I am creating this thread to post any big developmental updates for the Chain Scanner web app, created based on an implementation designed by Beeminder forum user alltom based loosely upon forster lists. I am also collecting feedback here so it’s all in one separate place.
For an explanation of how to use the app, please see the original Forster lists blog post, before opening the forum thread below, or open up my app to see how we implemented TrueSkill to serve up the next task, instead of having to manually scan down candidate tasks and use a lot of brainpower.
Forum thread where I initially post an explanation of how to use this web app.
Bug report that might be specific to me having tried this quite a while ago and then dusting it off today and only now trying to sign in via Google: When I do it seems to work, then it says syncing, but then ends up cycling between “synced”, “syncing”, and “error”.
Ugh, I thought I squashed that. Refreshing the page seems to fix that particular error for me on my end when it first popped up, but I thought the root cause was fixed but will take another look at it. It should be fixed by midnight tonight if I catch it. Thank you!
Forehead-smack. I think that fixed it for me too! I had thought I had tried that but maybe not. Also I was very confused that clicking that button logged me out, until now that I finally noticed the hovertext.
Bug should be squashed now. Users with data from before Sep 23rd will still need to fix the sync badge bug if they have it by refreshing their page once. Sorry about the bug you ran into, it’s a bad user experience to need an unprompted refresh to fix a bug, but I am confident this will avoid being regressed and I have setup harder gates to prevent future potential issues like this from popping up in the repo again. Ty for the reports and the rep in the newsletter!
Now that we have this dedicated thread…
I think I am doing something wrong in an attempt to create a chain of just work tasks, so that things that are in the list but are morning/evening routines won’t interfere. I thought I did that by selecting the contexts on top of chain creation, but things that are offered are from any tasks, not just those marked as “work”.
Could you please suggest the correct way to achieve what I am intending: that for this session, I only get tasks marked with a specific context as candidates? I can, of course, click “can’t” through all those that are not in the desired context (and will do this until I know the correct strategy)… but I thought that contexts exist specifically for this reason, so…
I think I understood what was happening. When a context is selected, this precludes those exclusively marked with other contexts to appear. So items marked with this context appear and those that are not marked with any context appear as well. Good!
Sorry about the late reply! That is a good solve, but it’s a little convoluted and unintuitive.
I am thinking about maybe adding a “no context” context, and being able to click on the context a second time to turn it from a check to an x, excluding this context from being scanned, while no context can only be toggled between check and x. The behavior you described there is yeah technically true! When a context is enabled, tasks with that context are eligible to be scanned.
On the flip side, users might want to make it so that a task is only eligible to be scanned if multiple contexts are true at the same time, but I don’t think I’m going to implement that unless someone asks for it. If you do want that, LMK, but contexts are a blind spot for me as I don’t use them as much, because of the hassle.
I am thinking of taking a page out of Omnifocus’ book and allow (optionally) to geotag contexts so they’re automatically enabled or disabled when you’re home or at work, but remains to be seen. Probably will implement if PWA location polling ends up not being too annoying by early next week.
No, I think it makes perfect sense the way it works now, it just took me a bit to understand the underlying rule - that the tasks I didn’t expect to see with the filter were the ones with no assigned contexts, not any random tasks. Maybe adding a sentence on this behaviour in the FAQ along the current response to “How do contexts and list filters work?” will be enough of clarification if anybody else also stumbles upon this.
