Palm Dishes on the PDK (Video) 37
You got most of the deets yesterday on the news on Palm's Plug-in Development Kit: it's coming to the Pixi, regular developers will have to wait for a couple of months before they're able to distribute PDK apps, and you'll be able to mix and match the PDK with the regular SDK.
Rene from TiPb spoke with Joe Hayashi from Palm yesterday, detailing the PDK in broad strokes in a way that the rest of us non-developers can make sense of. Palm's main message is a subtle dig on Android and iPhone: if you're on Sprint or Verizon and want cool 3D games right now, your best bet is the Palm Pre.
What interests us the most is how much thought Palm is putting into integrating 'native' PDK stuff with the standard webOS web-app architecture. It will be pretty cool to see apps that utilize a combination of quick-to-code standard elements with the fancier and shinier PDK elements. Right now PDK apps are just games - but we're excited to see utility and productivity apps.
We'd say we're pretty bummed that developers won't be able to release their PDK apps right away, but from how we understand how the PDK works right now, basically it looks like PDK apps completely un-sandboxed and able to do whatever they'd like to webOS' linux underpinnings - potentially very bad. So a future update to webOS will definitely be necessary (and welcome, as it always is) to get a wider distribution for PDK apps.
37 Comments
Hey Dieter,
Thanks for the heads up.
Just one comment.
"...basically it looks like PDK apps completely un-sandboxed and able to do whatever they'd like to webOS' linux underpinnings - potentially very bad."
I'm not following what you're trying to say here. Could you elaborate a little?
Keep up the great work!!
I believe he meant:
"...basically it looks like PDK apps completely un-sandboxed *ARE* able to do whatever they'd like to webOS' linux underpinnings - potentially very bad."
An example is that the current PDK has access to all of \media\internal\. So it is possible for me to release a cool game that also looks an sees if you have certain apps installed to modify their SDK files. Perhaps send all photos to a server somewhere, maybe steal all you contacts.
There is discussion going on currently in the Palm Dev forums about the level of access we will get for file I/O. All of \media\Internal would be nice, but dangerous. However, to write a text editor, you would need access to where ever the user stored the text file on the USB drive portion.
As always, a perfectly secure system is unusable and a perfectly usable system is insecure. The idea is to find the proper balance that supports both sides.
Thanks coasterer and saccherjj for the clarification and insights.
I just wasn't making the connection on what he was trying to say.
Thank you for the help.
Hopefully Palm will be able to find a "happy" balance so that developers can be free enough to do things that users will want while still keeping everything secure enough to minimize abuse.
Yeah Dieter,the completely un-sanboxed,& able to do whwtever they'd like to webOS linux underpinnings-potentially very bad.I don't get your point of view,like MotorDr said: can you elaborate a little more?
I believe he meant:
"...basically it looks like PDK apps completely un-sandboxed *ARE* able to do whatever they'd like to webOS' linux underpinnings - potentially very bad."
Why do we need to wait for few months? I bet they want it released with palm pre 2. Maybe Web OS 2. All the 3-D games/PDK apps are working right now. It will just push the development of more apps beyond 2-3 months.
i think it has to do with security issues with the current pdk.. that would explain why trusted partners are the only ones making PDK apps right now...
WebOS 2 is like 3-4 years away, I hope they don't wait that long to release next gen hardware. That's just what they need, two O/S's that are not fully developed.
Disappointing we have to wait a few months but i suppose its still good news. Looking forward to some decent apps. The 3d games are great but i just want some useful apps
Hopefully our KICK A$$ HomeBrewers will find a work around this non-sense, come on Palm get it in GEAR!!
Looks like a win, only time will tell
it means that just any developer can develop programs that can a. Steal your information, b. Listen in on ur calls, c. Take pics with ur camera and see you nekkid or doin ur gf, c. Steal ur nekkid gf pics, d. Turboheat ur phone n make it jihad ur pocket
Started playing with the PDK last night... I quickly remembered why I was happy to leave the world of C/C++ for web development... too much like math with symbols all over the damned place :)
But that's just long-dormant portions of my brain not fully waking up yet, so no big deal.
