Firstly, the example was given of "why does Photoshop ask you for both the name of the file and the place to save it?" In the context of interfaces being too overwhelming in the number of questions they present.
-
Show this thread
-
It asks you both because it needs to do both. It asks you both because both of those pieces of information are important to doing the job you're asking it to do. That's the task. The tool already exists.
2 replies 0 retweets 2 likesShow this thread -
And yes, I know, the whole deal is to somehow magically question whether or not the entire abstraction of files and file spaces is somehow inherently inferior to implicit clustering by temporal and task attachment.
1 reply 0 retweets 4 likesShow this thread -
Remember how I said there was decades of user interface design that were hand-waved by the writer? This is part of it. It's not like this is a new line of questioning.
2 replies 0 retweets 2 likesShow this thread -
-
Replying to @Conaw
The old answer is twenty years of semantic file systems and GNOME Shell where it's one of the first things people turn off once installed. So you're right, but not about what you expect to be.
1 reply 0 retweets 1 like -
Replying to @squidlord @Conaw
My own feeling is that the previous attempts AND Mercury both try to usurp too much. They try to replace and not adjunct, which provides no on-ramp. They forget people already have jobs to do.
1 reply 0 retweets 2 likes -
Replying to @squidlord
Agree here too iPads arent computers iOS isn't an empoweding OS MercuryOS is a good conceptual framework for **something** related to (programmable attention)TM-
@andy_matuschak2 replies 0 retweets 2 likes -
Replying to @Conaw @squidlord
My broad sense is that the desired end-state is an inversion, a dual of the current structure: instead of apps-as-nodes, we have apps-as-edges. Roam and LiquidText are nouns/places; I think their ideal form is as verb. Roam/etc-as-OS is roughly what’s necessary for this, I think.
2 replies 1 retweet 9 likes -
Replying to @andy_matuschak @Conaw
We've tinkered with "apps as service providers" (not in those terms, of course) before. Plan9 took it pretty far down into the Mach kernel, even. It has some useful bits but the multiplicity of desired ways to touch ANYTHING means things get unpredictable fast.
1 reply 2 retweets 4 likes
Indeed. I don’t know a workable model for this. It’s interesting to see TimBL try it again with Solid. The “no-code”/“Zapier-driven” trend is interesting to watch, too, since there’s a greater emphasis on end-user-usability. But I’m not optimistic here.
-
-
Replying to @andy_matuschak @Conaw
Maybe there's not a workable model -- at least while maintaining continuity with current workspaces and actual work getting done. The idea of externally-hooking pipes is probably the best possibility but dealing with API limitations will hurt them, always.
1 reply 0 retweets 0 likes -
Sooner or later, it BECOMES code, even if it's dragging lines and connecting pipes with filters and transformers and not writing text.
2 replies 0 retweets 2 likes - 5 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.