Not happy with BREW!

[elektrobank]

When I bought my new Motorola T720 from Verizon one of the things that really attracted me to it was when I saw that you could run fairly sophisticated applications on this phone through BREW. Services like Vindigo looked really cool and aren't too expensive, but what I find really irritating about BREW is that it was designed solely around making money for the cell phone companies. The fact that all software has to be obtained through services like Verizon's Get It Now and developers have no way of developing simple applications for their own phones is absolutely insane. I see these phone as just another computer platform and with the Get It Now service, Verizon has basically become the only reseller of software for this platform and independent developers who wish to develop personal applications for their own use or to resell on their own are completely forbidden from doing so. It's not enough that we will be using our air time downloading and in some cases using these application, they want to make a profit off of every application we use on their phones. This would be like if your ISP controlled all the software you used on your PC and you could only run software that was purchased through them and developers couldn't even run their own applications on their computers unless they were sold through the ISP. How insane would that be?

I was hoping that I could develop my own applications, which would allow me to do some very cool things on my phone that were specific to my own needs. I was also hoping that this would allow me to develop and deploy applications within my company that all employees who have one of these phones could use, such as an employee directory search. From what I understand the process for developing my own software involves the following: Getting a Verisign certificate (~ $500), sending my phone in to get the special firmware so I can download applications to the phone, purchasing Visual C++ (~$500 I think) and getting the ARM compiler (~$1500). That's pretty expensive to get started, and even after I have done all that, I still can only develop programs for my own use. If I want to actually distribute them to other people, I have to do so through my cell phone company, that's assuming that it was a commercial application I want to distribute. It's currently impossible for me to resell my own software or distribute it to my friends or within my own company, this is horrible!!

Anyone know what the future plans are for BREW and if they will ever give independent developers the ability to develop and deploy their own applications on their own? My feeling is that unless BREW applications can be downloaded from sources other than the carrier services like Get It Now and developers don't have to spend about $2500 just to get started, it will definitely fail.

Some of the best software is developed through independent developers, BREW is really limiting the variety and amount of software available to customers by excluding them. No platform can do well without good developer backing, every developer I talked to wouldn't even consider developing BREW applications, they think it's just as crazy as I do.

Please correct me if any of the information I have written is incorrect, I am just learning about BREW and these are my opinions so far.
What do you all think?

[d880640]

I see what you are saying about the cost developers must pay. You can add $2500. "more" plus expenses for travel for training.

Read this article it tells us why brew is doing it this way:
http://portals.devx.com/brew/article/6715

I would like to program for brew but I do not know
C++ only vb plus I do not have the money right now.

[charliex]

welcme to the world of embedded development, you are comparing the market of the PC i would imagine with the cellphones..

consider say nintendo or sony playstation development instead where you can pay as much as $25,000 for for the hardware to get started.

I disagree about the best stuff coming from indies, small companys perhaps, but not individuals, thats quite rare.

You are basically looking at a console style market place, where the publishers are in control of the content, marketing etc.

I'd say you probably should have researched your purchase a little better, not go blaming qualcomm/arm etc for that lack on your part.

Do you see an XBox or PS2 as another computer for you to develop free apps for ?

I don't know which developers you are talking too, but upfront costs for systems is part of any business plan. Have you got any idea how much it costs to lay an infrastructure down for a system like brew ? Or to develop software like ARMCC.

[arunbangari]

Originally posted by d880640
I see what you are saying about the cost developers must pay. You can add $2500. "more" plus expenses for travel for training.


I dont think you require special training to learn BREW. Its all about API's and very easy to learn. If you have little knowledge of C/C++, learning should not be a big deal. Also this community is growing larger day by day and u can get timely support from people who are already working in this field.

Regards
Arun Bangari.

[CodeFighter]

In order to distribute BREW applications via the Get It Now service, yes, I agree with your statements. But, while a novice at BREW (but a programmer for 18 years), it looks like you have some options.

First, VC++ can be purchased for as little as $80 or so, for the 'Standard' version. There are even Books that come with VC++ on a CD for $40 you can purchase. Even less if you are a student with a Student ID card. My guess however, is that you only need to create the DLL which even free compilers like Borlands C++ Builder, or the BloodShed compiler. My guess is that VC++ is the single *supported* compiler, but other free compilers can be used for creation with a bit more legwork.

Second, all those fees really involve *distributing* your application. If you only want to build something for yourself, then I think it is possible. I mention this because the BREW SDK seems to include tools for debugging and testing the application on a handset with the use of a data cable connection.

