Questions about convert jpg ocr

What you need is called OCR - Optical Character Recognition. This is highly complex and you shouldn't try to build your own library for it. You can however use one of many libraries available.
Using Microsoft Office OCR library
Asprise OCR library for .net
A discussion on stackexchange with several libraries listed

Look in Microsoft Package, there is a really good OCR there which is ultimately free (apart of your Microsoft Package that is). Just go to your scanner, right click, scan the document with Microsoft image scanner and then you will see a small Microsoft Word logo which you click; and there you go. It's in word.

If one exists, t'd be more complex by at least an order of magnitude or two than a standard OCR, since an accurate conversion would have to account for a symbol's position in relation to all the other symbols in addition to conversion of the symbol itself. A regular OCR can usually get by with a straightforward top-to-bottom and left-to-right conversion.

You can use Optical Character Recognition (OCR) software to convert the .jpg into text, perhaps even directly from the scanner. Some OCR software may well have come with the scanner. Check for it.
But OCR is an art, not a science. I can almost guarantee that you will have to fix:
1. Mis-converted fonts.
2. Mis-read characters (i for l, for instance, o, 0 or Q mixups)
3. Mangled layout.

You need OCR (optical character recognition) software in order to convert scanned images to text. Even in PDF form, they would just be large images without OCR processing.
If you have at least Business edition of Vista, it comes with Windows Fax and Scan, which can save your work as multi-page TIFF files.
Microsoft Word and OpenOffice.org (which is free) probably have ways to scan documents to PDF. You can use SimpleOCR (also free) to convert to text if needed.