a=[];until(a.length==5);n=1+rand(56);a<<n if(!a.include?(n));end;puts "your numbers are #{a.sort*','} and Mega Ball #{1+rand(46)}"

I ran into some problems trying to get the updatetool for Glassfish working on my VPS (at Rimuhosting) though. The updatetool is a Java Swing app and I’ve never run any GUI apps on this remote host before. It doesn’t have Gnome or KDE running so VNC was not an option. The trick is mostly just to set up your systems for running remote X11 applications over a SSH tunnel. Here’s how I got it to work…. I should say this server runs Ubuntu Hardy and locally I’m on Mac OS X so the X11 install from the Mac OS installation disks was the easy part.

Here’s what I had to do on the remote host:

open /etc/ssh/sshd_config and make sure you set

  X11Forwarding yes

in /etc/ssh/ssh_config  check that you have the following settings:

  ForwardAgent yes
  ForwardX11 yes

there are a few other dependencies that you may not have yet on a clean Ubuntu install. I was getting very confusing java link errors in my stacktraces until I’d installed all of them. So grab these with:

  sudo apt-get install xauth libxp-java libxtst-dev libxi-dev

at this point, log out and log back in with

  ssh -X  yourname@yourhost

The -X option is important here : it’s telling the server that you’re going to want to run a X11 application over this SSH tunnel.
On my system I got the following message after logging back in:

  /usr/bin/X11/xauth:  creating new authority file /root/.Xauthority

that’s all. I was now able to run the Glassfish updatetool simply by calling

  /var/www/glassfish/updatecenter/bin/updatetool

and the application came up nicely on my Mac desktop. These steps will probably work to run just about any Java Swing application on a remote Ubuntu/Debian server.

I hope this helps some people…

Here’s the basic steps on how to get this Trendnet TV-IP400 / IV-IP400W driver to work in ZoneMinder:

• Make sure you have indeed version 1.23.x. You can see this in the title of the main console window. If you are still using version 1.22.x, either upgrade your ZoneMinder or use my old driver.

• Download the driver file here, and store it in the right directory. For me (having compiled Zoneminder from source on ubuntu) that location is /usr/local/perl/5.8.8/ZoneMinder/control. If that location doesn’t exist on your system, you can do a search for another camera driver such as Visca.pm and store the TVIP400.pm file in the same location.

• Create a new camera for your TV-IP400 if you haven’t already done so. The source tab looks something like this:

Make sure you change the IP address and port number if necessary. You will need to make sure that access control is turned off by the way.

If you had enabled the ZM_OPT_CONTROL setting in the system options screen, you will also see a control tab in the camera config window. This tab should look like this:

Again, make sure you set the correct IP address (the same as what you set in the source tab). You won’t have a TV-IP400 option yet in the Control Type dropdown, so click on ‘edit’ there and let’s create one.

In the Control Capabilities window, click ‘Add New Control’. Then set the settings in the Main and Move tabs as follows:

In the Pan tab, only check the ‘Can Pan’ option and leave the rest blank.
In the Tilt tab, only check the ‘Can Tilt’ option.

All options in the Zoom, Focus, White and Iris tabs can be left blank since none of these features are supported.

The Presets tab should be set as follows:

Though you can’t actually set presets through ZoneMinder, you can move the camera to presets that you’ve configured through the Trendnet web interface. I’m only using 5 presets but I believe you could access all 24 presets that the Trendnet supports.

That’s all! Save the configuration, and give your new controllable camera a spin in ZoneMinder. Let me know how you like it.

Thanks to Armando Oritz for his help in testing this driver. 

I’ve been working with jQuery for over a year now and have always wanted to give something back to the user community by releasing my own plugin. Unfortunately it seemed like every type of generic functionality I might ever need had already been developed by others. Until last night.

