:laugh: that made me chuckle yer nothing like old school. At least the article is popular :).

Hi, OK I sorted an external hosting site that will allow hot linking and not delete the files after a period of inactivity. The problem I have now is that in such a case I won't receive any points for downloads. That's 60 odd already and it's only been up 2 hours. I take it I'm in a situation of like it or lump it??? :^) Cheers Chris

Thank god for spell checkers :laugh: . I'm an engineer and without them and a calculator I am 1st class paper weight. Chris

OK I will see if I can find somewhere I can have a direct link, thanks for your help anyway :) Cheers Chris

Alright, I have re uploaded the x64 source as this did not upload correctly and now both source files should download fine and extract, Cheers Chris

Hi, Sorry about the late reply I was away with work. Yes I'm sorry it appears the files did not upload properly I am trying again from work, I'll let you know when I've uploaded them, Thanks for your help in this matter, Chris

Yes, we can and it is my back up plan however this will redirect the user or open another window/tab to the sourceforge website and this is less than ideal. Cheers, Chris

I have uploaded them to source forge for the time being it should be available here: https://sourceforge.net/projects/emgufacerecog/files/[^] If not let me know and I'll try somewhere else. Sorry for the slowness my upload speed is dismal :wtf: Chris

Hi Guys, I have a problem my source code is 22mb (x64) and 70mb (x86) :^) granted this is large but it's face recognition using real time processing and parallelisation so it's bound to be large. Is there any way I can get this source code uploaded or am I stuck asking users to download several little zips and extracting them in the same place? Or using an external file hosting website. I would love to get this article up and on-line as it's some of my best work any help would be appreciated. Many Thanks Chris

Surprisingly not for which I'm truly thankful, It was a very silly newb of a mistake I think its nice to see the community didn't scold me to badly for it. It is wasn't for Chris Maunder quick admin reactions I think things would have got a little messy :s

Thanks for your help, Chris

Hi all, This is a post I mistakenly put in the Lounge section (For which I apologise). I have placed it here as suggested by others. On an additional note how can I now delete this post from the lounge forum as the delete option is blanked out? Does this require a certain amount of reputation points? I really hate repeating posts so why can't a user new or not delete there own post? But the my main question that got me here is.... I've written an article based on creating an EMGU project (Image Processing Library). Generally because I'm tired of answering similar questions about the correct set-up of EMGU. I wanted to create two new tags EMGU and EMGUCV which are the common references used for the library. Now I've read the documentation about just writing the tags in the Tag text area and new ones will be created. But this does just not happen! Can anyone of the administrators create them for me or someone explain how to create them as there are several questions that also require the tag. Many Thanks Chris

Hi, The Netduino is a great little tinker board however is limited in it's IO ability. As a clear frontrunner I would suggest the Fez Panda 2 the same website for GHI electronics has alternatives. I would strongly recommend staying away from Microsoft's ".NET Gadgeteer" it is simply the same FEZ based board however at 2 -3 times the cost. tinyclr FEZ (Frekin Eazy) boards have been around a little longer than Netduino and has a much stronger community. The advantage of both board are they are Arduino compatible which means the shields that plug into the the top will work. Great as you can get cheap ones of ebay however be careful as you must re-map pins for the much cheaper ones. The reason I suggest the Panda II as well is that you can use these shield however with the extra header you can reduce the space your require. To help sceptics here is a few comparesents

    FEZ PANDA II            |            NETDUINO

72MHz. 32-bit ARM7 processor | 48MHz, 32-bit ARM7 processor
|
148KB for user application | 128 KB for user application
|
62KB RAM | 60KB RAM
|
54x Digital I/O ports | 20x Digital I/O ports

and thats just for starters as the size of your project is quite substantial and complex the PANDA II will give you the extra where ever you need it and only for an extra $5 or £5. Hope this helps you on your way, Cheers Chris

Hi, I think I understand your problem I take it you are using interrupts with your serial comms as in my article:Serial Communication using WPF, RS232 and PIC Communication[^] . I'm an electrical engineer so I do a lot of work with micro-controllers and programming on UI levels. From my understanding at the moment the you can not determine which micro-controller has sent the data of which the interrupt loads and displays. There are a few solutions to you problem which depend on your ability to reprogram the micro-controllers. 1st I will assume this is not an option: If you can ensure that the micro-controllers respond one after each other the you could try the older approach of creating another thread and reading all the data into a buffer this more common in C# applications. I can provide an example of this if required however it's not used that often any more. This will attempt to read the data in a let you deal with it a bit at a time however is less efficient than interrupts and your likely to to loose data. The best option to ensure data integrity is simply to wait after polling each micro-controller until data is received however I understand this may not be possible if at time the micro-controller does not respond with data. If this is true you will have to introduce a timer to continue after if no data is received. 2nd If you can reprogram the micro-controllers: This is far easier, 1st simply ensure each micro-controller sends and ID tag before sending any data this way you can determine the start of each package and the device. An alternative which is the bus standard norm is a busy line. It requires an extra connection between your micro-controllers when a controller receives a request for data it will set this line high. Before any other micro-controller can send it's data it must wait for this line to go low. This method again relies on data always being sent on request and has inherent problems such as if one micro-controller takes a considerable time to transmit for example it has a long string of data to send you could find yourself polling for data from micro-controllers before they have had time to send there previous data. This timing issues are important to deal with and I would recommend the first option for ease as you could still end up