The left picture is a simple listview. The right image shows what we are planning to do in this tutorial.
Simple listview sample code here.

For various purposes, we might want to enrich the look and feel of a listview, to add pictures or other controls, or to better organize the items. There are a few samples available on the Internet, but they all require a static definition of an item, using the row.xml resource. I'm showing a different approach: build the items using only Java code, this is what is called a Dynamic Listview custom adapter, so no XMLs required here.

The idea is to use the standard Listview control, but build a different adapter to define the items.
First, let's focus on the items. Let's say we want each item to contain data regarding a network adapter. This means we'll record the name, the address , a type (maybe for wired/ wireless) . To define these details a new class is needed: Device.java

 
public class Device {
 
	    private String m_szDeviceName;
	    private String m_szDeviceAddress;
	    private int m_nDeviceType;
	    private int m_nDeviceStatus;
	    private int m_nDeviceID;
 
	    public Device( String deviceName, String deviceAddress, int deviceType, int deviceStatus, int deviceID ) {
	        this.m_szDeviceName = deviceName;
	        this.m_szDeviceAddress = deviceAddress;
	        this.m_nDeviceType = deviceType;
	        this.m_nDeviceStatus = deviceStatus;
	        this.m_nDeviceID = deviceID;
	      }
 
	    public String getDeviceName() { return m_szDeviceName; }
	    public void setDeviceName(String deviceName) { this.m_szDeviceName = deviceName;}
 
	    public String getDeviceAddress() {return m_szDeviceAddress;}
	    public void setDeviceAddress(String deviceAddress) {this.m_szDeviceAddress = deviceAddress;}
 
	    public int getDeviceType() { return m_nDeviceType; }
	    public void setDeviceType(int deviceType) { this.m_nDeviceType = deviceType;}
 
	    public int getDeviceStatus() { return m_nDeviceStatus; }
	    public void setDeviceStatus(int deviceStatus) { this.m_nDeviceStatus = deviceStatus;}
 
	    public int getDeviceID() { return m_nDeviceID; }
	    public void setDeviceID(int deviceID) { this.m_nDeviceID = deviceID;}
}
 

While Device.java defines the DATA for each listview item, we need another class to define the View of each item, meaning that we need to construct the items using standard interface components (imageviews, textviews, etc). For this, we'll be using another class, CustomAdapterView:

 
class CustomAdapterView extends LinearLayout {
	public CustomAdapterView(Context context, Device device)
	{
		super( context );
		/*setOnClickListener((OnClickListener) context);
		setClickable(true);
		setFocusable(false);*/
		setId(device.getDeviceID());
 
		//container is a horizontal layer
		setOrientation(LinearLayout.HORIZONTAL);
		setPadding(0, 6, 0, 6);
 
		//image:params
		LinearLayout.LayoutParams Params = new LinearLayout.LayoutParams(LayoutParams.WRAP_CONTENT, LayoutParams.WRAP_CONTENT);
		Params.setMargins(6, 0, 6, 0);
		//image:itself
		ImageView ivLogo = new ImageView(context);
		// load image
		if (device.getDeviceType() == 0)
			ivLogo.setImageDrawable(context.getResources().getDrawable(R.drawable.logo1));
		else if (device.getDeviceType() == 1)
			ivLogo.setImageDrawable(context.getResources().getDrawable(R.drawable.logo2));
		//image:add
		addView(ivLogo, Params);
 
		//vertical layer for text
		Params = new LinearLayout.LayoutParams(LayoutParams.FILL_PARENT, LayoutParams.WRAP_CONTENT);
		LinearLayout PanelV = new LinearLayout(context);
		PanelV.setOrientation(LinearLayout.VERTICAL);
		PanelV.setGravity(Gravity.BOTTOM);
 
		TextView textName = new TextView( context );
		textName.setTextSize(16);
		textName.setTypeface(Typeface.DEFAULT, Typeface.BOLD);
		textName.setText( device.getDeviceName());
		PanelV.addView(textName);       
 
