Safari will automatically execute some Unix shell scripts if the “Open ‘safe’ files after downloading” preference is checked. That’s the kind of mistake Microsoft would have done five years ago, and goes to show that, even within Apple, there are people seriously overestimating the platform’s immunity to malware. That’s frightening.

And, actually, heise online’s downloadable example [via] is much more far-reaching: it’s a zip file with a .jpg inside, which Finder displays with the same icon it uses for jpegs on my system (Xee’s icon, not the regular Preview icon), yet if you double-click it a Terminal window will open and execute the script’s contents (in this case, a simple ls). I don’t know about you, and maybe I just missed a security advisory when I switched, but I thought that, if you’d configured OS X to always display file extensions, you could safely double-click any document file as long as it had the extension it should have. Was I actually spoiled by Windows?!

Mind-blowing quote:

When the University of California at Irvine campus was first built, they just put the buildings in. They did not put in any sidewalks; they just planted grass. The next year, they came back and built the sidewalks where the trails were in the grass. That’s what haxies are to the Mac software market. Haxies are those paths in the grass.

HDR Photography

For a couple of days the web has been abuzz with links to the HDR pool on Flickr, so I figured that was a good opportunity to take my camera out again, after several months without touching it at all.

HDR (high dynamic range) consists basically of storing in a picture file much more information than a computer screen can display. A regular bitmap file (e.g. a JPEG you create on Photoshop or with your camera) stores each pixel as three 8-bit values (for red, green and blue) which, combined, give the pixel’s color — that makes for 16 million different colors, and it’s already slightly more definition than your average computer screen can really display. A digital camera’s RAW file typically stores 10 or 12 bits per pixel — that means four to eight times (one more bit equals twice as many possible values) more definition (I’m not talking resolution, here, but luminosity values — actually, the resolution itself, i.e. the number of pixels is three times less, if I’m getting this right, but that’s not the point), which is why it’s recommended to shoot RAW if you have enough memory card space, so you can correct the exposure afterwards and reveal highlight/shadow detail that would have been definitively lost if the camera had saved the picture as JPEG. Finally, in an HDR image, pixels are usually made of three 16-bit or 32-bit values, and they’re floating-point rather than integer, which means that they range from zero (pitch black) to infinity (supernova) rather than be clipped at either end (like zones being all white because there’s too much light).

Okay, that’s fascinating (or not), but how does it apply to real-life photography? Simple: as the most common example goes, if you’re in a cathedral and want to take a picture of sunlight through the stained glass, either the glass will come out completely white or the inside of the cathedral will be black, depending on which exposure you or your camera chose. A regular bitmap can’t capture the contrast between direct sunlight and the darker shadows (which is also why, if you’re not extra careful, landscape pictures often come out with washed-out skies); an HDR picture can:

Hold on, now: that was a JPEG, and it was displayed on my monitor, how could it be HDR? Well, it can’t, and that’s the trick: an HDR picture can theoritically store the full possible range of luminosity as seen in the real world, but there’s no way it can be fully reproduced. You necessarily have to convert it to something a computer screen can show (there are HDR displays, but even then it’s still an approximation — unless you have a fusion reactor inside, they can hardly emit the same amount of light the sun does). But HDR is still useful (beyond photo-realistic motion blurs in Half Life 2) in that all this information you can’t display is inside the computer, which means you can use it to improve detail in the highlights and shadows of your photographs, and fiddle with settings until you get the very best possible picture of your cathedral, with everything as clearly visible and as detailed as can be.

Now, if all you’ve got is 10 or 12 bits of information per pixel in your digital camera, how do you make a true HDR photograph anyway? Quite simply, you just take several pictures at different exposures: one exposure optimized for the stained glass, one for the detail on the walls, and one or several in between (which means it’s only applicable to still life and landscapes — no portraits or action photography unless you find, or build, or invent, a true HDR camera). Put them all together and you’re all set. Well, actually, not — that was the easy part. Now you have to convert the HDR picture into a regular JPEG, and that’s when you start to get headaches.