There's a lot to like about the PDK and what it means for webOS software, but I can't get past what I see as the negatives.
Primarily, what it says about the whole web technology approach of webOS and the nature of open web technologies.
Doesn't this in a sense validate somewhat the people who said you can't have a proper OS build on HTML, CSS and JavaScript alone? I've never believed that, and since the beginning I felt Palm and webOS has been a shining beacon of light proving that naysayers are wrong.
But, isn't the very existence of the PDK a tacit admission that there might be something to those arguments? I mean, if we could do Need for Speed with web technologies, wouldn't we have? (yes, I'm ignoring the point that the game existed already and was apparently a pretty easy port, which is a HUGE plus in the PDK column). Moreover, we now have to wait and see if Palm lets up on the gas at all in terms of enhancing what you can do with Mojo, and this leads to my second point...
Think about desktop browsers for a moment... there's a big push on right now to move away from plugins and proprietary approaches and use open technologies, HTML5 and what it brings mainly. The PDK is essentially embracing the approach we're trying to get away from!
I think what I'm really saying is I don't have any problem with the PDK per se, and like everyone else I'm excited about the apps it'll make available to me as a customer... but from a philosophical perspective I think there are some concerns that stem from it that worry me. These worries may prove to be unfounded, only time will tell as we watch and see what Palm does aside from the PDK. If this becomes *THE* way to write webOS apps though, or "great" webOS apps, then I see that as a big ding against the webOS platform as a whole.
I still see a majority of apps built on SDK. What the PDK allows you to do is have a small plugin for certain functionality that isn't a part of the API yet.
Agreed. I don't think the PDK is an admission that WebOS is "wrong". The "plug-in" analogy is perfectly fine ... there's plenty of things that we see on the Web that wouldn't exist without plug-ins, but that doesn't mean that basic HTML/CSS/Javascript isn't still good for a large majority of what we do.
To me though, it's the question of whether you want to perpetuate the plugin model we've had for years, or instead start going away from that. I agree of course that plugins have made lots of things possible that wouldn't have been otherwise, and I've benefited as much as anyone from that... but the promise of standards-based development is, IMO, much larger, and a vision worth pursuing.
With webOS Palm has, I think you could argue, done more than anyone else to show how you can have a web technology-based operating system, how that model of application development is powerful enough now to do that. I personally was hoping they'd take up that mantle fully and not, in a sense, muddy the waters by perpetuating the plugin model. I was hoping they'd say, loud and proud: "You can develop ANY KIND of application for webOS with just standards-compliant web technologies and we're going to do everything we can to make sure you really can do ANYTHING that way, including top-notch games". It was probably just overly wishful thinking on my part that *any* company would do that :)
Believe me, I realize what I'm saying is probably a fine line at best, and as others have said, I'm talking philosophy here, which often (probably even frequently) doesn't line up with business reality. I definitely see where the PDK helps Palm, and we all I think agree that they need whatever help they can get right now. If the PDK attracts some developers, both big and small, and that expands webOS' app ecosystem, if ports are quick and easy, that in turn all makes good business sense for Palm. I get that.
And I've said this before too: maybe the PDK will prove to really just be a stopgap solution. Palm has dropped a number of hints that they intend to make Mojo development on-par with anything you can do with the PDK. A year or two from now maybe we'll all be saying "Thanks for the PDK, but we don't need it anymore". I sure hope that's the case. I just worry that with the PDK here now that there's much less incentive to keep pushing in that direction... you could argue it wouldn't make good business sense :)
I wonder if Palm's vision is to always have these two alternative methods for developers to write applications... my point is that if you believe the web tech model is the ultimate destination, which webOS seems to clearly be saying by its very nature, then why would you ever need the other option? To be clear, you may need it NOW to build the platform up, but is that the ultimate goal? I don't know.