		TextView textAddress = new TextView( context );
		textAddress.setTextSize(16);
		textAddress.setText( device.getDeviceAddress());
		PanelV.addView(textAddress);    
 
		addView(PanelV, Params);
	}
}
 

As you can see, this class builds a nice item made out of an imageview, and two lines of text:

Now we only need to build the Custom Adapter class, that will use the CustomAdapterView private class and implement the getView method:

 
public class CustomAdapter extends BaseAdapter /*implements OnClickListener*/ {
 
	/*private class OnItemClickListener implements OnClickListener{
	    private int mPosition;
	    OnItemClickListener(int position){
	            mPosition = position;
	    }
	    public void onClick(View arg0) {
	            Log.v("ddd", "onItemClick at position" + mPosition);
	    }
	}*/
 
	public static final String LOG_TAG = "BI::CA";
    private Context context;
    private List<Device> deviceList;
 
    public CustomAdapter(Context context, List<Device> deviceList ) {
        this.context = context;
        this.deviceList = deviceList;
    }
 
    public int getCount() {
        return deviceList.size();
    }
 
    public Object getItem(int position) {
        return deviceList.get(position);
    }
 
    public long getItemId(int position) {
        return position;
    }
 
    public View getView(int position, View convertView, ViewGroup parent)
    {
        Device device = deviceList.get(position);
        View v = new CustomAdapterView(this.context, device );
 
        //v.setBackgroundColor((position % 2) == 1 ? Color.rgb(50,50,50) : Color.BLACK);
 
        /*v.setOnClickListener(new OnItemClickListener(position));*/
        return v;
    }
 
    /*public void onClick(View v) {
            Log.v(LOG_TAG, "Row button clicked");
    }*/
 
}
 

Using these resources it is trivial to build and use a listview. This code in our main example class:

 
ListView ls2 = new ListView(context);
// clear previous results in the LV
ls2.setAdapter(null);
// populate
ArrayList<Device> m_Devices = new ArrayList<Device>();
Device device;
for (int i=0;i<10;i++) {
   device = new Device("Network Device "+i,"13:B4:5C:0D:AE:67", i%2,0, 100 + i);
   m_Devices.add(device);
}
CustomAdapter lvAdapter =  new CustomAdapter(context, m_Devices);
ls2.setAdapter(lvAdapter);
 

Creates a simple but nice Listview:

Get click event on Custom Adapter Listview

Ok, there are a few ways of handling this.
Method 1
In CustomAdapterView constructor, you make each item clickable, set an id (that you can use later to identify the item), and set a click listener (in the main class, using the context):

 
public CustomAdapterView(Context context, Device device)
	{
		super( context );
		setOnClickListener((OnClickListener) context);
		setClickable(true);
		setFocusable(false);
		setId(device.getDeviceID());
		...
 

Then in the main example class you catch the click event:

 
public void onClick(View v) {
	int cmdId = v.getId();
	if (cmdId >= 100) { //first listview id
		int nIndex = cmdId - idLV1FirstItem;
		Log.i(LOG_TAG, "LV2 click "+nIndex );
		}
	}
 

This method works good, but it has a major disadvantage: because you've used "setClickable", the items will not show any change when clicked: they won't be drawn as "clicked". But you receive the click event.
Another disadvantage is that the listview will not react to key presses (enter) on items.

Method 2
The correct way of doing this, is to use the setOnItemClickListener method, without setClickable.

 
        ...
        ls2.setAdapter(lvAdapter);
        ls2.setOnItemClickListener(new OnItemClickListener()
        {
        	public void onItemClick(AdapterView<?> arg0, View arg1,int arg2, long arg3)
        	{
        		Toast.makeText(getBaseContext(), "You clicked on "+arg2, Toast.LENGTH_LONG).show();
    		}
    	});
        ...
 

This way you have the animation when an item is clicked, and you also get the click event. And it works with key presses as well. All in one, this is the method to use for getting click events on custom adapter listviews.

