End of J2ME for PDA 's from JDJEdge

[ersin eser]

I was at the JDJEdge 2001 one of the sessions was
SE10- Developing and Deploying Wireless Applications Using J2SE
Presented by SavaJe Technologies.
I am glad that picked this session. It was one of the best presentations of the conference.These guys have very different approach to wireless solutions and I think it is extremely logical and promising. This company is mainly after the more advanced wireless market which is PDA,RIM, IPAQ .... Their approach is: why change the language while it is perfect; change the device kernel ! They have written their own kernel and they have licensed the Java Virtual Machine from Sun.
What they do is they put their Kernel into these devices and use fully supported J2SE! It was very shocking to me to see IPAQ to run a Java applet as fast as desktop PC. They have a pretty good shot in this section of the market. J2ME is great for the limited devices as phones but the advancing PDA market might lean towards this direction since this kind of solutions brings in full J2SE usage.
I think this company is going to be bought by a bigger company soon .I also think bigger companies like Compaq, HP, SONY who also manufactures these devices will develop an interest for this kind of solutions.
And their usage of web start using .jnlp files on IPAQ was pretty impressive .They weren't doing any optimization for IPAQ they run the application on the PC download it to IPAQ and run it there, project it onto screen Damn ! I really don't care about using J2ME for PDAs or PALMs any longer while I can use J2SE.

This guys are very smart ( Computer Science graduates from MIT and MBA 's from MIT Sloan School of Management).
www.savaje.com

[Mark Herschberg]

I spoke to these guys back at JavaOne in June, 2001. I think they have a cool idea, but it's not going to make it, for the following reasons.

1) Java in not quite an OS
This is my biggest complaint about Java. Java straddles the line between language and OS. Often when writing J2SE apps, we find we either have to jump through hoops, or use JNI to do what C programs can do easily.
I haven't written an OS in Java, but it seems like it has some, limitations, simply because Java wasn't quite designed to be a full OS. This is starting to change with things like JSR-96 (daemons) and others, but I think we're still a long way from it.
2) Corporations like WinCE
Granted, WinCE's appeal isn't as strong as on the desktop. There's no way Java couple replace windows in the near future. However, corporate IT departments like WinCE. Why? Because it's MS and it's just like windows. First, you've got the brand name. No one ever gets fired because they went with MS for an OS. Second, you've got a lot of familiar apps, including Pocket Word, Pocket Outlook, etc. Pocket Outlook can sync with the desktop version, making it very easy to integrate into existing corporate infrastructures. Other OS will have to compensate for this, as it's pretty important functionality.

I agree that this is technologly a cool idea. And as a Java fan, I like that it's Java. But I think they face an uphill battle. (And, btw, I seen plenty of dumb ideas from MIT. Generally speaking I was not that impressed by the classes at Sloan.)
--Mark

[ersin eser]

Well Mark I have visited your web site a while ago and read a lot about your ideas & I respect you alot . I think your reply makes a lot of sense. I still have this wishful thinking about using J2SE on PDA's; it was very impressive...
I have sent an email to savaje about this topic I hope they respond and give us some appropriate answers.
thanks for taking the time to answer my post
sincere regards
ersin

[Lead Wey]

