Antipode

It’s odd how something that seems like a perfect piece of software will start to aggravate you if it goes stale for long enough. I would be thrilled with a Textmate 1.6 that fixes a few bugs¹, since it’s still a capable editor. So, two years after 1.5 is released, I go hunting for news:

There is no ETA, and I won’t speak about timing before I am certain I can provide an (alpha/beta) release within the next month… So put TM 2.0 up there with Duke Nukem Forever and be positively surprised the day it is released :)

Uh oh. I smell Second System effect. Peter Haza’s reports from the TextMate IRC channel earlier this year confirm it, listing many new features and saying it’s a complete rewrite. Hopefully Allan Odgaard doesn’t feel like Textmate 2 needs to be featureful enough to justify the wait, or the wait could never end.

1. Multi-file search and replace bugs, especially. Argh. [↩]

Today Jane Wells at Automattic posted a report on their usability testing for the WordPress admin panel. Not only are they addressing most of my Wordpress UI gripes, they’re also showing great leadership in usability work for open source software.

SurfaceWare is a research project to create glasses that can detect when you want a refill. As impractical as Microsoft Surface is, I love that it’s letting a lot of people play with new ideas.

I keep up to date on Internet Explorer development for two reasons. Firstly, most users are on Internet Explorer. Secondly, browser innovation and competition is fascinating. Anyone who thought browsers were mature a few years ago is now being proven very wrong. And so I present some of the most interesting additions in IE8.¹

Seeing Microsoft compete on browser UI is interesting. They’re incorporating ideas from Google Chrome so fast, it’s like they came up with them independently.² Ha ha, just kidding. IE8 users will notice:

• Chrome’s New Tab Page
• Chrome and Safari’s Private Browsing Mode
• Opera and Firefox’s Automatic crash recovery
  
• Somewhere between Chrome’s and Firefox’s Address Bar with search suggestions.³
• Tab Grouping with automatic colour-coding - probably the most original and the weirdest thing they’ve added to IE8

Okay, so it looks like they’ve taken enough from other browsers to get users to upgrade. This is good news, since the developer-side improvement from IE7 to IE8 will be a bigger jump than we’ve seen this century. Microsoft has been busily (and sometimes controversially) announcing standards-oriented developer features:

• Adding Compatibility View, where developers or users can tell IE8 to emulate IE7
• Adding the new developer tools that I posted about recently
• Coming much closer to CSS 2.1 than IE7 was; in my testing maybe 90% closer
• Prefixing all non-standardized CSS properties with -ms; even CSS3 ones like word-wrap
• Working on some performance gains, albeit nothing compared to The Other Guys

At the end of the day, it’s a great time to be a web developer, since the lowest common denominator is rising. It’s also a great time to be a web user, since Safari, Firefox, and Chrome developers will have to push that much harder to stay ahead in the new browser wars.

1. If you want to follow IE8 development, check out IEBlog. It’s nice to see Microsoft opening up their development process on such a vital piece of software. [↩]
2. Update: To be clear, IE8 Beta 2 came out 6 days before Chrome Beta 1 was released. It really is possible they came up independently, but doubtful. My bet is that Microsoft knew what Google was working on (a lot of people internally at Google had seen Chrome), and of course Google had seen IE8 Beta 1. [↩]
3. I should give them some credit here. Their implementation has what I think is a real innovation, which is the ability to delete any potentially scandalous history items from within the suggestion bar. I love to see location bar innovation. [↩]

A lot of people are very excited about jQuery. I hear a lot of, “jQuery has concise code, jQuery has less iteration, jQuery has a rock star on their website. A rock star!” I’ve always been kind of skeptical, though, since a few of the things people love about jQuery are actually in Prototype now, such as the ability to directly manipulate and chain some uses of $(), as well as the $$() function for returning items based on arbitrary CSS selectors.

With all these people talking about how sweet JQuery is, though, I wondered if it was on track to replace Prototype as the #1 JS framework. Well, according to Google Trends, interest in JQuery is blowing away interest in Prototype. What’s more notable is that these numbers include other uses of the word “prototype”! Very impressive. Next time I have a choice of JS frameworks, I’ll give it a run for its money.

Translink and my friends at Handi Mobility have taken the wraps off of a new iPhone-optimized Translink site. Check it out if you have an iPhone (or other mobile WebKit browser) and use transit in Vancouver. It looks good and feels good! Now all I need is the Maps application on the iPhone to give me trip plans, and my life will be complete.¹

1. There is a mobile Google Transit website, but it doesn’t work in Vancouver. It “appears to be outside our current coverage area”, even though the desktop version works fine. Grr. [↩]

My favourite CSS3 module, Backgrounds and Borders, published a new working draft this month. The working group’s new issue tracker says there are only a few open issues left. Ostensibly that’s good, but like many I’m getting cynical about the CSS3 spec. The first Backgrounds and Borders working draft was released in 2001. As in, the year the Sega Dreamcast was discontinued. Sigh.

Gecko and Webkit have been implementing these CSS3 features as they’re able from the working drafts, helping to form the spec. Unfortunately, Microsoft has refused to play ball until the spec is finalized. Meaning, until those responsible¹ get this out the door, millions of man hours will still be spent making rounded corner images and semantically meaningless wrapper divs.

Yes, 2001 as in the year Napster was shut down. I’m not joking.

1. The editors are Bert Bos and Elika Etemad. If you know them, feel free to give them a poke. [↩]

Four years ago Dave Shea gave us CSS Sprites, and it was good. Since then they’ve seen wide adoption, both for reducing HTTP request counts¹ and for doing away with Javscript image pre-loading.

As great as CSS Sprites are, they’re a bit more more error-prone than using traditional images. Google uses them extensively, and I’ve noticed a trick they use to keep everything lined up right. Between the images in their sprites on Google Reader, they’ve put a red line of pixels. That way, if something is one pixel off, it will be caught immediately.

As clever as this is, it has a problem: we’ve entered an era where one pixel at design time doesn’t always mean one pixel at display time. My favourite example of this is OS-wide resolution independence, but it’s here today in Firefox 3.

To the right is an example of Google Reader scaled up a couple notches using full-page zoom. It’s technically difficult to scale pixel-based designs seamlessly, and Firefox does a good job most of the time. Usually those half-pixel-off problems aren’t glaring, but Google stirs in some #FF0000 and kicks up the sucking.

So what do we have, in the end? A useful trick that reveals a bug in Firefox, or a Google engineer who was too clever for his own good?

1. Nowadays many websites spend more time receiving, setting up, and tearing down requests than they do actually sending the files. Considering that HTTP only allows two to four images to be downloaded at once, it’s easy to have network latency dominate your load times. [↩]