Change row colors

You can have one row of a color, and the next in another color to read the content easier.
In CustomAdapter, method getView, simply use v.setBackgroundColor based on position:

 
        v.setBackgroundColor((position % 2) == 1 ? Color.GRAY : Color.BLACK);
 

The result looks like this:

Here is the code for this tutorial:
Custom Adapter Listview Source code, 57KB

Hope this helps,

Since most of the NDK tools are coming from the Linux environment, that OS was my first approach in building native C applications for Android. See this tutorial for a start.

However, I also needed to be able to compile JNI Android apps using the Windows environment. It proved an easy task, but there were some issues, so I decided to write this guide.

Install Android NDK on Windows and use it to compile a JAVA/JNI application

You will need an internet connection to install Cygwin and the Android SDK automatically from the Internet.

Download link.
Current version of the file is: android-ndk-r4b-windows.zip 45MB

Download link. (Direct Link)
Select the mirror you need to download from then follow the installation guide.
On package selection, Open Devel category and check "make":

I've installed it to C:\cygwin

Download link.
Current version is android-sdk_r06-windows.zip 23MB

4.1) Unzip the zip content to a location on your harddisk. I choose:
D:\work_code\android\android-sdk-windows\
You will find SDK Setup.exe . Run it. I get a warning to install JAVA JDK
4.2) Download Java JDK
Download link.
Select platform "Windows". Current file is: jdk-6u21-windows-i586.exe (Direct link)
4.3) Install the Java JDK: simply run the exe . It takes a few minutes to complete.
4.4) Re-run the SDK Setup.exe from the Android SDK.
This opens the Android SDK and AVD Manager. If you get a HTTPS Error:

go to Settings and check the "Force https:// ..." checkbox:

4.5) Go to Available Packages, select all, and press Install Selected:

The installer will download the packages from the internet, and might prompt you to accept terms, which you must do if you want to continue with the installation. This step takes a few minutes.

5.1) Unzip the NDK zip file. My choice is: D:\work_code\android\android-ndk-r4b\
5.2) Configure Cygwin parameters
Go to the Cygwin folder and open Cygwin.bat. It contains something like this:

 
@echo off
 
C:
chdir C:\cygwin\bin
 
bash --login -i
 

Using any text editor change it to:

 
@echo off
 
set IS_UNIX=
set DEV_ROOT=d:/work_code/android/BlueInput
set JAVA_HOME=c:/Progra~1/Java/jdk1.6.0_21
set CLASSPATH=d:/work_code/android/BlueInput/obj
set PATH=D:/work_code/android/android-sdk-windows/tools;D:/work_code/android/android-ndk-r4b
set ANDROID_NDK_ROOT=/cygdrive/d/work_code/android-ndk-r4b
set NDK_PROJECT_PATH=/cygdrive/d/work_code/android/BlueInput
 
C:
chdir C:\cygwin\bin
 
bash --login -i
 

The path in the config, d:/work_code/android/BlueInput, is something I'm currently working on. Change to your Own project name. The project must be organized as follows:
A) The JNI part (the .c and .h and Android.mk files) in:
d:/work_code/android/BlueInput/jni
B) The JAVA part (assets,bin,gen,jni,libs,obj,res,src with the .java files) in:
d:/work_code/android/BlueInput
For these paths always use forward slashes.
5.3) Open Cygwin and browse to the NDK folder, using cd [enter] :

cd d:/work_code/android-ndk-r4b
5.4) Type ./ndk-build to compile the JNI project

The C files have been successfully built. Now using Eclipse, you can build the Java android project, and it will include the JNI part automatically. Easy.

To be continued.

Prototype
BOOL FindAttribute (UINT16 attr_id, SDP_DISC_ATTTR_VAL *p_val);
Parameters
attr_id The attribute ID being searched for.
p_val The location where the function should store the attribute.
Returns TRUE if the attribute was found. FALSE otherwise.

