(Note that I have no idea if this was the actual question they asked, but for my purposes, it’s good enough.)

From this they created WebOS, which ran on their Pre phones.

WebOS was pretty freaking awesome. Everything on the hardware was exposed through a pretty nice JavaScript API; you could play music, access location, dial the phone, check email, contact instant messaging services, whatever. The UI was created entirely with HTML and CSS, and they shipped a pretty nice set of “widgets” (basically, CSS and JavaScript) to quickly add UI elements.

There were tons of features that were really, really genius on the Palm, that has yet to be copied to Android, iPhone, or WinPhone. If nothing else, I still believe that the Palm/WebOS touch-based interaction was so much farther ahead and just damn better than all the others.

Anyway, it wasn’t all sweetness and light. There were … problems:

• They used a forked version of the otherwise wonderful WebKit, that never quite worked just like the iPhone or Android browsers, so you had to deal with yet another layout engine.
• Their JavaScript engine was JavaScriptKit, which is a good engine but at the time it wasn’t as fast as, say, v8. Today the benchmarks war is often one of increments but at the time, it was provably slower.
• At the time, people were raving over a bunch of games and apps (mostly on the iPhone) that simply aren’t possible with HTML5 today, much less 2009. Need proof? See this list. Show me HTML5 versions of those apps. Hell, show me Flash versions of those apps.
• jwz detailed his difficulties in getting one of the first apps onto their app store, among other problems.

Palm eventually gave in and shipped a sort of “native” runtime, allowing developers to ship C/C++/whatever apps.

You could make small complaints about the hardware, as well: AAPL was really hitting their design stride in 2009, and the Pre was a noble but ultimately failed effort to make a device that really stands out. It was really good but only when placed against the not-an-iPhone pack.

(Daring Fireball has some thoughts from mid–2009 that are pretty spot-on.)

Ultimately I think had Palm been competing in a world without the iPhone, it would be the #1 device today. They just didn’t get a critical mass of developers to attact enough users, because everyone was gaga over the iPhone.

So what’s this got to do with anything? Well, there’s this: Boot2Gecko. The tl;dr is simple: “Take WebOS and s/WebKit/Gecko/g”. It’s the same story as WebOS.

Seriously. Watch this presentation and tell me, that’s not the same stuff Palm said 3 years ago.

I really wish someone from Mozilla would explain how their thing is going to be any different, other then they have experience making “web APIs”. They’re not even targeting a hardware platform, like Palm did; so they are already 1 step behind Palm, who at least had a reference platform and “total package” right out of the gate.

I’ve reverted to Firefox and Tbird lately, because Safari is going through yet another spurt of growing pains and Google can kiss my ass. It’s pretty amazing how Firefox has been able to turn the ship and go from crashy bloatware to speedy; just a year ago Firefox was basically unusuable for me (and if Chrome’s adoption is any indicator, a lot of people) but I’m actually pretty content these days. I imagine others would, too, but they’ve already leapt on board of Chrome and one browser switch per generation is enough for most people, especially if the story is simply “hey, it sucks less now”.

So I’d really like to know why I should bet the farm on Boot2Gecko; why I should tell my bosses, “No way, you guys, the future isn’t Android and iPhone, it’s gonna be this MozillaPhone thing”. I’d really like to know how they’re going to overcome all the problems that killed Palm.

Let me translate some of this for you.

I’m becoming more and more certain that this means that Rails-style MVC frameworks on the server-side are going to end up being phased out in favour of leaner and meaner frameworks that better address the new needs of thick-client architecture.

Uh, yeah.

What that means is:

Part of the reason we have these monstrous frameworks is because at some point all software is just one giant if/switch around the seemingly endless list of bugs stemming from one of (client OS, client hardware, client browser, client plugins, deranged and clearly pants-on-head-stupid client workflows, and the vagaries of quantum physics.

But, because programmers think code rusts, and we love to rewrite the HELL out of it, we come up with new and exciting ways to do just that.

Much of the OP‘s complaints can be easily and succintly explained by simple things: IE6 is now dead, most browser vendors have some sort of agreement to not put us all though the same old BS browser war we went through last time. JavaScript is now a “first class” language and not an annoyance for “know-nothing” Web designers or inflicted on the company “rock stars”. Enough JavaScript exists “in the wild” that we have stable platforms to use to build apps.

So! It’s the trifecta of awesome rewrite-your-app time:

“Everyone’s settled on non-blocking IO as the only way to scale.” ”Client-side tempting is where it’s at these days.” ”We’ll less trouble since it’s all the same language and platform everywhere!”