Ersin,
Thank you for attending my presentation at JDJEdge, and thank you for your feedback!
Here's my response to Mark's comments:
1) Mark's comment that "Java in not quite an OS" and "Java straddles the line between language and OS" is precisely the problem SavaJe is addressing. Java is too big and too slow for smaller mobile devices, and running Java on top of another OS (WinCE, Linux, PalmOS, etc.) not designed to run Java is simply too inefficient.
2) I agree with Mark that writing an OS in Java won't work. In fact, Sun Microsystems tried it with their "JavaOS" effort and failed. SavaJe has learned from Sun's experience, and we believe we have done it right this time. SavaJe's OS has a real-time, multi tasking, multi threaded kernel written in C, not Java! The J2SE VM is then integrated directly to the kernel, and Java therefore becomes the native application environment of the operating system. (That is why we are able to run full J2SE, and demonstrate Java applications running so much faster on smaller information appliances than on other operating systems.)
3) I disagree with Mark's comment that "Corporations like WinCE". Having previously served as an IT strategy consultant ro many Fortune 100 companies, and manager of IT at a Fortune 10 company, I know that Corporations do not care much about WinCE.
- Over 80% of Fortune 1000 has adopted Java on the backend.
- Enterprises want to extend Java to mobile devices!
- Microsoft does not like Java.
- WinCE is not designed to run Java well, and never will!
4) I agree with Mark that SavaJe must support the basic set of applications, such as browser, PIM, email client, office suite, and synch. capabilities in order to meet the needs of the Enterprise. We know what it takes to succeed, and that is precisely what SavaJe will deliver in the product.
I hope my comments help to address the issues you have raised.
You may also read more about SavaJe at: http://java.sun.com/features/2001/06/savaje.html
Please do not hesistate to contact me if you have any other questions.
-- Lead Wey
(SavaJe Technologies) http://www.savaje.com

[Mark Herschberg]

Welcome to JavaRanch Lead. it's nice to have more people working on J2ME technologies join our forum to share their wisdom with us.

Originally posted by Lead Wey:

2) I agree with Mark that writing an OS in Java won't work.... SavaJe's OS has a real-time, multi tasking, multi threaded kernel written in C, not Java! The J2SE VM is then integrated directly to the kernel, and Java therefore becomes the native application environment of the operating system.

Well, clearly you'd have to write the kernel in C. I have no doubt about the ability of SavaJe to create such an OS.

Originally posted by Lead Wey:

3) I disagree with Mark's comment that "Corporations like WinCE". Having previously served as an IT strategy consultant ro many Fortune 100 companies, and manager of IT at a Fortune 10 company, I know that Corporations do not care much about WinCE.
- Over 80% of Fortune 1000 has adopted Java on the backend.
- Enterprises want to extend Java to mobile devices!
- Microsoft does not like Java.
- WinCE is not designed to run Java well, and never will!

When I learned of your company at JavaOne, I pushed our guys to look into partnering with you. We first went and talked to our backers and clients, to see what they thought. They were very against it, for the following reasons.
You are correct that enterprises like Java on the backend. Th like it because Java has lower development and mntanence costs than C/C++. Any minor performence drawbacks are easily compensated for with extra hardware, cheap compared to development costs.
Why doesn't Java dominate on the desktop, too? Well, let's ignore performence (which, while only slightly slower on desktops for many apps, is significantly slower on PDAs for the reason you mentioned). How come StarOffice hasn't taken over? Because people know MS Office. They have it, they like it, they are used to it. Why change? StarOffice has to offer something more in order to get comapnies to up end their IT infrastructure.
Looking at handhelds, we have companies telling us our product is exactly what they need on mobilde devices. we targeted enterprise applications and aoffered full security (authentication and encryption), as well as remote data synchronization. All their business apps will run on our platform. What was the one thing they didn't like? Our application takes up the whole screen, no surprie there. what this means is that it's hard for them to pop over to check their email through PocketOutlook. True, it's only a few clicks, and yet that's considered a lot of effort. And of course, if you're a mobiel user, getting obile emails is critical. Maybe it won't be with our software or with someone else's, but right now that's the mindset, because that's what we're used to. What did we do? We had to build in a bridge, so our application platform could work with PocketOutlook. Now you can check your email from within our application platofrm and our customers are happy. The moral of the story, customers like the programs which work with their current IT infrastructure.
In order to convince an IT guy to wipe WinCE from a device, you need to promise at least as much as what WinCE and it's apps offer, along with a compelling reason why they should choose a new OS (like a strong need for some program which is only viable under a Java OS).
I do not doubt the capabilities SavaJe's OS offers, or it's technical merrits. I just think it's going to be a hard sell. Again, it's not a like of WinCE, but rather that MS is the "safe bet"; moreover it's the default choice. When you make a product in which MS has an established name, OS, office suite, email, web server, etc, you're not competing on an equal level with them. hey're got the ball, you need to steal it.
--Mark
PS Ersin, I share your hopes for Java on PDAs.