I was working on a webapp in which my main browser window had to open a popup window, and the user would then perform an action in this window. Upon completion of that action the main (parent) browser window needed to be notified.
Though parent and child window can easily access eachother’s DOM, a common way to get a value from one window to another is by setting a value on a hidden input field of a form in the other window. This requires you to add form markup to your page, which is clunky. A second issue is that if you have several types of events or data to send to the other window, you’ll need to add an input field for each type. And a third issue I ran into is that ‘change’ events do not seem to fire for an input field if that field was updated by another window. ‘change’ or custom events using jQuery didn’t work either this way. ‘click’ events on links or buttons do work, but they would need to be hidden, otherwise the user could click on them too. I got it all working but this all started to look like an ugly hack, cluttering my pages with messy javascript and markup code.

Long story short, this is where my new skills of creating jQuery plugins came to the rescue. This could all be put into a nice little plugin that would hide all the messy details of adding a form to your page and using hidden form elements to transfer data back and forth between windows. Even better, my new plugin makes it very easy to use multiple callback handlers, so you can define your own events that you can send from one browser window to the other. And events can be sent both ways in almost the same fashion. To my delight, it worked flawlessly in all browsers I’ve tried (FF, safari on Mac, and FF and IE7 on Windows).

Here’s how to use it:

In any page that needs to receive window events, make sure to call the initialization function upon page load:

$.initWindowMsg();

The parent window can open a child window the good old way:

childWin = window.open('child.html', 'mychildwindow');

and then send it an event with data:

$.triggerWindowEvent(childWin, "myevent", "some data");

The child window can register a callback handler for this event as follows:

$.windowMsg("myevent",  function(message) {
    // do  something  with message from parent
});

The child window can send a message back to the parent with:

$.triggerParentEvent("parentevent", "message to parent");

The parent can register a handler for this event in the same way:

$.windowMsg("parentevent",  function(message) {
    // do  something  with message from child
});

That’s all.

You can see a simple demo here. (make sure your popup blocker is turned off).

Download the plugin and demo files
One thing to keep in mind is that it is of course a requirement that one of the windows opened the other, and that neither window has navigated to any other pages. Once they’ve lost the reference to eachother, there’s no reconnecting. The triggerWindowEvent and triggerParentEvent functions return a boolean which indicates if the event could be sent or not. I could pretty easily add a function that could check if the parent or child window is still there actually… will add that in the first update of the plugin.

As with the old Perl driver for Zoneminder (which I’ve been told doesn’t work anymore with the latest version of ZM, since the interfaces for the PTZ drivers have completely changed), this driver only controls the pan and tilt functions of the camera. Actually, you can also put it in swing mode and store or access preset positions, but none of the available functions actually do anything with the video stream. The video can simply be accessed at:

http://yourcam/image.jpg to grab a single frame of the current camera view
http://yourcam/video.cgi to get the mjpeg stream from the camera

There’s no real zoom control since the cam only has a digital zoom, which I believe is implemented in the client software only (so there’s no way to actually have it stream a zoomed image to your client).

Okay, onto the example script now. The code below demonstrates how to use all the features of the camera driver.

require 'Ip400.rb'

# substitute your host IP address or name and port number here.
# If you have turned off access control, you can simply use:
# ptz = Ip400.new("192.168.0.140")
# if your camera is accessible at a port other than 80, simply
# specify the address like this: "192.168.0.140:90"
ptz = Ip400.new("192.168.0.140", "ruby", "ybur")

# go to the home position
ptz.home
sleep 3

# pan right 3 steps
ptz.pan(3, Ip400::RIGHT)
sleep 3

# pan left 5 steps
ptz.pan(5, Ip400::LEFT)
sleep 3

# tilt up 3 steps
ptz.tilt(3, Ip400::UP)
sleep 3

# tilt down 3 steps
ptz.tilt(3, Ip400::DOWN)
sleep 3

# pan and tilt at the same time
ptz.move(4, Ip400::RIGHT, 2, Ip400::DOWN)
sleep 3

# go to 1 of the 24 preset positions
ptz.goto_preset(1)
sleep 3

# visit all preset positions with about 5 second intervals
ptz.swing_preset
sleep 20
ptz.stop_swing # stop the swing mode again

# go into continuous scan mode (scan left and right with 1 second intervals)
ptz.swing_scan
sleep 10
ptz.stop_swing  # don't forget this!

