I built the extension in full opinionated-mode, building exactly what I wanted, including the semi-arbitrary decision to exclude red goals. On reflection (four years later!), eh, I kind of buy it that it’s probably better to be consistent and have the button on every goal.
But on a similar note of opinionated software and anti-magic: I don’t super like the way Beeminder sorts goals. It’s kind-of sort-of intentional that the extension has a simpler and less magical method of sorting. This wasn’t so apparent before, but with the new super-magical PPR sorting it seems to have become more salient.
I personally don’t really use PPRs on my do-less goals, so I don’t have much of an opinion about the change, other than my general opinion that the sort order was to magical already.
Thinking about it now, perhaps my ideal sort order would be by integer days until derailment, then alphabetical by slug. A simile, non-magical ordering with no guessing about where a goal is located relative to other goals of the same color/days until derailment.
I am actually tempted to implement this sort order in the extension (yay opinionated software, right?), but at this point I think I’d feel like a spoilsport if I didn’t support this fancy new feature (even though it’s one I don’t use.)
Well, whatever. For now I’ve released a new version that allows collapsing red goals and uses the urgencykey sort order. (Maybe at some point I’ll make a version that adds my proposed sorting as an additional sort option…)
Oh, I am excited to be able to collapse red goals! On the one hand, it was helpful that at the start of the day if anything was already collapsed and red, it popped back out. On the other, it was such a pebble in my shoe not to be able to collapse a goal I stopped thinking about. Sometimes I wanted it to work, and sometimes not… I’ll have to see what behaviour I prefer in practice!
Thank you for the new version. When possible can you include the Chromium version as well so that it can be used in non Firefox browsers (e.g. Chrome, Edge, etc)?
That makes a ton of sense to me! And I’d love to hash out the sort-by-urgency question and understand your ideal sorting criteria better. It sounds like PPRs-first isn’t actually skin off your nose if you don’t have such goals in practice? Is it by chance the case that all your goals have the same deadline time (mine mostly do)? If so, sorting by deadline next would also not affect you. (And it sure seems natural to me to have goals due sooner ahead of goals due later, when deadlines are different.)
Finally there’s sorting by pledge, which also seems pretty natural. Oh, and note that the current sort-by-urgency does use alphabetic-by-goalname as the final tiebreaker so the sort is stable. You’re sure that doesn’t suffice?
Like if you use neither PPRs nor variable deadlines, sort-by-urgency is just sort-by-color/safe-days, then by pledge, then alphabetic. That’s not toooo magical, is it?
One other thing to clarify for those maybe just skimming all this: You can still sort purely alphabetically or purely by pledge by clicking the corresponding column header on the dashboard. I’m now wondering if we could expand those sort options to get the best of all worlds. Like sticking in “SAFE DAYS” as a header and using @zzq’s ideal sort for that, if there’s enough demand for that. I worry it’s confusingly similar to sort-by-urgency though.
You can download the ZIP version (as posted by zzq above), extract it into a new folder, then go to chrome://extensions (or replace chrome with edge or brave depending on your browser).
Enable developer mode and click the button load unpacked that appears. Navigate to the folder into which you have extracted the extension and select that folder (the one that includes the file manifest.json). The process is a little bit tedious, but not too bad.
PPRs first indeed isn’t skin off my nose, as I don’t use that feature. (This January I gave it another try, and had a goal with PPRs for around 2 weeks before decided that it was superfluous and annoying and removing the PPRs.)
Perhaps similar to you, most of my goals have the same deadline, but that’s precisely the issue. The three that don’t (one with a slightly earlier deadline, two with a later) are “out of order” realtive to expectations. Of those things that I need to do within 3 days, does it matter that one of them needs to be done by 10pm and the others by 12:30 am the next day?
That’s an order that feels very arbitrary even though I know what’s going on.
It’s worse (more magical) if there are only a few exceptions to the otherwise sensible order. That you can mostly rely on the order, but with a few exceptions, is precisely what the anti-magic principle is against.
Strongly agreed. That’s the problem with sort-by-urgency! It’s almost right, 90% right, but then it throws distracting curveballs.
The Chromium version exists, but I probably should make the link more prominent. I’ve been updating my previous zip download link (which @felixm has also now linked to), but probably it’s best if I keep the first post of this thread pointing to whatever the latest version is, not only of the Firefox extension but also the zip for Chromium. I think I’ll do that.
Ah, sounds like you’re one of these weirdo non-edge-skaters! If, hypothetically, those goals were all due in a small number of hours, would you not want them in deadline order? And if not, why give them different deadlines at all? They all default to midnight, right? If you never change that default then this too would be no skin off your nose sorting-wise.
Well… With my current setup, I find that in practice some 5 or so of my 30 goals are red each day. What’s the definition of edge-skating? Does that count? It’s definitely true that both a) I expect to have a nontrival amount of red goals each day; and b) most of my goals aren’t red most days. (A while ago I scaled way back to focus on the goals that were most important to me: before that I had ~60 goals, with maybe 10 or so of them red each day.)
If there are a single-digit number of goals in the red, I am aware of them all, and the order between them matters less. It’s only further down the list that the order actually matters, as I scan it to see what the status of my not-immediately-pressing goals are.