In the rest of this article I’m detailing the process a bit and posting examples made with Photoshop CS2’s “Merge to HDR” feature and with Photomatix. Spoiler: most experiments were pretty disappointing.

The old Nintendo DS’s design was clunky and lame; the DS Lite totally makes me want one. For when I’m doing laundry or something. Or whatever. I just need to find the right excuse.

Especially if it’s going to have a TV tuner and a wifi web browser [via] (although I have some doubts about the idea of browsing the web on two tiny screens, zoom or no zoom).

But then, do I really want to look like a fool, petting my Nintendog in the laundromat?

LED Throwies [via] [via]: one dollar’s worth of material per unit and the simplest possible assembling instructions (put things together, roll in a bit of tape) gets you magnetized LEDs you can put or throw as decoration or graffiti onto any metallic urban surface. Insanely cool — too bad it’ll only last a week or two, it’ll quickly get expensive.

La presse envoie sa facture à Google :

In Belgium and the Netherlands, where the service was just launched, Google News has opponents, too. The Dutch Publishers Association considers that Google should remunerate the online newspapers it quotes. As a retaliatory measure, the Belgian newspaper De Standaard has decided to block its links to Google News.

Heh. They didn’t mind linking Google News as long as it wasn’t their content that was copied?

More seriously, even if “35% of Google News visitors won’t go any further on the information websites”, that’s still 65% of them liable to go and visit, plus 35% who wouldn’t read you anyway but see the source name displayed — in other words, it’s advertising, lots of, more than you could ever afford. And, of course, you complain.

It’s really amazing how clueless all content-production businesses can be.

XPize [via] replaces all the old 16-color icons and graphics Microsoft never bothered to update in Windows XP, for a more consistent, polished interface. It’ll still be Windows, but cleaner. Not tested, but the screenshots look nice.

It’s creepy how the full-screen iPod rumors make sense with all the tablet-PC patents that cropped up in Apple’s portfolio lately. Can’t you totally imagine Steve Jobs deciding to disguise the future iPod’s patents as tablet PC stuff, so nobody could see it coming? This could be big — much bigger than just an iPod video.

Or they could not be ready at all, and we’d have to wait for the next generation before we can carry a Mac in our pocket. Steve Jobs seems to embrace planned obsolescence like a madman lately.

SetAlphaValue is an input manager (yeah, evil, I know) that makes OS X windows become partly transparent when the associated application is in the background. The interface is ugly and not quite configurable enough, and the ergonomic benefits of the idea are pretty dubious anyway, but that’s the kind of toy I played with on Windows XP, and I do intend to play a bit more with it now that I’m using an OS that can actually handle transparent windows without choking.

I was going to write a detailed post on imeem [via], the Flickr / Blogger / Myspace / MSN would-be killer with a ridiculous name, but I’m not so sure it’s worth my time (and support didn’t feel like answering my email requesting some clarifications). The website looks good, the convergence works pretty fine (besides blogging / photologging / podcasting on imeem, you can also import your blog and Flickr account’s RSS feeds to display on your public profile), but it relies on client-side software to generate the pages, only loading from their servers encrypted data rather than HTML pages, and the OS X client lags far behind in terms of functionality. Considering the client essentially a web browser that generates its own pages, I don’t see how hard it could be to move the HTML generation into a cross-platform library that could be updated simultaneously on all OSes.

It’s a pity, because the web pages and functionalities are actually well-designed enough that it could stand a chance, if it were usable without the proprietary client. (For now your public profile exists on the web, but you can’t do much in the way of administration or communication on a regular web browser. Not even write blog posts.) All so your data can be encrypted — hey there, ever heard of https?

 

There’s been a lynch mob over at MacDailyNews because some tech columnist complained about his difficulty installing iTunes on a Windows PC. The author started his rant with an “Apple cultists” joke, because that’s what columnists are paid to do, write rants and try to be humorous; sure enough, Mac fans were quick to prove him right.