# get the current absolute position of the camera
pos = ptz.abs_position
# store this position at preset 15
ptz.set_preset(pos['pan'], pos['tilt'], "myposition", 15)

These examples should get even a Ruby novice going pretty quickly, and the Ip400.rb driver class can be used in or outside of a Rails app. The camera you’re controlling can be in a different room or even in a different country.
I’m releasing it under the MIT license so it’s open for further hacking. If you end up using this code anywhere I’d love to hear from you, and I’ll release it as a real Ruby plugin if enough people show interest.

Have fun!

Download the above example file and the IP400.rb file.

The above image of the view from the East Bay hills is at most a couple of minutes old (unless the camera is down). Click to view yesterday’s timelapse movie.

I came across a hilarious essay by David Sedaris on youtube that (fairly accurately) describes the Saint Nicholas tradition in Holland. You can view it right here – highly recommended. (it’s in 3 parts, make sure to watch all three of them…)

Feel free to use the information in this page in any way you want. I make no guarantees as to its accuracy or completeness. If you write any software based on this info, please let me know!

control commands:

The following CGIs are supported.

/PANTILTCONTROL.CGI

This is the main CGI for camera movement control, and should always be a POST request. The commands are very simple, there’s one parameter for specifying the direction of movement, and both for horizontal and vertical movement you need to set the degree at which it should move. This degree should be an integer of 1 or higher. (The web control interface never submits values higher than 10, and I haven’t tested any either).
I don’t think these commands can be used to do absolute movement.
Theoretically, if you move 3 degrees up and 3 degrees left, then 5 degrees up and 9 degrees right, then 8 degrees down and 6 degrees left, you should be exactly at the point you started – I’m not sure how accurately this would work, but you might be pleasantly surprised. (preset positions don’t seem to wander over time, so moving relative distances might be accurate too).

On to the commands now. To move the camera 1 degree up and 1 left, you should send the following command in the POST body:

PanSingleMoveDegree=1&TiltSingleMoveDegree=1&PanTiltSingleMove=1

To send the camera 3 degrees to the right, send the following:

PanSingleMoveDegree=3&TiltSingleMoveDegree=0&PanTiltSingleMove=5

I hope you’re catching on. Here’s the complete table of movement directions that should be set with the PanTiltSingleMove parameter:

0 up left
1 up
2 up right
3 left
4 home
5 right
6 down left
7 down
8 down right

If you’re just tilting the camera (vertically), you can leave out the PanSingleMoveDegree parameter. Likewise, if you’re only panning the camera (horizontally), you can leave out the TiltSingleMoveDegree parameter. If you’re sending the camera to the Home position, you can leave out both.

Note that there’s no way to control the zoom! The IP400′s zoom is digital only. The java/activex interface and windows application all implement the zoom on the client side; the camera itself doesn’t know how to do this. So if you need the zoom (or the suggestion of a zoom), you’ll have to implement this yourself by cropping and resize the center of the video stream.

SWING MODES:

There are two swing modes that can be activated simply by passing the PanTiltSwingMode parameter and value to PANTILTCONTROL.CGI. The first scan mode will simply scan the room horizontally by panning it one step in every approx 1 second intervals until it reaches it’s left or right limit, then reverse direction. The swing mode will simply scan all preset positions and hold each position for a few seconds. I don’t think there’s any way to change the interval time for either swing mode, at least the decompiled camera java applet didn’t reveal any such options to me.

values for swing mode:

0 stop
1 horizontal scan, move every second
2 scan all preset positions

example:
to put the camera into horizontal scan mode, send:

PanTiltSwingMode=2

PRESET POSITIONS:

To move to a preset position, send the following command:

PanTiltPresetPositionMove=3

substitute 3 for the position number that you want to move to. The camera supports 25 preset positions, 0 to 24.