Time to rewrite those apps!

Fast-forward 2 years when “I’m becoming more and more certain that this means that Node.js-style single-threaded engines with client-side logic are going to end up being phased out in favor of leaner and meaner megathreading frameworks that better address the new needs of the quantum-core architecture”.

“A plug-in free Web benefits consumers and developers and we all take part in the transition.”

I’m sort of cynical about this whole “no plugins” thing, when it comes to Microsoft. IE was a submarine made of Swiss cheese for the most critical years of the the web – crashy and a security nightmare.

The reason it was a security nightmare was because of plugins. And it was those plugins – mostly ActiveX controls and Flash, among others – that Microsoft needed as a critical cornerstone of its “embrace and extend” policies.

So: the web community in general looks forward to a world without plugins, it’s sort of annoying to have you guys act magnanimous and not acknowledge that you played a pretty big role in creating the mess.

I don’t get why Apple hasn’t seriously entertained a “next generation” scripting engine, in particular one based on JavaScript. All the things are there:

• They have an excellent JavaScript engine already
• JavaScript is more common than AppleScript (lingua franca of the web and all that)
• JavaScript is pretty popular

There’s the “canonical” JavaScript OSA component but it doesn’t fit the criteria: it uses Spidermonkey, not Squirrelfish/JavaScriptCore. Still, its API is pretty spot-on (since it’s an OSA component). It’s also pretty old.

JSTalk is another contender. It’s sort of the opposite number of the JavaScript OSA: it uses entirely modern, native components – stuff Apple almost literally has laying around – but does not feature an OSA-compatible API (as near as I can tell anyway). But, it’s modern and very friendly with Cocoa.

The sweet spot is the fusion of the two; the “already works on top of an OSA layer” of the former and the “all the modern stuff” of the latter.

Someone at Apple please make this happen! Make scripting the Mac awesome.

I failed.

In part I failed because of Safari, and the way it structures extensions. Briefly, Safari uses a sort of separation-of-concerns model, so what a particular piece of an extension can access varies. Getting the selection is really hard because you need to use proxying and messaging to communicate between the current page, the extension’s “page”, and any resources.

I gave up because it sucked and wasn’t fun.

I had another thought: use Applescript to get the selection and pass it to the Delibar app itself! GENIUS!

The tl;dr here is “Delibar isn’t scriptable” so none of this matters, anyway.

Safari’s Applescript dictionary doesn’t include the term “selection”. To the best of my knowledge, you cannot use Applescript to get the selection of the current tab. Oh ….. kay.

So, the next logical step is to use JavaScript, because I know it has the ability to get the selection AND I know that there is a ‘do JavaScript’ command in AppleScript for Safari. Now we’re cooking!

Here’s the relevant bits from the dictionary:

do JavaScript (verb) Applies a string of JavaScript code to a document.

and the example:

set theResult to do JavaScript unicode text ¬ in tab

Makes sense, right? Sure. So you’d pretty much imagine that it works just like this:

tell application "Safari"
    set selectedText to (do JavaScript “window.getSelection()” in current tab of window 1)
end tell

Always returned null. Always. I then tried hundreds of combinations of more-or-less JavaScript, including:

• adding a return
• wrapping it in an anonymous function
• using document instead of window

• banging my head against the desk and cursing myself for ever having these sorts of thoughts After lots of searching, I discovered the completely obvious answer:

  do JavaScript “(”+getSelection())”

That’s a double-quote, left paren, single quote, single quote, plus, then the function and a closing quote.

Of course. Completely obvious. Note that removing the single quotes and concatenation operator doesn’t work. So “(foo())” doesn’t work, but “(”+foo())” does. I found two places that say it does work, but I can’t make it happen in the latest Lion and Safari.

How the hell do you even arrive at that as a logical place. And how do you expect regular people, ostensibly the target market for Applescript, to arrive at that conclusion?

Applescript is goddamned awful.

The biggest change of that last bit is the switch to Python.

Python is fucking weird.

Yeah, there’s The Whitespace Abomination. The reality is, you get over it. It’s a HUGE deal at first; just awful, terrible, crap. It’s just the dumbest idea anyone could ever have, there’s no sanity to those arguments you see about readability, it’s some magical bullshit thinking. Lisp’s paren soup is NOTHING compared to this bullshit!

It just sort of goes away after a while. It’s still bothersome and weird but I’ve got my editor configured properly now, so generally it’s not really a problem. In fact I tend to end statements with a semicolon more often than making some indentation-related boo-boo!