The structure SDP_DISC_ATTTR_VAL is defined in BtIfDefinitions.h. as following:

 
#define MAX_SEQ_ENTRIES     20
#define MAX_ATTR_LEN        256
 
typedef struct
{
    int     num_elem;
 
    struct
    {
        #define ATTR_TYPE_INT      0       // Attribute value is an integer
        #define ATTR_TYPE_TWO_COMP 1       // Attribute value is an 2's complement integer
        #define ATTR_TYPE_UUID     2       // Attribute value is a UUID
        #define ATTR_TYPE_BOOL     3       // Attribute value is a boolean
        #define ATTR_TYPE_ARRAY    4       // Attribute value is an array of bytes
 
        int     type;
        int     len;                        // Length of the attribute
        BOOL    start_of_seq;               // TRUE for each start of sequence
        union
        {
            unsigned char  u8;                      // 8-bit integer
            unsigned short u16;                     // 16-bit integer
            unsigned long  u32;                     // 32-bit integer
            BOOL           b;                       // Boolean
            unsigned char  array[MAX_ATTR_LEN];     // Variable length array
        } val;
 
    } elem [MAX_SEQ_ENTRIES];
 
} SDP_DISC_ATTTR_VAL;
 

As you can see the "array" member of the union can only take 256bytes. The SDK has a note:

the structure allows for an attribute that is a sequence of up to MAX_SEQ_ENTRIES (=20), each with a value array up to MAX_ATTR_LEN (=256) bytes long. Nested sequences up to 5 levels deep are supported.

Assuming you need to retrieve an attribute that is bigger then 256 bytes, the first 256 bytes will be stored in elem[i].var.array and the following bytes in &elem[i+1], &elem[i+2], and so on.
Please note that the first set of 256 bytes is stored in elem[i].var.array, but for the rest, the whole elem structure is used, meaning you will have 268bytes (type, len and start_of_seq are included).

A quick calculation shows you can receive an attribute consisting of an array of maximum: 256+19*268bytes.

Good thing there is a way of getting more then 256 bytes!

 
BYTE *pRDData = (BYTE *)malloc(MAX_RD_SIZE * sizeof(BYTE));
memset(pRDData,0,MAX_RD_SIZE * sizeof(BYTE));
nRDSize = p_val.elem[i].len;
if (nRDSize < MAX_ATTR_LEN)
{
	memcpy(pRDData, p_val.elem[i].val.array, nRDSize);
 
}
else
{
	BYTE nBlocks = 0;
	memcpy(pRDData, p_val.elem[i].val.array, MAX_ATTR_LEN);
	nBlocks++; //we've just copied the first block
	for (int j=0;j<nRDSize / MAX_ATTR_LEN;j++)
	{
		memcpy(pRDData + nBlocks*MAX_ATTR_LEN, &p_val.elem[i+nBlocks], MAX_ATTR_LEN);
		nBlocks++; //another block has been copied
	}
	memcpy(pRDData + nBlocks*MAX_ATTR_LEN, &p_val.elem[i+nBlocks], nRDSize % MAX_ATTR_LEN);
}
 

It is bigger then my Radex 1706 or the Terra-P MKS 05, but smaller then the CDV 700 or the CDV 717.

The dosimeter is intended for use at temperatures from -10°C to 40 °C, relative humidity up to 80%, normal atmospheric pressure (650-800 mm Hg). Dampness will interfere with its high voltage generating circuit. Keep the radiation dosimeter clean and protect it from impact, dampness, and harsh environments.
One set of batteries will operate the radiation dosimeter continuously for at least 100 hours under natural background radiation conditions.
It came with two pages of instructions, you can see them here: DRSB instructions.

My uranium glass marbles, give higher-than-background clicks, and the green led lights frequently:

The CHECK Source DU on the CDV 700 casing, brings the RED led on:

As I said, this is a very simple Geiger Counter device. You can see the schematics here:

I plan to add a Microcontroller and an LCD to "enhance" this dosimeter. But more on that later!

