No doubt, DevonThink is a wonderful piece of software. My impression, though, is that it creates a “walled garden” and all of your data and metadata is then stuck in there forever with no hope of ever getting it out. It would be nice if their web interface could run on any old LAMP server, but it’s not that easy.
You can’t (really) virtualize OSX under Linux, sadly, as this would give an elegant solution if you’re trying to run a Linux server.
Some suggested alternatives are easily found by googling, but most fall short. A commonly recommended one is KnowledgeTree as they have a FOSS version. Unfortunately, it’s not what you think and the install is cumbersome. I’d recommend you try out the online demo before getting into installing it yourself – you can sign up via their website.
So I am still looking. What I’d like is the following workflow type thingee:
- Docs scanned to network share in PDF
- Periodically OCR’d
- Manually sorted, if need be
- Available on a local network in a searchable, retrievable form
One potential option is Zotero, but it doesn’t do any kind of previewing and it doesn’t search contents of OCR’d PDFs. For organizing scientific articles it’s the SHIZ though. I regularly use it for that purpose. If you keep your collections sync’d across the webs with dropbox, it can be really cool.
Referencer is pretty cool in that it offers a preview of PDFs, but it wants to make everything a scientific journal article.
Nautilus – the file manager – can do PDF thumbnails which is handy. There’s also a nice search feature in there. I’ve tried using gscan2pdf, but the OCR is laughable and it doesn’t create proper searchable PDFs. Did I try tesseract in addition to GOCR? Yes – tesseract is only marginally better.
There’s just not a good Linux solution I’m afraid, though just using Nautilus may be your best bet.