To clear a position, send:

ClearPosition=12

this would clear position 12 (which is the 13th preset)

To set a preset position:

send the following command.

PanTiltHorizontal=a&PanTiltVertical=b&SetName=myPosition&SetPosition=12

Where a and b would be the absolute values for the horizonal resp. vertical positions. There is some numbers juggling going on in the java applet of the webcam control application, and I haven’t made any attempts yet to reverse engineer this. All movement (except for movement to preset positions) is relative with this camera, and I am not sure that an absolute position can be calculated accurately from relative movements. As far as I know there’s only one way to get the current absolute position, which is actually through the mjpeg cgi, see below.

———————————–
/IOCONTROL.CGI

It appears that this CGI can be called to trigger the camera to upload or email a current camera snapshot. I only know this from decompiling the xplug class, and haven’t played around with it at at all.
The parameter string should look like this (or at least that’s how the java applet sends it)

Trigger1=0&Trigger2=1&ImageUpload=0&ImageEmail=1

where values of all 4 parameters should be either 0 or 1. You should call the CGI with the POST method, however, if you use the GET method (eg by simply entering the CGI url in your browser), it will answer with the current values of these parameters, like this:

Trigger1=0&Trigger2=0&ImageUpload=0&ImageEmail=0&CurrentTime=2007-08-16 15:27:34

(There’s no point in trying to set those properties in the GET request (with something like /IOCONTROL.CGI?ImageUpload=1) as the IP400 seems to ignore those: the camera apparently needs a POST request to receive parameters. )

/MJPEG.CGI
/VIDEO.CGI

(use GET request)
Both these urls appear to return a MJPEG stream, which is basically just an endless stream of jpeg images with a short header in between. That header data looks like this:

--video boundary--Content-length: 10089
Date: 2007-08-16 14:56:19 IO_00000000_PT_136_046
Content-type: image/jpeg

but the ‘Date’ header isn’t present when using the VIDEO.CGI variant. The interesting thing about this Date header is that the last bit, the PT_136_046 actually indicates the current absolute Pan/Tilt position. The Ruby driver I’ve written for this camera can grab those coordinates from the stream, which is really only useful when you want to set a preset position.

You should be able to view the stream in a browser simply by using the address http://192.168.0.50/MJPEG.CGI (substitute the correct IP address of course).

Other than the missing Date header in the VIDEO.CGI response, there don’t seem to be any differences between these two streams. If you find any, please let me know.

By the way, the exact request that Trendnet’s xplug java applet sends to get the video stream is:

GET /MJPEG.CGI HTTP/1.0\\r\\n
User-Agent: user\\r\\n
Authorization: Basic abcdefgrn\\r\\n

where the authentication token is optional (but needed when you have user access control enabled in the camera). This token abcdefg should be the base64 encoded string of ‘username:password’ as is standard for http basic authentication. It’s probably best if your request look exactly the same, and do make sure you end those headers with double carriage return and line feed, as illustrated.

/IMAGE.JPG

simply returns a jpeg of the current camera view, which is handy if you want to save a snapshot or display on clients that can’t handle streaming video.

If you need to grab a bunch of frames, want to convert the stream to another format or do any other magic with your camera view, I recommend you look at ffmpeg and imagemagick.

Here’s an example of how to save the first image from the mjpeg stream to an image:

ffmpeg -f mjpeg -i http://192.168.0.50/MJPEG.CGI -vframes 1 -an -s 320x240 cam%d.jpg

It will save multiple jpgs if you increase the vframes value, but if you want to do that it would be better to output to another video format then to output to jpg.

Having written all this: you may not need to know any of this since I’ve written two drivers, one Ruby and one ZoneMinder driver that can be used on any platform that supports Perl or Ruby. I’ll post the code for the ruby driver as soon as it’s ready.