Currently I have WAMP installed. Extremely lightweight application, this package brings everything you together: apache, php, mysql and a nice control panel interface.

I was surprised to see the server will not start. Also the status ICON was showing that the server is off, despite my attempts of restarting it. All I was getting in my browser was an empty page. But the browser seemed to be connecting to something (no connection error):

Intrigued since I was aware of no "server" running on my machine, I used netstat to see a list of open ports and the PIDs of the processes using them:

 
netstat -anop TCP
 

Here are the results:

As you can see Skype was holding the port 80 open! Why? How? I never acknowledged this when I installed Skype! So I quit Skype and probably will be uninstalling it too. For now, no more Skype Autostart, that's for sure.

Killing Skype solves the problem immediately:

You draw the conclusion. It would be a good idea to think twice when installing "bells and whistles" software like Skype, Messengers, and others.

The specs are:
Mini ITX D201GLY2 motherboard
1.3GHz Celeron CPU
512MB RAM
IDE2CF and CF2SDCard adaptors for a 2GB SDCard based DOC (disk on chip)
M2-ATX DC-DC PC power supply 160W, so I can power the board using 12-24V DC.

The idea was to build a case and add a big 320GB harddrive. See what I managed to come up with:

The server works good, for various purposes, from file storage to permanent internet clients, and the connectivity is easily assured with a USB-WLAN card. To administrate it I use a VNC software.

This effect can be used to generate electricity, to measure temperature, to cool objects, or to heat them or cook them. Because the direction of heating and cooling is determined by the sign of the applied voltage, thermoelectric devices make very convenient temperature controllers." Source.

I recently got a 150W peltier element on Ebay:

The parameters are:
- Operating voltage: 0 - 15V DC
- Current: 0 - 10 Amps
- Temperature interval: -60°C to 180°C

This device works like a heat pump: it has two sides, the cold and the hot. Between these sides there are several thermo couples. All those thermo couples are interconnected and to the two wires that comes out. If voltage is applied, the heat is transferred from the cold side to the hot one.

The device does not "generate" heat nor cold, it just transfers heat. It is also called a thermo-electric cooler or TEC for short.

Because of the internal proximity, a part of the pumped heat can radiate back to the cold side. So the temperature difference is limited. Sometimes it can reach a temperature difference of 80 degrees Celsius.

As indicated in the datasheet, when applying DC 12V, one side becomes cold and one very hot, but the heat radiates and in a few moments the side that should be cold receives heat as well.
To maintain the cold side, the heat must be evacuated from the TEC hot side.
Theoretically with an ideal heat-evacuation system, I should be able to bring the cold side down to -60°C (room temperature ~ 22°C - 80°C ~= -60°C).

So for a start, I've added a heatsink:

I've also used Thermal Grease for a better thermal contact of the TEC and the heatsink.

To improve the results even further, the heatsink has been merged in a volume of cold water. By doing so , the lowest temperature I got was -3.0°C, and I was able to instantly freeze a few drops of water placed directly on the cold side of the TEC. See this video, I've added a thermometer to observe the temperature evolution:

The BtSdkCE50 is compiled for THUMB while BtSdkCE30 is for ARM.

You'll need BtSdkCE50 if you're planning to use:
GetBTWCEVersionInfo, AddServiceName, SetSecurityLevel and other APIs.

So because of the need to use these functions it is almost compulsory to link your Bluetooth application with parameter machine:THUMB . Unless you can go with what BtSdkCE30 has to offer, only.

An even bigger issue is with a set of functions, out of which I've identified: IsDeviceConnectable and IsStackServerUp. Simply calling any of these , results in "Not a valid Pocket PC application" critical error message, before the application has any chance to start:

It makes no difference if only BtSdkCE30 or BtSdkCE50 or both are included. For some reason the resulting linked code is invalid. This behavior has been tested on several Windows Mobile devices, ranging from WM4.0 to WM6.5.