I do want to clarify one thing: I don't at all think Palm is saying that webOS is a failure, or that standards-based development is a failure, and *I* am certainly not saying that! However, I can see how someone could interpret the PDK that way.
Now, what I will say is that if a majority of application development a year or two from now is PDK-based, then I *will*, I suspect, claim webOS is a failure IN TERMS OF PHILOSOPHY. I think webOS is a great, forward-looking technology that places a big bet on the web development model (which a plugin is not just because it hangs off a browser engine). If most people are doing PDK development down the road then I really would interpret that as a failure of the vision webOS is based on. It may still be a business success, could even be the preeminent mobile operating system in the world at some point, but it will have failed to deliver the promise it started with from a philosophical standpoint IMO. Only time will tell of course, but it's fun to think about now :)
I think the issue is that the market per say, isn't moving in Palm's direction. In a perfect world they would have liked to stay focused on the full web platform based upon XHTML/css/JavaScript, however let's be real. Windows Mobile 7, Android, iPhone OS are all pushing their OS to a mobile gaming platform ALONG with meshing all their services into a cloud environment like Palm did. They should be still sticking towards building their current applications on the very fundamentals they originated on however this allows people who want more functionality to come into play. The development of the PDK was there for people to be able to port in applications to draw customers towards their product. They need to do something or they are toast. You can't expect HTML5 to be ready while Palm sits there and flounders. Let's be honest here, I don't see the implementation of any new technologies for the web being feasible on an mobile platform for another year or so. Even when they come out, it'll be lagged when developing.
If they kept to their philosophy they envisioned, while revolutionary, it's not the way the market wants to go. If they didn't release these options, they would have been dead in about 6 months to a year.
I see you're concern, but i think the point Palm is trying to make is that the web based platform is quite flexible. i think the browser analogy is very appropriate. We are able to do so much more in our browsers than was ever intended thanks to the plug-in model. I think the PDK speaks to that same quality in WebOS.
I wouldn't say the OS is built on JS/HTML/CSS - the UI is, sure, but there are lots of background libraries/services (mostly written in C) that the UI talks to to actually get stuff done. I just see this as a different way to access those services, via natively compiled code instead of JS.
I think with HTML 5 and other web technology like portable flash and web opengl, will later have the tools to keep it more specific to web development. But for the time being I think it important to have other flexible ways to bring in powerful programs and more importantly to get MORE apps and the best way to do this is if they can easily port from 80,000 iphone apps. Even if WebOS doesn't currently have a huge user base, if it is easy to port to it,it essentially creates a way for iphone developers to get a few extra dollars on a smaller platform. This might help Palm get a bigger user base. Very important.
From my perspective, one could say that was (and is) great from Palm to embrance the web and trying to attract current web developers, because the future is there, but the thing is that we already know that the current web applications on Internet are not written really just on javascript.
The ruby on rails people, the php one, the java and .net one, etc... they still program server side with "compiled"-ish code (lets say they are a little bit more "native" than javascript), and javascript is used just on the client side (notable exceptions can come with things like node.js and so, but I hope you get the point).
So I see the pdk as the "server-side"ish technology we cant avoid yet for some things. But is good cause I assume we can create our on "services" to use from our shiny MojoSDK app and that can give us more freedom (even if more complexity).
About philosophy... sometimes a philosophy alone doesnt make business sense, unfortunately. But I still think they are trying to show that they are still pushing web technologies. Not only by talking about WebGL and css transform, but... how else the very same day they release the PDK can they release an article about making games...with CANVAS :P (that you must really know :P)
There is a single problem I see with the PDK, is not a real problem by itself but Im kind of worried about it:
If palm wanted now to really made the backend of their core webOS applications (mail, messaging, phone, etc) in C for possible performance improvements, and so release them compiled straightway, I wonder what would happen with the homebrew movement that grew up around webOS preciselly because we could access and PATCH the code of those applications cos it was all javascript.
