Wanted: naive C# developer

I just realised that a potential goldmine application for Windows Phone is possible, and it involves porting an Apache 2.0-licensed C++ library over to C#. The application is a sure-fire seller.

The library itself is about 530 C/C++ files (though most of them are boilerplate codes and other library functions already available in the BCL). The vast majority of the files are under 8KB in size, though there are about 5 "monstrous" files sized between 90 and 170KB in length.

If anyone is willing to assist me in porting this library over let me know. I can pay for this, or you can take less money from me and instead get a bigger share of the massive profits we'll be reaping

Also, you'll need to be in possession of a Windows Phone 7 device.

Any takers? I'll give out details privately, I don't want to reveal too much on the public web

UPDATE: I've just been informed that the upcoming Mango update for WP7 includes functionality that pretty much does what I'm after.

Oh well.

I wonder if you can patent your business model? Take Apache 2.0-licensed C++ libraries, port them to some other language, and then sell them?

-Josh

20 minutes ago, JoshRoss wrote

I wonder if you can patent your business model? Take Apache 2.0-licensed C++ libraries, port them to some other language, and then sell them?

-Josh

I wanted to make an app like Word Lens but for Windows Phone 7, it would revolve around the Tesseract OCR library. The C# port is necessary because WP7 doesn't allow native code to run. The existing CLR wrapper on Tesseract is CLI++ and I don't know if that's allowed to run on WP7.

Slightly OT:  I've played around with Tesseract and found it to be highly inaccurate relative to the commercial libraries that are available.  Have you been successful in training it so that it is much more accurate?  Just wondering if you are confident that you could build a product around Tesseract that would compare favorably to the offerings that are already in the marketplace on a mobile platform.

@W3bbo:In my experience, unless you have a team of specialists, small software projects are vulnerable to being copied, oftentimes better than what people originally create. Most software I have worked on has patents so we know we have a product that is unique and it usually is tied to some hardware

I think there will be people that make a fast buck making phone applications, but the minute people know you are successful, your days are numbered for the most part.

But don't give up looking for ideas.  Any fool can learn to code and code well (I'm proof).  Coming up with worthwhile and marketable ideas is hard, IMO.

People don't pay for software ideas anymore. They pay for an audience.

http://img197.imageshack.us/img197/7004/awsbill.jpg

I have ramped up millions of impressions per month and the project is gaining hockey stick growth.

Once you have an audience, you can sell them anything you want, including total crap.

The audience is worth money, not the software. Facebook is an extreme case of this. The facebook software is largely worthless and easy to clone other than it's scale with its new data center. But the audience can not be cloned, nor can it be easily taken away.

Had Mark Zuckerberg been a kid from a community college in New Jersey and not at Harvard the very same software would have failed miserably. It's not about specs or technical achievements for sure. You can make the best, and most useful software in the world and it will not grow virally and nobody will buy it.

@beerinbelgium:

You seem desperate to want us to think you're a great developer when you post links like this. Did your mother not hug you enough while growing up?

@cbae:

I don't want you to think anything. I just wanted to let W3bbo know he's wasting his time trying to set up a "software business". People in the industry treat these mobile app devs and software companies like a novelty commodity.

 
The best possible outcome is that you get acquired at bottom dollar and thrown to the curb. If you're big enough your team will be acquired and will end up working on software they never wanted to and will end up quitting.

There are no good outcomes from doing something like this. Apple, Google and MSFT rely on people's stupidity to drive their communities and stores.

You're way better off not participating.

The 2nd point I was making is that if you spend your time building an audience instead of software, you have a product that a company like Microsoft can not clone and superset no matter what they do. They have to acquire you if they want the audience, and they have to pay top dollar.

I've spoken to a friend of mine who is willing to work on porting Tesseract over to pure C# in exchange for peanuts, we'll then split the proceeds from any ensuing WP7 and desktop applications.

A translation app is still useful for desktop Windows, especially laptops as they all have webcams nowadays (and the upcoming W8 tablets); but we have thought up other innovative uses for live OCR on WP7. We'll keep you guys informed.

@W3bbo:

Doesn't mango do OCR natively?

http://www.youtube.com/watch?v=Qq0yVxoVY1w#t=104s

http://www.youtube.com/watch?v=WERWFXQqgZo#t=97s

Yes, but I'm, not aware of any OCR API being made available to developers. I understand due to performance constraints it's all done natively.

15 hours ago, beerinbelgi​um wrote

People don't pay for software ideas anymore. They pay for an audience.

...

Once you have an audience, you can sell them anything you want, including total crap.

This is why you'll always fail.  You contribute nothing useful and aspire to do so for a much larger audience.

The best possible outcome is that you get acquired at bottom dollar and thrown to the curb. If you're big enough your team will be acquired and will end up working on software they never wanted to and will end up quitting.

That's an awfully comfy curb populated by billionaires and household names like Bill Gates, Steve Jobs, Zuckerberg, Eric Schmidt and so on who all started out with a product and then got their audience, rather than starting with an audience and capital and thinking of a product later.

Well looks like we got well off the rails here.  Original topic was talking about converting Tesseract to C#....so many problems here:

1)  The codebase for Tesseract is huge.  Even if you could automate most of the conversion it would take you a very long time to convert the code.

2)  Converting Tesseract won't be like converting some string conversion code; you are going to need a deep understanding of the way the code works.  What happens if something is broken after your conversion?  You won't be able to just say "oh, we forgot to initialize this variable properly".  You'll need to know what the code does and how it does it to convert properly and fix any problems.

3)  It's written in C/C++ for a reason:  performance.  OCR is a hideously brutal operation on a CPU.  Converting to C# will cripple it out of the gates, then throw in the fact that you are running it on a slow mobile platform(slow compared to a desktop PC) and you are looking at a performance nightmare.

4)  Tesseract is a pretty poor OCR engine these days compared to commercial offerings.  Even if you get it ported and you can get it to somehow perform reasonably well you are still looking at a dog.  As Beavis and Butthead once said "You can't shine a turd".

What Word Lens is doing is something special if the demonstrations are to be believed.  But it does have that feel like many of the "compression" vendors in the late 90s where they claimed you would be able to download full length movie in 5 seconds with their new "compression technology".  I hope it's real as it opens up a whole host of new types of applications.

...

2 hours ago, beerinbelgi​um wrote

...

Is that your neuroimage?