You know what? Apple is widely known for making cool machines (from computers to music players or even defunct PDAs); they’re insufficiently known for making a great operating system; what they’re not good at, however, is making Windows software. (And why should they? They’re Apple! Well, they should because they want to sell iPods to PC users.) More than once I’ve been embarrassed after I recommended to a Windows user to install iTunes, because it’s so cool and convenient and well designed or whatever. Sure, the program’s design is the same (except for iTunes 6’s menubar that ridiculously ends up in the middle of the title bar), but the most striking difference, besides resource usage (never really compared, but I wouldn’t be surprised if the Windows version was even more of a resource hog), is the installer: it takes eons, it forces QuickTime Player on you, and more often than not it fails.

Oh, I’m sure on a brand-new PC, or a very clean Windows setup, the install would work just fine. (When I was on a PC, I never had a problem with it. I actually used Windows XP for years without a real hitch.) But the thing is, real-life Windows setups are anything but clean. That’s the way the world is, that’s the reason we bought Macs, there’s no changing it, and, more importantly, most other programs manage to get around it. How is it not Apple’s fault that iTunes fails to install on a system that manages to install and run Office, the Adobe CS suite, Firefox and Thunderbird, and whatever other applications average Windows users have? What’s so low-level about a freaking music player that it’ll trigger unfathomable incompatibilities with anything but the cleanest Windows systems? (Oh, I know what: either QuickTime — and if iTunes for Windows is coded in such a way that it actually requires QuickTime, then it’s poorly designed — or the Rendezvous / Bonjour runtime — and then why don’t they just disable it, make it optional or something, if they can’t get it to work? it’s not really the most important functionality for someone who just brought an iPod home.)

Everyone’s been talking about the iPod halo effect: making the iPod so representative of Apple’s perfection that it’ll make PC users want to switch. Well, if PC users are judging the quality of Apple’s software from iTunes for Windows, there isn’t going to be much more switching than there’s already been.

(And, of course, when Apple users do behave as cultists, it doesn’t help, either.)

Apple is possibly maybe preparing a widescreen iPod with a virtual touch-screen click-wheel. What do you think would be most annoying: the perpetual smudges, or the inevitable scratches on a surface even more fragile than the nano’s?

Not to mention that a virtual click-wheel provides no tactile feedback whatsoever, so unless you pay for and burden yourself with a remote control, there’s absolutely no way you can navigate your music library while the iPod is in your pocket^*. But then, maybe that’s exactly why Apple’s FM adapter doubles as a remote.

I still think my idea for a widescreen iPod was better: a regular screen (though smaller and of lower quality than the current model’s) and click-wheel on the front, and a full screen on the back. You don’t really need that Apple logo there that much, do you? But there’s no way His Steveness could ever consider that: it’d make the iPod a hair thicker, and these days it looks like he’d almost remove the battery altogether just to make the device a couple microns smaller.

Anyway, such a drastic redesign would certainly account for the 1GB nano being too unimportant to make it to a Steve Jobs keynote, so I guess it’s pretty credible.

But what would be the difference, on the hardware side, between this iPod video and a born-again Newton? Could Apple justify using touch-screen technology without providing additional functionality, or would that actually a way to reintroduce an Apple PDA without much fuss or emphasis on the productivity aspects?

^* A nice suggestion, though, in a Think Secret comment: gestures, rather than the virtual click-wheel, could let you control the iPod without taking it out of your pocket. Drag your finger left or right to change tracks, up or down for volume, double-tap or something to pause… that could work. But I don’t think we’ve ever seen Apple express any real interest in gestures. Or have they? Oops, I’m so not interested in gestures personally, I forgot that OS X’s own Inkwell (the handwriting recognition technology mostly inherited from the Newton) does support them — only I never used that.

Apple silently updates the Apple Store to include a 1GB nano for the same price the shuffle went. Sure, it didn’t deserve an Apple Special Event all by itself, but how could His Steveness not want to drop a “And now the nano is more affordable and cooler than ever!” in his next keynote?

