Just a few things I’ve thought of that would make life easier.
Since some PDF’s import as images, the text isn’t searchable. Yet, I have to import the PDF, then display the ‘info’ for the PDF to discover why it isn’t showing up when I do searches. It would be nice if the non-content-searchable documents were visually identified in the list views. Maybe by color or icon?
It would be nice to be able to search content, name, url, comment simultaneously rather than only one at a time. I find myself having to do several searches to find all the docs I’m interested in, particularly since some of them cannot be found by content.
Since the only way to add descriptive text to an image doc is using the comments, it would be nice to have the editable comment field available in image document windows.
If text that looked like URLs in RTF or plain text documents could be recognized and selected, contextual menu options could process the link in various interesting ways, such as opening in DevonAgent, Safari, downloading to disk, importing text from the linked doc into DevonThink.
I use DevonThink to maintain bookmarks with associated notes. It would be really nice to be able to open a DevonAgent crawler already setup to search one or more sites selected in DevonThink.
As I mentioned in a different message, it would be convenient to be able to open documents stored in DevonThink, perhaps read-only without having to make a copy of the document outside of DT. This could be implemented using WebDAV. Any application supporting WebDAV, such as OpenOffice, could then open docs directly.
The menu item Tools->Properties is a little confusing. I keep opening it expecting to see document properties, particularly since the document ‘Get Info’ item is right beneath it. How about calling it something like “Database Properties”?
I hadn’t looked at IceCoffee before. You are right, it does most of what I want. However, if the ‘import linked doc into DT’ service was implemented in DT, could it import the all the metadata about the doc, such as URL and timestamp from the HTTP header, and title from the HTML header?
With the DevonAgent idea, it seems like the simplest behavior would be the first. In other words, create a new crawler set containing the selected links. I can always delete the crawler set after I use it if I’m unlikely to need it again. Maybe there is a way to make this more elegant, but the basic idea is to be able to search documents (web sites) that are stored as links in DT.
Anyway, I hope these ideas are useful…
P.S. I wonder if the AppleScript support in DT Pro will be rich enough to support a reasonable subset of WebDAV operations? If so, it shouldn’t be too hard to build an Apache based WebDAV front end for DT. Seems like a reasonable 3rd party opportunity.
Basically this command is in the "Content" menu because all other similar commands are over there and because this command does not save files. File operations on the other hand are of course in the "File" menu.