I was initially attracted to Yojimbo because of its iTunes/iPhoto-like Library, but found it too limited and restrictive in other ways. No hierarchical organization, as you mentioned, and no user-definable Smart Collections.
When I tried EagleFiler it was sluggish (especially launching) on my 2GHz iMac G5, maybe because parts of it use python. It claims:
• Like iTunes and iPhoto, EagleFiler keeps your data in a library.
• Tag sources behave like playlists in iTunes or albums in iPhoto.
Those comparisons with iTunes/iPhoto seem kind of stretched. Unlike iT/iP, EF’s library is essentially a file/folder structure, which didn’t appeal to me. And tag sources can’t be hierarchically organized; iT/iP have playlist/album “folders” (a misused term for them, IMO). Not sure how buggy it is, though Michael Tsai was responsive to issues (not too serious) I reported during testing. More potential than Yojimbo but not enough to sway me from DEVONthink Pro.
Together has some appealing “iTunes/iPhoto for documents”-style features but it’s definitely buggy. Tried it about a month ago for the first time since it was renamed from KIT and within 15 minutes I’d discovered at least a half dozen problems without any effort. Plus it was much too slow on my wife’s 1.25GHz eMac (our only Leopardized Mac, for now) with barely any data. 'Tis unfortunate the implementation is relatively fragile, compared to DTP that I can punch in the mouth without breaking its teeth.
I’m mostly a piler (at least initially), as my wife will attest to. Piles have more visually diverse cues for me, making them easier to put and find certain things in than attempting more formal filing/organizing strategies. Some piles can eventually migrate to files, though. How that’s applicable to DTP in a moment.
I’ve purchased Notebook, too. Still more in the tinkering than productive stages with it. I wouldn’t consider it a replacement for DEVONthink. I’m satisfied with it as a robust note taking app (it’s primary intended usage?), not expecting it to be a general “digital shoebox” like I mostly use my main DTP database for. I’ve got several smaller, more purposeful, well organized databases, too.
DTP still does a pretty good job at simultaneously catering to my piler, filer, and clutter tendencies towards information (dis)organization simultaneously. For too long I let its hierarchical structure get in my way, but eventually created fewer groups that now work more comfortably like piles. Some information I still prefer more strictly filed/organized, either within my main database or in separate ones. And there’s the clutter that either doesn’t seem to belong anywhere or isn’t enough of a priority to pile/file.
Another thing I really like about DTP (e.g. compared with the aforementioned apps) is the different views. Those help support the database-within-a-database usage of my main database.
And there’s comfort in knowing that if something’s in DTP it’s usually a matter of moments to find it, regardless of its age, content, or location.
Downsides? I still wish DTP were closer to the iTunes/iPhoto Library and playlist/album method of storage and organization (which is one Yojimbo, et.al. have been temptations). Treat all documents like they’re contained in a single library, hiding any hierarchical storage it might use (unlike EagleFiler); DT’s History window view comes close. Treat groups more like virtual hierarchy-supporting iT/iP playlists, with documents they contain behaving like aliases (Finder terminology) or replicants (DEVONthink terminology); sort of a superior form of Finder’s Burn Folders. Add a flexible layer of virtual organization for documents that doesn’t rely on or normally expose any underlying file/folder storage, which is something I love about iTunes/iPhoto (that other people hate). Being able to create and destroy different organizational structures without caring about the physical location of the data (other than knowing it’s safely stored somewhere in the Library until explicitly, intentionally deleted) would be usability bliss for me with DTP. DTP 1.x databases have that kind of backend, but their frontend group/document hierarchies aren’t virtualized from the user’s perspective.
I’ll wrap this up (finally!) with thumbs up to Katherine’s post. 