Obviously, for testing and debugging, though the emulator does most of the work, as the project nears completion programmers want to test on the actual device, without actually publishing the software on Get It Now. I believe the SDK tools provide the ability to download your apps directly to the phone (assuming it supports a PC connection to begin with).

So, if you can download for testing purposes, then you can run the program. And if you can run the program, you can put it on the phone without the major 'distribute' task.

What this form of development certainly does kill though, is the shareware model of the PC. Now, Joe Programmer can't create an app at home in his spare time and distribute it to the 'Net in the hopes they will decide to fork over $25 for his 6 months work.

As always, there are ways around this as well: I would imagine that if you developed an App in your own time using BREW, tested, debugged, and completed it (for a total of maybe $40 for VC++), then you could post a message here or in a newsgroup offering it to a registered BREW partner for distribution under their name. You could then arrange to have them purchase it outright, or have them pay you a certain percentage of the sales profit of the App.

There is almost *always* some way to accomplish what you want. Thats thinking like a programmer.

[Louie]

CodeFighter,

While I agree with some of your comements, you are assuming a lot.

1.) Yes you can get cheap copies of VC, but you need an ARM compiler for the handset - $US1500 minimum (GCC cross compiler can be used, but I haven't got that to work yet)


So, if you can download for testing purposes, then you can run the program. And if you can run the program, you can put it on the phone without the major 'distribute' task.


You can not download to your phone without a test enabled phone, the BREW apploader and a sig file for your device.
These items are only available via the BREW Developer extranet (NOT part of the SDK), which costs $US400 a year. Thats right - develop something for use ONLY on your phone it still costs at least $US400 + complier.

I agree, there is almost always a way to do most things, but in this case Qualcomnm appear to have thought this through quite well.

[warpig]

Quote snipped for brevity....
Originally posted by CodeFighter
...
Second, all those fees really involve *distributing* your application. If you only want to build something for yourself, then I think it is possible. I mention this because the BREW SDK seems to include tools for debugging and testing the application on a handset with the use of a data cable connection.
...
Obviously, for testing and debugging, though the emulator does most of the work, as the project nears completion programmers want to test on the actual device, without actually publishing the software on Get It Now. I believe the SDK tools provide the ability to download your apps directly to the phone (assuming it supports a PC connection to begin with).
...
There is almost *always* some way to accomplish what you want. Thats thinking like a programmer. [/B]

There's a small exception here. None of us here have been able to successfully port an app that ran flawlessly on the emulator to an actual device without a lengthy recoding process. So we're talking about using the ARM compiler, which really has nothing to do with distributing the app but yet costs an amount that single developers rarely have access to.

Contrast that with the J2ME and other models (RIM pager for example) where single developers can reliably develope applications for a fraction of the costs associated with BREW.

Now granted, BREW is nice in that billing and provisioning are an integral part of the system as well as license management but with the exception of the latter, the others mostly benefit the carrier's ability to manage a library of BREW apps while the developer still has to pay for the priviledge of developing apps for BREW.

As a developer, it seems to cost of entry includes:
[list=1]
Purchase of Microsoft VC++.
Signing up for the Developer Extranet.
Purchasing the ARM compiler.
Purchasing a phone.
Getting the phone test enabled.
Passing test certification.
[/list=1]

I think BREW in and of itself is a great idea. I think the API is well thought-out and very thorough and the emulator is a great environment to code against (even if deploying the apps to actual devices can be somewhat nightmarish).

But to compare BREW development with console development is a bit premature. BREW is still in its infancy while there are millions playstations sitting in households guaranteeing to a developer that if his/her app is sound and appealing and distributed properly, then his/her app stands a good chance to generate revenue. Such is not the case with BREW which is still in it's infancy and is better tasked to attracting application developers rather than throwing barriers in their paths.

Until BREW is able to prove itself against competing products and technologies, it is at best a crap shoot although perhaps an intelligent one. Maybe I'm dating myself here but as an owner of an Atari 2600, I clearly recognized that the Mattel Intellivision was a superior product but it failed to sell to the masses. One of the reasons for that is that while the Atari 2600 was an inferior product, it had much better developer support than did the Intellivision.

In a perfect world, the cost of development of a good BREW app would be quickly compensated through the revenue that application generated but ultimately I believe, consumers will flock to devices that offer the greatest array of choice in applications.

If that technology turns out to be J2ME then I, as a developer would much rather develope against J2ME and then only port my application to BREW if the application was highly successful under J2ME.

