Steve Jobs Announces iPhone SDK

An anonymous reader writes "It finally happened. Steve Jobs announced an iPhone SDK today. The plan is to release it in February, and the suggestion is that apps will need to be digitally signed (not unlike digital signing in Leopard). Here's hoping that developing for the iPhone/Touch will be cheap (or free) enough to allow the folks who have been writing apps to continue doing so. Says Jobs: 'It will take until February to release an SDK because we're trying to do two diametrically opposed things at once--provide an advanced and open platform to developers while at the same time protect iPhone users from viruses, malware, privacy attacks, etc. This is no easy task.'"

Digital signing

[PlatyPaul]

TFS got it wrong: Apple did not in fact say that digital signing was going to be a part of the SDK or making sure that apps are kosher.

Here's the quote that may have misled:

Some companies are already taking action. Nokia, for example, is not allowing any applications to be loaded onto some of their newest phones unless they have a digital signature that can be traced back to a known developer. While this makes such a phone less than totally open, we believe it is a step in the right direction. We are working on an advanced system which will offer developers broad access to natively program the iPhones amazing software platform while at the same time protecting users from malicious programs.

So, what they're really saying is that they're hoping to do something along the same lines as signing, but not signing per se. This actually may be the most interesting part of their announcement, in that it could signal the next step forwards in indicating trust and providing clarity of who worked on what. Here's hoping it's not just repackaging.

Re:Digital signing

[ceoyoyo]

Possibly. One of the new features in Leopard is digitally signed apps though, and Apple is setting up some kind of infrastructure so you can verify the signatures. It would seem likely they'll use the same system on the iPhone.

Re:Ipod touch

[BlakeReid]

Will this apply to the ipod touch as well?

Yep - FTA:

P.S.: The SDK will also allow developers to create applications for iPod touch.

Re:final pieces ...

[ceoyoyo]

My Touch has a PDF reader, SSH client AND an SSH server. ;)

It's great. Best gadget ever. Hacking the touch is pure software too, so you can just restore it with iTunes if an update you have to have comes along.

Re:February is kind of a long time, isn't it?

[goofballs]

no, it's not hard to say; this had nothing to do with at&t- they've long allowed 3rd party apps across a wide range of their phones.

Re:Finally!

[timster]

Well, as an iPhone owner (and, apparently, a Certified Fanboi(tm)), it's plainly obvious that the software wasn't finished in June and is still not finished. While the core features work well for the most part, any iPhone owner can name a dozen obvious omissions off the top of their head. MMS, copy/paste, SMS to multiple recipients, Safari stability, etc, etc. Not to mention an RPN mode for the Calculator ;)

I'm personally happy to have the device now, as it's extremely useful in a variety of ways (hence the fanboi status). But an SDK is only one of many things that are a tad overdue.

Re:Security weakness of their own making

[suv4x4]

The risk of damage would be a lot less damage if every app on the iPhone didnt run as root

They made the apps run as root due to lack of time to figure out the security properly. This is the same reason they didn't release a SDK.

By February, we'll have a firmware with reengineered OS and apps that don't run as root. The SDK will only support this firmware and newer.

Re:Digital signing

[RzUpAnmsCwrds]

Possibly. One of the new features in Leopard is digitally signed apps though, and Apple is setting up some kind of infrastructure so you can verify the signatures. It would seem likely they'll use the same system on the iPhone.

I'm just going to point out that Windows has had digitally-signed apps since (at least) Windows 98, and that nearly every system library and executable in Windows XP and Windows Vista is signed. Vista even checks the signature before you see the UAC dialog, and the dialog for signed apps looks completely different (and has different keyboard shortcuts).

Windows Mobile also has signed apps.

Of course, I'm sure that some Mac fan is going to point out how this is another Apple innovation.

Re:Finally!

[n8_f]

No, your post is just wrong and should be moderated as such. Here is Jobs at the D5 conference on May 30th, about a month before the iPhone was even released:

Q: All indications appear that the iPhone is closed, we'd love to develop apps...

This is an important tradeoff between security and openness. We want both. We're working through a way... we'll find a way to let 3rd parties write apps and still preserve security on the iPhone. But until we find that way we can't compromise the security of the phone. I've used 3rd party apps... the more you add, the more your phone crashes. No one's perfect, and we'd sure like our phone not to crash once a day. If you can just be a little more patient with us I think everyone can get what they want.

Re:Finally!

[Sparks23]

As some of the hacker community will readily point out, splitting open Springboard (the Finder/shell equivalent) in the iPhone, you discover Springboard always had some support for additional applications... and going forward, more was added. In 1.1.1, Springboard even added code added that supported multiple pages of applications... a pretty clear indication that either Apple was planning to add a LOT more apps, or were thinking of third-party dev.

There were lots of other little clues people found that the iPhone had either had plans for a third-party SDK which was scuttled, or had a third-party SDK in the works but not yet announced. So I admit, I am with the folks who are saying that Jobs probably had this planned from day one, but held off on the announcements until closer to the SDK/security methods being sorted out for marketing/publicity/spin reasons.

3 months after the phone was released is not a huge waiting period, but if he'd announced ahead of time that the iPhone would have a native SDK in February, lots of folks would have waited both on buying phones and on doing iPhone development. Instead, now we have hackers who have already worked on third-party native apps, there's all kinds of web-apps to keep those who won't jailbreak busy in the meantime.

Love him or hate him, one thing Jobs knows how to do is build anticipation, and manage publicity. He'll take bad press for a while simply so that he can sit on some announcement to greatest spin effect.

