unifying capture methods

[sjk]

Here’s part of my recent post from the beta forum, re-posted here to possibly spark some general discussion:

@sjk

Lately I’ve been thinking a lot about how the interface for capturing data might be unified and simplified. Is it necessary to expose so many methods (Services, bookmarklets, scripts, etc.) directly to users or could there be a proxy interface that makes or guides choices based on specific actions in different contexts? For example, I’d prefer to use a relatively small number of hotkeys for different capture methods without having to separately define them for every context. In cases where a non-interactive hotkey is unavailable there’d be an interactive fallback, similar to the new Clip to DEVONthink bookmarklet. When interactive, the destination selector and capture method could be combined into a single dialog. Maybe even borrow ideas from the Sorter (which, even without its problem with Spaces, is too cumbersome to me).

Summary: I’d like data capturing to be primarily choices of formats and destinations without being burdened by too many choices with the methods.

EagleFiler’s Capture key (non-interactive) and Capture with options (interactive) shortcuts attempt to do that.

Last thought for now:

When tags are implemented I’m certain there will be requests for some way(s) to assign them when data is captured rather than be forced to do it later.

[acl]

@sjk

Lately I’ve been thinking a lot about how the interface for capturing data might be unified and simplified. Is it necessary to expose so many methods (Services, bookmarklets, scripts, etc.) directly to users or could there be a proxy interface that makes or guides choices based on specific actions in different contexts? For example, I’d prefer to use a relatively small number of hotkeys for different capture methods without having to separately define them for every context. In cases where a non-interactive hotkey is unavailable there’d be an interactive fallback, similar to the new Clip to DEVONthink bookmarklet. When interactive, the destination selector and capture method could be combined into a single dialog. Maybe even borrow ideas from the Sorter (which, even without its problem with Spaces, is too cumbersome to me).

Summary: I’d like data capturing to be primarily choices of formats and destinations without being burdened by too many choices with the methods.
EagleFiler’s Capture key (non-interactive) and Capture with options (interactive) shortcuts attempt to do that.

I also find all the different ways to capture things to DTPO confusing; my solution is to use fastscripts and assign application-specific keys to a script that does whatever it is I want done in that application. Of course, one still has to define it for each context.

Doesn’t work if the application isn’t scriptable though (but I avoid relying on such applications, except when I have no choice).

[dbyler]

I think we’ve almost got ubiquitous capture via the Services menu items.

I’d like to see the Services handled by the DEVONthink Sorter, not DT itself. When DEVONthink is running, clipped items would be sent straight to DT; when DEVONthink is not running, items make their way into the Sorter. That way one could always clip without worrying whether DT is open or not (or being forced to wait for its launch).

[sjk]

Thanks for the feedback.

The Services capture commands/shortcuts are close to being ubiquitous for my purposes since they’re available in all apps I want to us them. I’ve also thought about making them behave similarly to what you’re suggesting. Of course if you don’t use Sorter (I don’t) DT would still launch if it’s not currently running when its Services are invoked. At least that launch time is more acceptable for me when only the relatively small global Inbox db opens instead of my larger one(s).

Backtracing a bit …

@acl

my solution is to use fastscripts and assign application-specific keys to a script that does whatever it is I want done in that application. Of course, one still has to define it for each context.

Ideally, by default, I’d prefer not to rely on scripts or third-party apps. Main goal is keeping it simple while serving most users. Obviously it would still be possible to create/use more specialized solutions.

Maybe a bit ironic that DTech has stated they want to minimize the number of preferences in their products, yet seem less concerned about the dizzying numbers and combinations of capture/import methods/formats/contexts.

[dbyler]

Agree, on most counts anyway. I don’t use the sorter either—the global inbox handles multiple databases nicely. I would love to see the Sorter jump into action when I invoke Cmd-) and DT isn’t open, though—much better to use the global inbox as a holding pen for those times when (which is when I probably need my RAM for other tasks).

[sjk]

I wouldn’t mind Sorter being a proxy for Services captures, but don’t need its UI.

My application usage is constantly influenced by the 2GB RAM limit on my iMac G5. I can sustain the current four swapfiles awhile longer (about five days now), and still comfortably use DT with my largest database, if I’m careful. Game over if I’d do something like run Toast to burn a Video-DVD of EyeTV recordings; it doesn’t play nicely with virtual memory management.

[fharvey]

I like the idea in theory, but the practice with so many users, so many with different backgrounds, so many ways of working, and so much diversity in needs, I think multiple ways is the best. What would help is a tutorial for each method that highlights key activities of the process.

Also, when tags work, there better be an option to add them when capturing data.

[sjk]

@fharvey

I think multiple ways is the best.

That needn’t contradict what I’ve suggested, which is basically a simpler front-end for capturing that could use existing methods (in certain cases) which would still be available.

What would help is a tutorial for each method that highlights key activities of the process.

Uh, huh. Better documentation of which capture methods are available, with recommended and example usage that could help answer repetitive questions like:

Best format for importing a web page?

Also, when tags work, there better be an option to add them when capturing data.

You mean like the last sentence of my first post mentioned?