Python is often called “easy”, and I think it lives up to that description. That said, I think idiomatic Python isn’t easy. Python wants to make it such that there’s one obvious way to do it; but is that one way actually obvious?

In the Python I’ve seen lately, they tend to do the Ruby thing of, “module M contains class C and also a grab-bag of functions”. I’ve spent rather a lot of time in the past few years doing what amounted to Java Lite: large applications are structured by interfaces, abstract classes, and implementations. So, the idea of “here is a class and also some functions” is just … odd. I sort-of get how it works with Python’s namespace/module system, and I’m sure it’ll make more sense as time goes on, but as a switcher it’s positively strange.

Also, tools like virtualenv and the more complex elements of the Python ecosystem aren’t immediately obvious. Coming from a background in PHP or Perl, Python can actually be pretty intimidating if you’re not working in an established platform like Django: what do I do about MySQL? What do I do about web server integration? What do I do about etc etc etc? Scoff all you want, but the fact is hitting http://www.php.net/manual/en/funcref.php is mostly all you need to get rolling on PHP projects, and of course CPAN remains one of the gold standards for code repos. (In fact, CPAN might be too large at this point.)

Anyway. I’ve written a couple of non-trivial Python scripts now, and I’m trying to get more into designing things to be “Pythonic”. We’ll see. I can’t say I’m happy with Python yet, but it’s still pretty new (it’s been all of a month) and so I’m at that strange place in the learning curve.

I feel like I’m leaving behind code and systems that were in better shape than when I arrived. Hard not to when you arrive to a giant pile of Sun hardware, running ancient versions of Solaris. I leave behind 100% Linux, running on commodity hardware, a change I brought to the company.

Our primary application, a horrible mess of PHP4 spaghetti when I started hacking on it, is a slightly less horrible lasagna of PHP 5.3+, and there’s a good first pass at a modern PHP app in the repo as well.

There’s more than a few things I didn’t accomplish that I set out to do. Some of them, we were simply “overtaken by events”. The business has a way of turning underneath you – it’s part of the fun. In other cases, the current business climate simply isn’t right. Austerity is the order of the day, and fewer and fewer customers were willing to spend money to make money (to name but one example).

I’m proud of some of the work I’ve done, but I won’t be sad to see PHP go (the new job is a Python shop). A couple years ago I went all-in on PHP; I even obtained a Zend certification. PHP often deserves much of the hate people throw at it, but there’s also a lot of smart, innovative people who are really trying to push the language/platform in the right direction.

I’m super-excited about the opportunities at the new place. I’m really looking forward to kicking a lot of ass for them. All the skills I bring, I honed in the salt mine.

Contrary to what you’d probably guess, I didn’t get an iPhone; I got a Blackberry, and the free one at that (a 9300 series).

“But Blackberry sucks! They’re going out of business!”, you may think. You’re right, sort of.

One thing I learned since I got an iPad was I don’t like the smartphone form factor. Increasing the size of the device (what seems like the current Android device solution) to me just seems to create a form factor that’s further inferior to both the slightly smaller iPhone and larger iPad.

The things I want most from a phone are the 3 “T’s”: Talk, Text, and Twitter. I’d like decent email support. And that’s … pretty much it. Maps or a web browser or Angry Birds or ssh or 10 billion other things, I simply don’t do. I made myself angry (and spent too much money) trying to convince myself otherwise.

For all of the OMG BLACKBERRY SUCKS talk, it is still a pretty good email and text device. From my point of view, the recent problems with RIM are more about “how did you idiots make so many bad decisions in a row, when you were on top?” more than “your technology sucks”. They’re making bad decisions with an otherwise workable tech platform.

One day, when I go places and do things, maybe I’ll upgrade to an iPhone. Until then, my iPad is all I need for most everything, and the Blackberry is my ideal handset.

I disabled Google Reader sync last night and switched back to NetNewsWire on the desktop, doing local sync.

On the plus side, there’s not a HUGE amount of feeds I follow any more. Also, I never go anywhere so the lack of Google Reader sync isn’t crippling.

BUT: my main use case is, Mac downstairs as “digital hub” and iPad everywhere else (couch, bed, porch, etc). Unless I’m sitting at my Mac to do work, I’m using my iPad; I watch TV on my Apple TV; I listen to music streamed from the Mac via Home Sharing from iTunes.

What I really really want is to have NetNewsWire/iPad to sync to the desktop, much in the way the Yojimbo iPad app does.

I know it’s not some “look at me I’m SUPER MOBILE”, work-at-the-coffee-shop thing, but it works great and is probably (hopefully?) less error-prone than the scenarios Brent describes.