Re:final pieces ...

[Angostura]

Well, I can solve one of them for you, without hacking the Touch.

Have a look at FileMark Maker [insanelygreattees.com]. It is an app that that runs on the Mac and lets you sync and store any .doc, .xls, .rtf, .txt, .jpg, .jpeg, .png, .gif, .pdf, or .html on the Touch (or iPhone) for offline viewing via Safari. It seems to encode the document actually into the bookmark datastream, which seems a bit hackish - but it works.

It's a joy to use on my Touch, and I've tried it with a 100 page PDF. However long filenames screw up the bookmark display seem to be a bit of a problem, so trim them down a bit before syncing.

I'm just surprised that more people don't seem to know about this app.

So, how long before Java?

[DdJ]

So, even if apps require formal signing and they all cost money, I still kinda expect that one thing we'll get is IBM's WebSphere Everyplace Micro Environment.

It exists for PalmOS, it exists for Windows Mobile, it exists for other handhelds, and I imagine that both IBM and Sun would explode with joy at the possibility of getting it onto the iPhone and iPod Touch.

For those who don't know, this is IBM's J2ME/JavaME runtime for small systems. If you have Java on your PalmOS, Windows Mobile, or even many Linux handhelds, it's probably due to this being loaded on or embedded into it.

If we don't get that, maybe we'll get a port of the open-source reference implementation of JavaME:

http://en.wikipedia.org/wiki/PhoneME_(software) [wikipedia.org]

It already builds for both ARM (current iPhone) and x86 (rumored future iPhone) instruction sets.

Either way, looks to me like once there's a general dev kit, a JVM isn't going to be too far off. Anyone want to make predictions about how long it'll take or what form it'll come in?

Misinformation

[Space cowboy]

Have you actually written any iPhone apps ? Or are you just postulating ?

The iPhone essentially runs a cut and trimmed version of OSX, so getting an SDK for it is NOT some massive undertaking

• The SDK is *not* the same as the Cocoa SDK on Mac OSX. They use UIKit (all the classes start with 'UI' not 'NS'). They use CoreGraphics directly (so you have CGRect structures, not NSRect structures). The port of the Foundation library is incomplete (there's no NSNetService or NSTask that I can find, for example, though basic things (collections, iterators, etc.) are there).

I mean, look, despite Apple's attempts to keep people from using their own phones, random hax0rs got a working SDK up within days

• Those random hackers didn't "get an SDK up in days". They ran classdump on the libraries that *Apple* created, and made the headers available.

A iPhone SDK would use a gcc cross-compiler (since the iPhone isn't running PowerPC or Intel chip -- by the way, gcc makes it easy to build a cross-compiler so this isn't a big deal)

• Interesting. Those "random hackers" got the gcc compiler to cross-compile. Oh, but you can't have any methods that return a float (*) like, er, just about every UI class since co-ords are floats in UIKit. Oh, and it can crash with internal errors in cc1. And they're onto the third incarnation of the compiler now. Perhaps it's not so easy after all. The ARM chip is an established supported target for gcc, so building a cross-compiler itself is relatively trivial. Writing the bootstrap code is presumably harder, and writing the support libraries (libarmfp for example) needs to be done as well.

Not a massive undertaking at all.

• And here's where you lose all credibility. Not a massive undertaking to write a new 2D-accelerated UI framework from scratch, trying to be as compatible as possible with the 'Mac' way of doing things while incorporating a completely new input method ? On a new hardware platform for the OS ? With a very aggressive release schedule ? And design it so there will be no frequent (ahem) updates in the future ?
  
  No, that's trivial mate. Tell you what, we'll do you two, in case one breaks - have it to you next Tuesday... Not.
  

Writing whatever they needed for the initial (general public who don't give the shake of a rat's tail about the SDK) release, then writing/polishing a general developer release is so obviously the way to go, I can't believe people are still talking about it. And if you expected 'The Steve' to lay out all his plans ahead of time, you've obviously been in a coma for the last decade. Welcome to the new century.

Simon.

(*) I think this is actually resolved in version-3 of the compiler. I'm still stuck with v2 because I can't get the LLVM part to compiler on my mac for some reason.

Re:February is kind of a long time, isn't it?

[jd3nn1s]

Interesting, however typically there is no necessity for an application to be compiled from source for it to be signed. People could just sign a binary.

Re:Finally!

[Guy Harris]

OS X as a platform has been around for long enough, and Apple took pride in announcing that their phone and new iPod runs on the same platform

Well, sort of. They didn't mention that the iPhone version of the OS has UIKit rather than AppKit, for example. hello.c would Just Work (if you have Mobile Terminal in which to run it), but J. Random GUI App wouldn't.

If the API's have been settled, they wouldn't and shouldn't have released the product.

You need more than stable API's, you also need stable ABI's, and they could well have released it without stable ABI's - all the bundled apps would have to be recompiled if the ABI changed, but that's doable. You might believe that they shouldn't have released the iPhone without stable ABIs; I believe otherwise, and, quite frankly, think arguing that they shouldn't have released the iPhone without stable ABI's is bogus.

Even API changes, although they're more disruptive to the code base, wouldn't be out of the question.

Unstable API's and ABI's, however, do cause problems for third-party apps, so they need to stabilize those before releasing an SDK.

Re:bug report

[Guy Harris]

I wonder if they will limit all things internety to WIFI only, as AT&T might complain about random packets flying over their EDGE

Yeah, it's not as if AT&T sells cards for PC's that support EDGE [att.com], so that any packet your PC could send out over Wi-Fi could also be sent out over EDGE.