I have downloaded DevonThink Pro Office and am evaluating it for purchase.
I have happily imported my existing PDFs, and was starting to scan
some more documents with my Fujitsu ScanSnap.
The first two documents I scanned worked just fine, they scanned OCRed, etc.
Every other document I have scanned since scans, seems to OCR, prompts me for a title, and then after I hit the ‘Set’ button… disappears. It appears to be nowhere
in the database. I can search for very distinctive parts of the title (like 2007-02-10) and DevonThink cannot find them in the database. I can look at the Tools->History and I do not see the document I just scanned listed there. It’s as if I never did it.
So my question is: Is this some sort of limitation of the DevonThink trial version?
Or is there something more mysterious going on?
Are the non-working documents over 50 pages in length? The OCR license that DevonThink has only allows documents up to 50 pages. I can’t remember the behaviour when it hits over 50, I think it just throws away all the work it’s done (but I might be wrong on that).
If you want to do documents over 50 pages, you’ll have to first split it into multiple documents, OCR each of them, and then merge them together again.
If you search through the forums, I posted some Python scripts that can do the splitting and merging. If you can’t find them, I’ll post them again here.
There is a limit of pages per day for OCR testing in the beta or demo period.
As I remember, it’s 20 pages per day.
When you do searches are you using the Tools > Search window? It’s easy to inspect in that Search window for the selected operators, including whether or not the search is database-wide, whether you are searching in the Name or Comment or All, etc.
OK… so I looked through the history… I’d only OCRed about 4 pages total…
However, after midnight rolled by, I was able to scan and OCR 2 additional pages (all I’ve tried so far). So it does indeed seem to be a per page limit in the demo.
I have purchased DevonThinkPro Office and am extremely pleased with it. Not only did I just scan a whole stack of documents into it, but it correctly figured out things as intricate as my health insurance notifications, where one page is portrait and the next is landscape. It correctly OCRed the portrait page, and then rotaited the landscape page and correctly OCRed it as well. This was a MUCH better behavior than I had expected