When you think of it, it’s a bit unnerving to see they can sell a nano, with its cute (fragile) screen and cute (fragile) glossy finish, and compatibility with the Apple FM tuner and all iPod accessories, for the same price as they used to offer the shuffle. Makes you wonder what kind of markup they made on these lately, and when you start thinking about Apple markups it always makes you itch a bit.

Of course I want one, even though I’m starting to feel cramped in my shuffle’s 1GB. But if money was no object — and it isn’t, considering I haven’t got the money for a 1GB nano either — I’d rather they introduced an 8GB nano instead. Which, if they’re non-announcing the 1GB variant today, isn’t likely to happen for a while.

“A University of Chicago study shows that people overestimate their ability to convey tone in email messages.” Which is why my e-mail/IM communications are so chock-full of smiley, you can have a hard time believing they’re written by the same person as my blogs.

(Well, it could also be useful in blogs, when I’m ironic and a certain part of the audience doesn’t grasp it, but I can’t get around to it, it’s just wrong.)

World of Warcraft: Alone together?

The social factor at work here clearly has little to do with direct interactions and camaraderie in the context of quest groups or guilds. Instead, it looks as if other players are mostly: […] social presence (the constant flow of chat in the wide-reaching “general” channel, the movement of other avatars around you, make playing WoW somewhat analogous to reading a book in a densely populated café)

Speaking of WoW, I was going to mention in passing the thing about forbidden gay-friendly guilds and link to Kotaku and the gay legal organization’s “ You can’t tell gay and lesbian people that they have to be quiet so other folk won’t harass them. If you want to stop harassment you have to stop the harassers not the victims.” But then I read comments on both posts and that reminded me why MMORPGs annoy the hell out of me — people.

The story of Mozilla and Firefox by Ben Goodger.

Xee

Xee 1.1 is a freeware image viewer and browser for OS X. On Windows, I usually had to use ACDSee, which isn’t free and, more importantly, become atrociously bloated over the years; since I switched, I had to make do with Preview because none of the alternatives suited me.

Xee keeps Preview’s simplicity, but adds the most important functions you need: elementary file management (okay, you can Cmd-Delete files from within Preview, and use the titlebar proxy icon to move them around, but specialized commands are better — especially those ACDSee-inspired ones that remember a list of folders you usually copy/move images to), lossless JPEG rotation, automatic rotation of photographs according to the camera orientation (works with my Canon G3, while OS X’s Image Capture ignores it), a highly informative status bar and… and… wait for it… previous and next buttons! You know those? That doesn’t just mean you don’t need to use Preview’s drawer anymore (I don’t hate it, it’s cute — I’m not against drawers in general); you can also browse files ACDSee-style, i.e. opening one (just one) in the Fider and using Xee’s previous/next commands to view all other images in the same folder. If you’ve only ever used Preview, you may not realize how much more natural that feels.

Icing on the cake, Xee is evidently the work of a clever programmer. It’s fast and lean, you have to manually ask it to associate itself with each file type, and the shortcuts customization window is a pleasure for the eyes (although the way it informs you a shortcut is already taken isn’t prominent enough). Unless some unexpected issues crop up during daily use, this is an absolute must-have.

 

While we’re at it, you should have a look at his downloads directory: a bunch of widgets, some great screensavers (especially LotsaEscher and LotsaWater) and a simple but surprising game (Swear, basically Snake on a sphere/torus/etc.). Goody.

