. > > > > Funny thing is, it works fine (ie without using pop-ins) when JS is > > off. But I can't turn JS off in iPhone Safari. > > Argh... iPhone/iPod Safari is one of the worst horrors there are. The > UI is simplistic to the point of daimbramage, which makes it unusable > for anything but the most trivial of tasks. Nothing is configurable, > no privacy settings, can't control Javascript, the maximum number of > tabs is ridiculously small, You can't tell it to override all "target:blank" (I think that's what it's called, I never make them, so I don't remember) and always open in the same tab unless *I* say otherwise. That's one of my biggest annoyances with it so far. > scrolling a long page is really horrible, Yea. Needs directional buttons. Swipe is overrated and only suitable for minor infrequent uses. > wide images get clipped with no way to unclip them when using the > mobile stylesheet (probably the same bug you describe above), etc.. > And Apple has the audacity of forcefully banning all other browsers > from the app store, for the simple reason that they are superior > browsers, and oh no, we simply can't allow customers to have a > superior experience! I thought Chrome was available for iOS? But if what you say is true, then that's interesting to compare to "evil M$": Microsoft: Installs their browser by default. Allows any other browser to be installed and set as default. People are pissed. Gates is demonized. DOJ sues. Apple: Installs their browser by default. Bans other browsers entirely. Everybody's happy and praises Jobs as a great designer and savvy businessman. No lawsuit. > > About the only commendable thing with iPod Safari is the lack of Flash > (good riddance!). > Yea, I was always ambivalent about that. On one had, I felt it was a bone-headed decision and that it should be left up to the user. OTOH, I can get behind almost anything that helps bring an end to Flash. So I've always been torn ;) > > > > The problem with that is you're creating excess vertical scrolling. > > Just to read linearly it's "scroll down, scroll up, scroll down", > > etc. (Of course, that pain is hugely compounded when the > > multi-columns are on page-based PDFs, like academic research > > papers.) > > That's why I said that multicolumn support needs to be natively > supported in the browser, NOT hardcoded into the page itself. It > should be the browser that decides whether something should be > multicolumn, and how tall the columns should be. There's no way the > author can possibly account for every possible browser configuration > out there to make this kind of decisions. > I see. I'm not sure how even the browser would really make it work though, unless maybe you make the whole page scroll horizontally with as many columns as it takes? > > > The root problem there is that the need for multi-column on the web > > is artificially created by manufacturers and consumers who have > > collectively decided that watching movies is by far the #1 most > > important thing for anyone to ever be doing on a computer. Hence, > > "decapitated fat midget" 16:9 screens for everyone! No matter how > > bad it is for...just about everything *but* movies and certain > > games. Which, I suspect, is also the main reason we can't have > > browsers anymore with nice traditional UIs - because they have to be > > shoe-horned into a movie-oriented half-screen. > [...] > > I avoid those height-truncated monitors like the plague. I only ever > buy monitors with 4:3 aspect ratio. Seriously, if all I wanted to do > was to watch movies, I wouldn't be using a PC in the first place. > I would do the same thing. In fact I had sworn I would never get anything wider than 5:4 (and even then I prefer 4:3). Unfortunately, when I was shopping for a laptop, there was *nothing* but 16:9. Not one single model, in or out of my price range. So it was 16:9 or no portability :( At least this has VGA output though (and HDMI, but anything that takes HDMI is going to be 16:9). > But still. Sometimes you have a long list of narrow items, and > multi-column makes it more readable without excessive scrolling. > > Maybe I should start a new trend: side-scrolling webpages with > *multi*-columns. :) (Though this probably only makes sense with > vertical writing systems, like the vertical variant of Chinese > writing. Which is in vertical columns *and* read right-to-left. > Bwahahahaha...) > Heh :) Traditional Japanese is like that, too. (Not surprising since their writing system is derived from Chinese.) Weird thing is, after studying that, and reading a lot of manga, anytime I see vertical English text, I keep trying to read it right-to-left out of habit :)