That patches thing is something no other mobile OS that I know could have in such an easy way.
I'm sure they're developing a way to make the user authorize certain access to their phone, such as is done with gps location services
That is the suggestion many Devs are asking about.
"This Application Requests access to Read and Change the contents of your USB Memory section."
Even with this, certain parts of the \media\internal will still be locked down, such as others source code.
I'm fine with them waiting a few months. As long as the details get hammered out and bugs fixed. I just hope that waiting too long doesn't mean someone else comes in and takes the moments away, like how Verizon was excited and saying they were going to get the pre six months after it came out then the droid took all of that thunder.
If these PDK apps function like that apps that are available then no thank you. I have to restart my phone EVERY TIME i was to play a 3D game. Not only that but they don't save my progress. I was hoping this problem would be resolved with the 1.4 update but that didn't happen. How long do I have to wait before the apps in which i paid work properly?
The reason you have to restart your phone is because you don't have enough free memory to run the apps. When building a PDK app, it requires the developer to specify the maximum amount of memory required for the app. So let's say your device has 60 MB free and the developer specified that the game requires 64 MB to run. In that case, you'll get a "Sorry not enough memory, close some cards" error.
Keep in mind that most of these early PDK titles are ported from the iPhone that does not allow 3rd party apps to run in the background or multitask. So their apps may not have all of the necessary memory optimizations required for webOS which does allow background apps and multitasking.
Hopefully developers will start creating new games with memory optimization in mind and their memory footprints will become smaller and thus decrease the risk of "Sorry not enough memory" errors.
And for regular apps that are using the PDK to implement smaller plug-ins, I doubt this will be an issue since those types of apps should not require too much memory.
this video is not being played in my palm pre.
...that would require the Flash Beta Plug-In my friend ;) Don't worry though, it should be available for free in the App Catelog very soon!
It is a Flash video, not YouTube. We can't play it... yet.
so when can we expect shazzam then? 2 months? This new pdk should allow mic access correct?
The PDK does not allow Mic access ATM.
I know you are going to kill me, but building 3D games based on Web Technology is really sick.
Of course, there are some WebGL demos, but as long as interpreted languages (JS) are slower then native code it will hopefully stay this way.
The overall performance of the Pre ( opening apps, using the calendar ) proves that using web technology for everything is not the way. GPU support for JS won't solve this problem, it will just make it look nicer.
Don't get me wrong, I love many concepts of my Pre and use it in everyday life, but you cannot cross the ocean with your car just because everyone can drive one. Sometimes you have to use a faster transport.
Just my 2 ".
Are you trying to say that part of the sluggishness of WebOS is due to the nature of the OS? Does this mean that it could never be faster? If so, could they change over to "native", or is it too late?
Also, "sick" is a good thing in today's slang. Like cool, or awesome, or groovy or whatever. You could also say raw, sweet, or tight.
I'm not saying it won't be faster, I just think its slower then it could be. They won't change it as they take the advantages and their progress over these disadavantages. You might compare it to web Services ( Google Calendar, Docs ), they are slower but people use them.
And "sick" is translating german slang into english, but in the future I might use raw or tight as they sound groovier :)
For the codeheads: While I don't know for certain, I suspect that existing PDK-based games are statically linked. This pulls the library code into the binaries that are being distributed. It makes sense to use this mechanism when you're developing a new set of APIs and the implementation is fluid.
I haven't looked at what is available in the developer PDK yet, but my guess is that when support makes it into webOS, dynamically linked libraries will be used. These are probably available now in the PDK, but with no redistribution capability.
There should be at least three added benefits. First, the code size of PDK-based apps will shrink. Second, overall stability will improve. Third, Palm can choose to accelerate APIs used by the user interface itself and Palm-supplied applications, leaving intepreted code in-place for less critical background processes.
Man im excited!!
It will be more interesting when people develop libraries in the PDK that can be distributed to people who only can code in the sdk.