Opinons on Mobile Devices

[Pallidrone]

Ok here is the scoop.

I am in the process of trying to program a mobile computer. This device is similar to what you might see a worker in like a supermarket lug around to check prices/stock. I have decided to take my current program and move them into this realm.

The first time this was attempted, it was programmed in EVB with a mobile access backend and it really was a complete disaster. So with that said I am looking at a couple things -

1) All these units use Windows CE 5.0 or above. That means no Java, no flash, no android or anything else useful.
2) They all have volatile memory, so unless there is a SD card placed in them, anything that is loaded will go away if they let the battery die in it.
3) They do have WiFi and Bluetooth
4) They all come with an embedded SQL Express version

With that said I am trying to figure out the best way to go. I could develop a .Net app and place it on the unit, but I am afraid these idiots will let the battery die and I will be on the phone ever other day trying to get it reloaded. The other way is to develop a mobile site, that is limited to one address and loading Apache on the machine, but I have very limited experience with HTML or any other web tool. I still would have no clue on how to connect to my database and also to limit the size of the site to fit on the screen. Plus I am worried about security.

What do you guys think? If you were handed a project like this, would you develop a stand alone app the required downloads to the unit and the possibility of reloaded it, or try to develop an internal website that would provide real time data entry/retrieval, but also has more security risks?

If there is something else that you can think off that I am missing, please let me know.

[Unforgiven]

I think having a central server and more or less using the devices as dummy terminals that connect to a central server. That way you have low overhead and all of the heavy lifting for the device is on the backend server.

BTW, do you have a patent on your ideas? Who is your market/audience? More details would be helpful.

[BuddhaMan]

One thing to possible look at doing is writing an app on the central server that exposes a "web service" to the devices through the web server. The devices interact with the server via this web-based method of communication.

https://secure.wikimedia.org/wikiped...ki/Web_service

.NET Web Service example: http://www.15seconds.com/issue/010430.htm

SOAP intro: http://www.w3schools.com/soap/soap_intro.asp

http://www.w3schools.com/ngws/ngws_webservices.asp

Example of SOAP being used to request the stock price for an item: http://www.w3schools.com/soap/soap_example.asp

Another .NET Web Service example: http://articles.sitepoint.com/articl...rvices-5-steps

http://social.msdn.microsoft.com/Sea...nement=26&ac=2

http://msdn.microsoft.com/en-us/library/ms950421.aspx

http://msdn.microsoft.com/en-us/library/aa480728.aspx

[=DBA=Ronin]

I'm with rarry on that one. Centralize it if you can. Our old NT apps were all localized and such a PITA for end users and admins whenever updates needed to be applied or problems occurred. Our .NET app all runs on centralized servers and all users need is a web browser to access it.

[Unforgiven]

Web based is the way of the future. Apps are merely a shortcut to a web based product in my current job. It allows the flexibility to make all the back end changes you want as the piece doing the actual interaction remains constant. As long as you have connectivity, you have a product that is low maintenance.

[BuddhaMan]

I need more info as well. You mention an SD card...is there actually one installed & available for use? That's unclear.

You also say load Apache "on the machine". On the handheld unit or a separate server? There's no way I'd want Apache running on the mobile devices, so you must mean to load Apache on a central main server and use WinCE's built in browser. (It does have one right?)

If WinCE has a built-in browser, does it allow HTTPS connections over SSL or TLS (encrypted)?

What levels of encryption can the Wi-Fi client on the device do? WPA2 + AES? WPA + TKIP?

Can you do RADIUS authentication on the device (WPA or WPA2 in Enterprise mode) etc?

Point being, if you can encrypt all communication via the Wi-Fi using Enterprise mode, you get encryption + authentication so you know that device is authorized to access the server. If you have that level of security to the server, encrypting web traffic isn't as important (although you could go with the "double condom approach" and do both). If you can only do Wi-Fi with no encryption, you'll HAVE to use HTTPS for a browser or roll encryption into your app. That means you'll have to encrypt and decrypt at both ends.
That sounds like a LOT of asspain.

Constant Wi-Fi use will drain the battery quicker, but if all the data is on the server, that will be your trade-off for doing that. (plus you care about data integrity, the user can worry about battery life)

All good project planning should take idiots into account. Glad to see you did that.

[Pallidrone]

Yeah I was a little vague in the description.

As you know I still work for a company that develops POS systems for the liquor industry. Right now our current product was written in Foxpro. We are in the process of moving that over to Postgres and .Net (still annoyed by that, but it is what it is - I wanted to go Java).

I have a few mobile computers/guns that I can use in conjunction with the program. Here is a sample of one of the guns -

http://www.honeywellaidc.com/Site.as...e/?product=319

I want to develop an app that talks with our back end. Right now I will develop it to talk to our Foxpro database, and then when the new system comes out some time next year, switch over the back end to Post.

I want to start with the basics and if I can get that part done, then the rest should flow easy. So what it would do is load up, the customer would scan a UPC and it would show the product description, price, and current stock levels.

I could go with making an app and placing it on the unit, but that would require the person to upload and download the data from the gun. While this is probably the easiest way to go, like I mentioned before, it has volatile memory in the unit which would blow out the programming and data if the battery dies. It has an expansion slot for a SD card so the programming does not get blown out of there, but it still would require a reload to re-establish the links to the program if it does die.

I was thinking about just running a web app on their server that directly links to the data. That way as long as they are connected to a wireless connection within the store, it should always run as well as give real time statistics. However, since this is Windows CE and it has very limited options, I cant run a Java app from the server or anything of the sort. I would need to go strictly with HTML and PHP. I looked into ISAPI as well, because it would be very easy for me to turn around a form and connection, however it seems like Apache doesnt play well with it. Plus I am not very strong in HTML/PHP design and programming. For the most part it looks easy, but we all know what looks easy usually is not.

On top of that I am concerned about security with a wireless network. I dont know if you guys know much about the new credit card standards set forth by Visa and other CC companies, but it really restricts what you can do on a network. On top of that, I want to make sure that I pigeon hole a specific device to the system, so that way I know what is connecting and that a random rogue computer on the network cant just log in.

Also I am not worried about speed or optimizing the back end. There will probably be a max of 5 users connected at one time, so speed should not be an issue.

That is where I am getting stuck. I know the database back end through and through and can whip up something very quickly, but before I start putting a bunch of time and resources into it, I dont want to get to far without having as much information as possible.

[Unforgiven]

Dude Foxpro? Really? We were talking about this almost a decade ago LOL. Foxpro is dead, MySQL with a custom API for purchases to satisfy the credit card vendors need for security is the way to go.

[Pallidrone]

Unfortunately it is what it is.

It is sad because Foxpro was actually a fantastic product. It reminds me of the Amiga, a fantastic product that never got it's due.

Remember it is just the back end which means as far as connectivity I can go any which way I want to. Would I love to go MySQL, Post or almost anything else? Of course. But I have to satisfy what we currently have out there.

[Unforgiven]

FoxPro 7 and 9 were really great products. Much better than the alternative of Access at the time.

[Pallidrone]

So I made a decision.

I loaded up Apache on a machine and did some PHP code on the backend. It was the first time I have ever screwed with PHP,HTML,CSS and my god it is a lot easier then I ever thought. I was able to put a quick 3 page thing together in about 4 hours that got real time information from the database.

I think I am going to continue with this method.

[BuddhaMan]

Yep, HTML and CSS are pretty easy to get going with. PHP is slightly harder, but it's a scripting language, so it's not THAT hard.