Thanks for the response!

I'd probably not cover mime type handling in the class, so here is a code snippet for how I would do it. This relies on the mime type being known to the machine so I'd try and integrate your custom collection with this.

        public static string GetMimeType(string extension, out bool wasFound)
        {
            wasFound = false;
            string mimeType = "application/octet-stream";
            if (!string.IsNullOrEmpty(extension))
            {
                if (!extension.StartsWith("."))
                {
                    extension = "." + extension;
                }

                try
                {
                    bool setMimeType = false;
                    RegistryKey key = Registry.ClassesRoot.OpenSubKey(extension.ToLowerInvariant(), false);
                    if (key != null)
                    {
                        string value = System.Convert.ToString(key.GetValue("Content Type", string.Empty));                        
                        if (!string.IsNullOrEmpty(value))
                        {
                            mimeType = value;
                            wasFound = true;
                        }

                        key.Close();
                    }
                    
                    wasFound = setMimeType;
                }
                catch (SecurityException)
                {
                    // No rights to key.
                }
            }

            return mimeType;
        }

I stumbled across this thread when trying to determine if the Laserfiche OCR engine can process more than one document at a time.  I am investigating building multi-threading into an OCR scheduling application but I am unsure whether or not this is possible from an engine stand point.  

I am trying to avoid building a multi-threading capability to process OCR on documents more efficiently only to find out the OCR engine can only process one document at a time.  

Does anyone have any insight here?