"OCR engine initialization error, or there is no appropriate license"

I have created a .Net Windows Service to help us OCR documents as a 'background process'. On my development machine the Service works perfectly, yet when I deploy it (via a Setup Project built in Visual Studio 2008) to the test server, upon install I get the error "OCR engine initialization error, or there is no appropriate license" when the Service attempts to start.

Searching Google and this forum, I discovered there are more files that are to be distributed than just the Leadtools.(whatever).dll's I referenced in my project:
http://support.leadtools.com/SupportPortal/cs/forums/5846/ShowPost.aspx

However, I have similar questions to the authoer of the above post, and since thy were *not* answered in the reply by Support, I shall ask them again....

My Service simply loads a document as a Raster Image, uses the Recognize method to collect any OCR text, which it passes on to another application. Nothing more is done, no spell check, no PDF creation, nothing.

As such, is it necessary to redistribute the hundreds of files to accomplish this task? Do I need to include the dozens of language files? I am not utilizing any of the BarCode features, so con those files be left out?

In other words, what is the smallest set of files I absolutely need to accomplish basic OCR so that the Service will have the smallest footprint on the machine as possible?

-- Andrew

Andrew,
I noticed in a previous post by you that you are using LEADTOOLS v15. Please note that LEADTOOLS v15 OCR does not support multithreading. Therefore, your web service will fail to execute OCR because web services send multiple requests to the engine.

If you can guarantee that your web service will only use the LEADTOOLS OCR features in one thread, it might work.

About deploying your application with minimum set of DLLs, you will need to follow the steps in the following help topic:
- Files to be Included with Your Application
For OCR, please read the paragraph "OCR Support" in the help topic mentioned above. Also, you will need to deploy all the engine DLLs with your application.

Adnan, I appreciate your attempt to answer my questions and provide adequate customer support, but if you want me to take your responses with any measure of seriousness please look up the difference between Web Service and Windows Service... the former makes remote method calls and queries, where the latter is something that runs in the background on a local system... such as a print server or file watcher.

So, while you've blown it out of the water by discussing a topic completely unrelated to my post (threading), failed to understand the context I was inquiring about but blazed on ahead as if you did (Windows Service), you also repeated the same oversight I have noticed in virtually every other post on here that asks about reducing the footprint of the OCR files needed to simply OCR any given document by offering the overwhelmingly helpful suggestion: "go read the help manual".

I have looked up the phrase "Files to be Included with Your Application" in the help documentation, of which there are no less than three occurrences. One of them is under the heading of OCR Support, which lists a couple hundred files (including language files, Spell Checking modules, FRX Language Recognition Files, barcode support, and even one file named "MiddleEarth.dll". All of which total nearly 100MB of files that must be redistributed with any OCR project.... according to the help topic you and your cohorts so quickly refer to, time and time again.

So for my little *Windows* Service that is no more than 15Kb in size, I am required to include over 100MB of redistributables for it to function. Which of course after doing so saw the application function flawlessly as it did on my development machine.... I just never suspected that in order to have an OCR application work on the production box, I had to bring with it virtually everything that existed on the development box. I guess the term "lightweight" is not in the LEADTools vernacular?

I must apologize Adnan, I do not mean to take my rant out on you for a failing squarely resting with the development engineers at LEADTools, but I do expect a better class of support. Next time, let's take a couple extra minutes, re-read the post as written, and then make every possible best effort to answer the questions that were asked. I know it's easy pulling out the Customer Support Q&A Handbook, finding an answer to the question that most closely resembles the one in the book... but let's try to resist that in the future, hmm? Thanks

Oh, and by the way.... I'm still waiting to hear if there is a reduced subset of files that can be redistributed with OCR projects? This is *not* mentioned in the "Files to be Included with Your Application" at all.

-- Andrew

Andrew,
I apologize for my previous post. My reply was not accurate about your problem.

However, about using LEADTOOLS v15 OCR engine in a Windows Service, you need to keep in mind the following 2 things:
1- You will need to register the OCR Thunk server. The OCR Thunk folder should be installed in this folder:
[LEADTOOLS 15]\Bin\Common\OCR Thunk
The instructions to register the thunk server are mentioned in this link:
http://www.leadtools.com/help/leadtools/v16/dh/to/leadtools.topics~leadtools.topics.ocrthunkserver.html

2- You need to run your Windows Service in an Administrator User account. The reason for this is to be able to write the temporary RDF files on your system.

About the files to be included in your setup, you will need to deploy the complete set of OCR engine files.
The OCR engine used by LEADTOOLS 15 is not designed by LEAD engineers; it is from Nuance and its documentation does not specify any optional files. If you want to try to isolate a subset of files by trial and error, this might work, but we will not guarantee it.