[arunbangari]

It is obvious that QCOMM will expect to make money out of BREW when it has spent so much effort and money to develope such platform. There is no doubt that BREW is an excellent platform for mobile deveices. It is very easy to learn and developers can develope and publish their application in no matter of time.Ofcourse the cost involved is more.

BREW is a small community compared to J2ME. J2ME can do most of the things that BREW can do but still few things such as Billing system, and thorough testing such as NSTL will do for BREW etc. are missing from J2ME or not robust.Such features not only will attract developres but also carriers. No doubt J2ME is picking up and one day people may opt for J2ME rather than BREW. As the BREW community grows, it is not easy to replace BREW too. Right now BREW and QCOMM requires lot more to do. One best thing would be reducing the cost of development of applications. This will not only attract more developers towards BREW community but also makes BREW more popular..

Thanks
Arun Bangari

[warpig]

I reserve the right to withhold on calling BREW an excellent platform for mobile devices. My experiences is that BREW is an excellent platform for the BREW emulator. Actually deploying it to mobile devices is another story entirely. :D

[drvirens]

well...well...well...i do think that it has been largely insane on the part of qualcomm to put a complete control on the 3rd party applications that would be developed using BREW technology. i guess the model should more have been like Palm Inc guys where they believe in the extensive third-party support and have taken the steps accrodingly. so..a palm developer can write his own app and market it without giving any part of his sales to the creator of technology directly.

also, the tools that i need to develop a palm app is not so expensive matter ( u get code warrior for under $360 something ). the question is - why didnt qualcomm do something like that? like elektrobank, even i have been thinking that all this technology qualcomm has created to make "rich" handset comapnies "richer" with no bothering for developers - at whose cost they really dream to make the tech popular. their is an element of antagonism in that approach, bad sadly, as i see it, its true!

-viren

[charliex]

While i agree ive had trouble with brew emulator vs hardware, i've also equally as much trouble with j2me emulator vs handset, and phone to phone, theres no differences there.

I'd like to see JTAG made more available to us.


as for comparing it to consoles, its perfectly valid, consoles did exactly the same thing when they were introduced. it was even more expensive and difficult to get in, theres no guarantee of sales, no matter how big the market is.

As a 2600 developer, the 2600 did better because it had better games was more available and cheaper iirc. I'm not sure how you know atari gave better tech support than mattel as an owner, perhaps i misunderstood you.

The 2600 was a bitch to code for


Most people wanting free dev tools, are just programmers who bought the phone wanting to tinker with it and put their own apps onto it, theres no real viable business model there. They should have researched a bit more before buying the phone, it tends to be the see it buy it, then discover they cant do free stuff for it. there are phones that allow you to do it for nothing.

you still need test enabled phones to do some java stuff, especially with the brain dead OTA stuff, you still need the phones to test the apps on, if you are relying purely on the java emulators you are asking for trouble, java write once, debug forever

theres also j2me certificates as well..

VC is well worth the cost, which can be had for $99 or less
The armcc is very reasonably priced and if you think your app is good enough to sell, then you can get the costs back.

if youve got a good app go to a publisher, they'll help you out.

consumers go for shiny things when it comes to phones

As for the emulator, what does anyone expect its a win32 app runing inside another win32 app compiled with a different compiler running on a different OS ?

[flarb]

There is sort of a JTAG debugging solution for BREW. It only works on the Sharp Z-800, but Sharp will modify your phone to work with a box from Lauterbach and their TRACE32 software that allows you to step through code line-by-line. Unfortunately, the box costs $5,000 or so--and this solution only works with one phone (Z800).

Because the box is made by a third party, other handset manufacturers could provide support for it, but they don't. If every phone supported it, $5,000 might not be so bad.

[charliex]

interesting, i know the chipset exposes a JTAG conector and the OEMs use it, bu they so far havent come up with the goods, whether or not its DRM paranoia or general hocus pocus red tape i dont know. but it'd be really useful instead of all this heath robinson/mickey mouse emulator rubbish.

[flarb]

Yeah, they had it at BREW 2002. Or at least the Sharp guy had a pamphlet with info and stuff. If you want to blow $5k, email this guy at sharp for the phone mod (takes 1 month they say--and I think it's free):


> Dean Bergstrom mailto:dbergstrom@sharplabs.com
> Sharp Labs of America http://www.sharplabs.com
> 5750 NW Pacific Rim Blvd., Camas, WA 98607 USA
> (360) 817-8547 voice
> (360) 834-8696 fax

And this guy at Lauterbach for the $5k box:

> Jerry Flake
> 13256 SW Hillshire Drive
> Tigard, OR 97223 USA
> (503) 524-2222

Or, check www.lauterbach.com