 

Objective-C First Contact

It all started on Windows 3.1, with Borland C++. The language was simple enough, the Borland classes were pretty well thought out (I never had the misfortune of touching the MFCs), and for the time it was a rather good development environment. (Ironically, most of my time programming Windows was spent working on a NeXTstep interface clone — at a time when Windows customization and OS emulation weren’t as common as it is now.) Later on, I tried BeOS for a bit, and you could tell it was really designed from scratch to be programmed easily, unlike a certain antiquated Microsoft system. But you know what became of BeOS, and I only started programming again when .NET was beginning to get traction. Although it could get really hairy whenever you wanted to do something that wasn’t specifically included in the .NET classes and had to reach for the Win32 APIs, C# was amazingly convenient to write: just like C++, only simpler, with garbage collection and all. Fun! (That’s when I made Ghrone, a skinnable translucent desktop clock that isn’t obsolete yet.) Oh, and obviously I routinely do PHP, too.

At the time I began programming for Windows, the computer magazines (and that was when there were only a couple of them) were filled with stories about NeXTstep, how it was modern and well designed and a breeze to program and a developer’s paradise, and it was going to revolutionize application development. Needless to say, when I finally got a Mac almost twenty years later, knowing that OS X had inherited so much of its programming platform from NeXT, I was filled with expectation. Until I started eyeballing Objective-C code.

Oh boy.

I guess I approached it from the worst possible angle: a quick C++ to Obj-C tutorial (more of a cheat sheet, actually) that shows how to write Objective-C when you’re coming from a C++ world. What the hell was that now? Brackets all over the place, and yet there’s still some C syntax around? If there’s still C in there, why is the syntax for objects and methods so creepy? And what’s with method definitions? How could that ever make any sense to anyone? Those method calls? This is all designed by a madman!

I guess the secret to learning OS X programming is, don’t start out that way. History is your friend; I found it helped a lot to realize that Objective-C was designed many years ago (roughly at the same time as C++, but with a very different philosophy) and that you can’t blame it for being less advanced and modern than C#, which was born twenty years later. And that the insane syntax stems from a decision to keep the language and compilers 100%-compatible with C. And that the NeXT engineers didn’t choose it because the syntax was cute, but because it has very advanced and intelligent mechanisms for simplifying the programmer’s life on complex projects, only they’re not really obvious at first glance.

I haven’t really started programming it, as I’m still knee-deep in documentation (I have this weird habit of reading all the docs before I get started, rather than try out examples as I go — I know you’re not supposed to do that, but it’s always worked for me), but I must say the way the language replaces garbage collection (so very convenient in C#, but admittedly a bit too bloated for computers of the 1980s) with “reference counting”, and adds subtleties like ‘autorelease’, is pretty clever — it took twenty years for Windows programmers to get rid of the “whoever is supposed to delete this pointer that I sent/received?” dilemma, and even now the .NET classes aren’t extensive enough to completely shield you from those system calls moving pointers around. After .NET and PHP I still can’t believe I’m going back to programming without a garbage collector or equivalent, having to keep track of my pointers, but I feel much better knowing the language conceptor had a clear and clever idea of how to help managing them — smart enough that I don’t feel the urge to wait until garbage collection comes out in Leopard.

I’m still not a fan of some discrepancies between .NET and Cocoa development — .NET and Visual Studio are much more automated, and it’s really weird to see Interface Builder doesn’t generate a single line of code (by design, it turns out) — but despite my initial reaction I now do look forward to getting my hands in it. I only need enough free time, which will be a tad harder now that I have rent to pay.

[02/02] Here are some pointers to help you if you’re going from C++ to Obj-C, in the order I read them:

• The C++ to Obj-C cheat sheet — like I said, lacks explainations and disregards everything that’s conceptually different, but you’ll be roughly able to read Obj-C code, and that’ll allow you to understand the rest more easily

• Love, Hate and Objective-C will help you understand many design choices of the language

• Wikipedia has a good page teaching you a bit on Obj-C (although it’s not always 100% clear)

• More about memory management in Obj-C on MacDevCenter and Stepwise (that seems like the most important thing to learn when you’re coming from another language)

• The other articles in the “Cocoa Basics” section on Stepwise, with good information (you’ll notice some articles date back to the very first OS X beta days, so some details may be obsolete, but not the basics)

• The Apple Developer Connection tutorials (may require sign-up, but if you’re going to program for the Mac you’ll just have to): Cocoa Application Tutorial, Getting Started with Apple Technologies, Window Programming Guide for Cocoa, etc.