The code that results in this error is plain simple:

Not sure how to fix this, and unfortunately despite the high popularity of the Broadcom stack among newer devices, there is little documentation available to developers.

"CD-V700 radiation survey meter is a sensitive low-range instrument that can be used to measure gamma radiation and detect beta radiation. It is recommended for (1) monitoring of personnel, food and water when used in a shielded facility or an area of low radiation background, and (2) follow-up monitoring or areas for human habitation and food production. The instrument is also used in training programs where low radiation dose rates are encountered. "

(From Radiological Defense Planning & Operations Guide SM-11.23.2 Revised March 1967)

The case was a bit scratched, but checking the Counter with the left side Radioactive source, showed 0.22mR/h on the 10X Scale meaning 2.2mR/h, so good to see it is still calibrated and working correctly.

The emissions of the Check Source were double checked with a modern Radex 1706 dosimeter, details that can be seen in the picture above 2.46uSv = 0.246mR (because 1mRem = 10uSv).

Good thing to have a Radioactive Check Source, however I prefer to keep dangerous things away, so my idea was to cover the Check Source area with a Lead Shield. An online article, indicated the check source for this model (Victoreen CDV700 6B) is Depleted Uranium and the emissions are as follows:

Depleted uranium consists primarily of the radionuclides uranium-238 (with a half-life of 4.5 billion years), thorium-234 (with a half-life of 24 days), and protactinium-234m (with a half-life of one minute). There is also a small quantity of uranium-235 present (about 2% of the activity of the U-238), and an even smaller quantity of uranium-234. Like a Ra DEF source, a DU source emits two beta particles for every alpha. The most prominent gamma rays emitted by DU are at 63 keV (thorium-234), 93 keV (thorium-234), 186 keV (uranium-235) and 1001 keV (protactinium-234m).

At 5 cm away, the DU source still produces a 0.20uSV indication on the Radex 1706, that would be 0.1mRem.
Should see if the radiant energy calculations correspond to these numbers.

Shielding the Check Source

I want the Check Source but only for a few tests. So I decided to cover it with a Lead Shield. First, I had to melt a few small pieces of Lead and have everything in a nice plate shape, good to put in front of the DU.
I used a small aluminum foil recipient to host the melted lead (since aluminum has a higher melting point) , my butane torch (that can give > 1000° C) and a screw driver to mix the content:

I placed this Lead Shield over the Check Source area and the radiation was completely blocked.

The Audio Connector Mod

The CDV700 6B has a built in audio connector, but it is old and unusable. So I took it off and simply connected a modern 3.5mm audio jack so I could plug in any headset currently available.

And it works!

The BNC Connector Mod

The Geiger Muller probe is connected to the CDV 700 via a coaxial cable. However this cable is directly connected to the device and cannot be unplugged.
My idea was to add a BNC connector so I could disconnect the probe or connect other probes easy.
After removing the cable, directly soldered to the circuit board of the CDV700, I had to drill a hole and force-insert the BNC connector. On the loose cable I added the corresponding connector as well.
Here's the result:

A quick test shows there are no signal losses and everything runs good as before.

The result

The last picture shows the CDV717 , a Huge dosimeter based on a ionization chamber, that is unfortunately calibrated for enormous amounts of radiation , making it useless for my experiments. Someone said the CDV717 actually measures how dead you already are, a good joke indeed.

The schematics for the CDV700 and a few circuit pics are attached below:

According to this source:

Various ,,6993" Geiger-Müller Tubes were used in the CD V-700 (Anton, Lionel, Victoreen, EON). Only the Chatham CD V-700 model 3 uses an Amperex 85NB / 6980. These GM tubes need 900 Volts to work and they are filled with Halogen gas. Instrument accuracy on any of its three ranges is within ± 10% of the true dose rate from Cobalt 60 gamma radiation. This accuracy is maintained throughout a temperature range of -28° C to +51° C, and at altitudes from sea level to 7600 meters. The CD V-700 used also a headset (4000 Ohm) for audible detection of radiation. The newest models produced were the 6Bs