I have a default deadline of 12:30 am, but there’s one goal which I know there is no way I’ll do it if I’m too tired, so I gave it a deadline of 10 pm so as not to be tempted to put it off until late in the evening (and then not do it.) For goals related to a bedtime routine, I have a later deadline (4:30 am), so that I can do them just before going to bed even if I’m up late. So I’ve got a total of three goals with unusual deadlines, one early and two late, for what I think are sensible reasons.
This is making me think the use case is pretty corner-case-y. You want to make sure one goal is done by 10pm and the others you want to allow more time for. But then you don’t need that 10pm goal to show up first on your dashboard because you have few enough red goals that the one due early won’t ever be out of sight, out of mind. What you do want is for it to be sorted alphabetically when it’s in a sea of green goals, to make it easier to find.
Maybe the really ideal sorting for your use case is to sort red goals by urgency but non-red goals alphabetically. [EDIT: technically: first sort by color; then among the red goals, sort by urgency, and among other colors, sort alphabetically.] That’s another inelegant if-statement but I do see the logic. As a user that one would be no skin off my nose. Either a goal is red or it could be anywhere and I’d always search for it explicitly.
I agree that my use is corner-case-y. Part of it is that I have relatively many goals, and so my main use for the order is to go over the long list of non-red goals I typically have, whereas a user who only has a handful of goals probably has a significant fraction of their goals in the red each day. This, combined with my minor use of goal deadlines (but not doing waterfalls or anything), makes sorting by safe days make more sense to me than by “urgency”. But I don’t claim that safe days are a better sort order for most of Beeminder’s users. (I guess that means it’s the kind of thing that should go in a browser extension!)
I would suspect the website (and apps) could offer another entry in the how-to-sort dropdown:
‘Urgency’ (via Beeminder’s special sauce)
Deadline (next upcoming deadline)
Pledge (amount, most to least)
Recent data
Goalslug (alphabetical)
For my use cases I would really appreciate a text input box I can use to enter some text of the slug to have those matching filtered. I would use that and prefer it over scrolling the list as I am usually looking for specific goals anyway.
As for the aforementioned sort strategies, I probably would prefer sorting say Pledge by most to least (highest pledges topmost) and would expect that tapping/clicking it again would then sort by pledge least to most.
That’s pretty much what the website sort columns are! The only substantive change to sort order with the change of column header from ‘deadline’ to ‘urgency’ is the interjection of any goals that Beeminder knows you missed entering data on yesterday (i.e. right now, just pessimistic presumption datapoints on do-less goals)
Also, @zzq’s extension lets you press ‘f’ to filter the goals shown, which to me is nicer than just using the browser search.
Note to self: I suspect we ought to move this sorting discussion to another thread, to leave this one for the extension.
This is nice as a kind of ‘uncle!’ button, but it doesn’t uncollapse new beemergencies any more, which means that some of my eep! goals were out of sight, way down the page this morning.
I was about to report back about the effect of this for me! Because I already have a workflow of hitting o to uncollapse everything at the start of the day, it hasn’t been a problem for me, and I’ve been pretty happily using it as a sort of uncle button. Also it reduces the amount I have to refresh my page (I don’t have to clear the beemergency, then refresh the page, in order to be able to collapse a completed goal). So for me being able to collapse goals in beemergency has been okay. (I can definitely see that this would depend strongly on workflow, though.)
I can never remember the keyboard shortcut for that. Nor for clearing the filter!
I’ve just now hacked the plugin to ignore collapsed status and just use urgency. This means that I see collapsed and uncollapsed goals intermingled, but that seems to be fine. We’ll see if I still like it tomorrow.
I have so much api-driven data that I constantly refresh the page anyway, and have recently set up a Keyboard Maestro macro to hit refresh whenever I switch back to the browser and it’s showing the gallery page.
This was exactly my original fear, the reason why four years ago I made the decision not to allow red goals to be collapsed…
You know, all this is making me want to do the unthinkable, and implement (heavens forbid), a settings page (the horror!) But that’s what happens when you start down the slippery slope of not being opinionated!
I have, on previous occasions, mentioned how the anti-magic principle and the anti-settings principle are often in strong conflict: either you have opinionated software that does something a certain way, just because that’s the opinion of the author even if it’s “magic”, or you have unopinionated software that allows the user to do whatever they like… with settings.
@dreev, this is the part where you try and convince me this is a bad idea. Still, it feels awkward to freeze in this in-between state, neither opinionated nor configurable, the worst of both worlds.
I suppose another in-between solution is contained in what Philip said, emphasis mine:
I don’t know if this is possible, but this would be my ideal solution, neither a setting nor a “my way or the highway” solution, but just, on reload, automatically uncollapsing something that’s changed to be in the red since the last page load, while still allowing it to be collapsed again.
Because of my workflow, I’m happy with the way it currently works (actually, it reduces a lot of annoying friction for me). But this solution might be even better, and I thought it was worth putting the thought out there.
I was mulling my response but maybe @shanaqui has just solved it for us? I’ll keep mulling the philosophical question of whether anti-settings and anti-magic are in tension. (I know we’ve debated before whether “minimize if-statements” could be the common thread in these Principles.)
What about this generalization and less if-statement-y version of @shanaqui’s proposal:
You can collapse any goal but every goal uncollapses when its number of safe days goes down.
I’ll have to see what behaviour I prefer in practice!