Hope this helps.

DrunkenBlog:

The obvious solution [for creating a link blog] that probably comes to mind is sucking in something like del.icio.us bookmarks in some form, but that won’t really happen because I have zero desire for the world to see all of my bookmarks, which means I’d have to maintain two sets, and I’m way too lazy for that.

Exactly. It always amazes me when people mention del.icio.us as a replacement for the browser’s bookmarks system (e.g., Flock).

Optimus mini three is out (in May)

The Optimus mini three pages are online at artlebedev.com, and they could take a hint from Apple PR: no surprise to be excited about, not even a tiny detail — and Steve Jobs would tell you: with all the hype there was before, that means disappointment. It looks like something most people would surely like to have on their desktop, but not at this cost.

OS X drivers aren’t written yet, which is certainly a mistake — who’s more liable to shell out $100 for a three-key remote control than people who bought 33% more than average for their computer? (Despite the Optimus mini three being aluminium rather than white.) And it looks like the designer should know.

In my opinion, the second (and worse) mistake is that, if “it’s possible to configure Optimus mini to display additional functions with modifier keys (Ctrl, Alt, Shift) pressed and held on a ‘regular’ keyboard”, why didn’t they include a couple modifier keys on the device itself? Like, press a left-side modifier to cycle between active button sets, press a right-side modifier to access a set of emergency shortcuts? It would cost half a dollar more to build, but the device wouldn’t look that useless.

According to the Answers page, the full Optimus keyboard should reach pre-production in late 2006. Sure, whatever — but if three keys cost $100 (preorder price only until before April 2nd), what’s it gonna be for more than a hundred of them?

OhGizmo!:

There [would be] a little over a hundred keys on [the Optimus keyboard], many of which go unused [uh…]. It wouldn’t have been much of a stretch to program three of those to do exactly what you’re being encouraged to do with the Mini Three. So why release this at all? To give us a preview of the real deal? Or is this a sideways admission that development on the Optimus is nowhere near completion? Is this some daft PR spin? If the release of the full keyboard goes as planned (end of 2006), then what will be the use of the Mini Three?

Well, whatever the case, the device goes for $100 on pre-order, or “more” later. Even that sounds like they’re bootstrapped for cash.

The Optimus keyboard design does precisely include a dozen programmable shortcut keys on the left side. But then, if the mini three (ah, they did take something from Apple: lowercase initials) is, essentially, a fundraiser to enable production of the real thing… doesn’t that make you want to participate?

(Oh, it looks like the site is being slashdugg to death. How surprising.)

P.S. Slashdot:

From the FAQ [artlebedev.com], the expected lifetime of these displays is 5000 hours. That’s a little over 200 days. Even with a “key saver”, this severly impacts the usable lifetime of this device.

You probably won’t need the displays to be on 24 hours a day, but that still makes it die after one year and a half of intensive use — and if you’re gonna spend $100 on this you’ll want to use it intensively, won’t you?

WhyReboot [via] displays the list of file changes a Windows software installer plans to do after reboot, so you can decide for yourself whether you do need to reboot — if you don’t know, nine times out of ten an installer forces you to reboot for no good reason. Freeware, must-have, untested.

Currently, it does not report changes in driver, program, or service state that require a reboot in order to take effect: for instance, if an installer copies new drivers, programs, or services onto your system and configures the system so that these will be started automatically at boot or login, your system will have to be restarted. These changes are not detected by WhyReboot.

 

SmallWindows [via] is yet another Exposé clone for Windows, but this one has the advantage of being free and open-source (which, for a carbon copy of another operating system’s functionality, should actually be the rule). Untested as well (the less I turn my PC on, the better), but I still think the Windows XP graphics system isn’t adapted to this kind of functionality — probably only good for gamers having souped-up video cards with not-too-crummy 2D drivers (which is uncommon, as I understand it).