Where should new items created by extension/widget/API go?

andrew

The current popup pane (e.g., the one shown when invoking global search in a popup) is fixed and modal, meaning it’s not designed to be able to interact with the board/panes in the background while open so you can’t drag out of it. For it to function as an Inbox, it would need to be non-modal and movable so you can drag from it to a background pane. It would probably also need to be resizable so that you could see the target pane more easily. So not sure doing that is any easier than creating a fixed side panel, which would be easier to drag from as you wouldn’t need to mess with moving or resizing.

Jerud

andrew All great points, in line with what I mean about how popup panes seem like a promising start but definitely aren’t adequate for the job. I’d say they aren’t really adequate for much of anything as-is — my point is that they already exist and could be fixed up to be very useful indeed, both for this and numerous other uses. So it’s about getting the most bang for your development buck. I assume that modifying an existing feature is easier than creating a whole new one — but really have no idea what it takes to make any of this stuff happen!

Jerud

andrew The current popup pane (e.g., the one shown when invoking global search in a popup) is fixed and modal, meaning it’s not designed to be able to interact with the board/panes in the background while open so you can’t drag out of it.

I just reread this and realized your statement here is actually false, and I didn’t catch it.
I, at least, can absolutely drag items out of a popup pane, which then conveniently hides itself while I drop the item into a background Pane.

So in that regard as well, Popups could be leveraged to make a rather nifty “inbox” or “new item creation spot” or whatever we should call it. I agree that feature names which push users toward a specific use should be discouraged in Legend since it is oh so versatile and that is one of its strongest aspects.

Maybe something cute like “item nursery” or “item spawn point”.

mondayrobot

Hey, sorry to revive an old thread, but I’m wondering if this has been further developed? Thanks!

wecanseeformiles

I’d also be interested too. I like the idea of a calendar type inbox where every day is essentially a card that stores Legend items. Into this inbox could flow:

Emails dragged over from the email pane.
Web sites from the chrome/browser extension.
Links to local and cloud files (even better if we can save direct from the Windows file manager or Finder in Macos without having to drag the file over, or by clicking something in the originating app itself - so button in say Word that says save to Legend).
Dictated notes (using whisperai would be another bonus as some other apps like Reflect do.

Once in this inbox they could either be left after being prioritised or they could be dragged/linked/copied into other projects in Legend.

SamMarrocco

Hopefully I’m not too late on this….I missed the start of this thread!

Option 3 seems the closest to what I would like to see: I would also stay away from the term “Inbox” because it is too easily associated with email.

In Legend’s prefs, the user selected the Document (possible the entire item hierarchy?) into which all “External Item Adds” would go. For example:

Legend Prefs / New Items are sent to Document: SamIncomingItems
….or better still, adding the hierarchy with
SamIncomingItems / ToBeSorted

As far as sanity checking, don’t lock the document. If the user tries to delete a document, it should check the preference setting and provide a warning “This document is being used as the default for incoming items! Are you sure you want to delete it?”. If yes, the preference would be cleared and you lose that ability.

I’m not a big fan of adding “default inboxes”, and other things that clutter my organization of documents, panes, etc. I’d rather be able to determine said locations of things myself.

I look forward to an API implementation of adding items like this….Not a day goes by when I don’t hope for the mobile ability to say “Hey Siri, Legend add Mow the Lawn” without some discombobulated paid IFTTT hack.

justyannicc

Hi
Is there a date by when the API will be implemented? I cannot switch to Legend without it.

SamMarrocco

Another good thing about option 3 is that it pretty much lets you simulate 1 and 2 if you like those options.

3 is definitely the most versatile.

Jorge

So to revive @Jerud ’s comment… it has to be quick… Just the item entry description and possibly a note… and then Enter…. that’s it.
So I think Option #3 is the one that allows everyone to make it work to their taste.

justyannicc

This discussion has been going on for 3 years. I care less where it goes and more about the timeline on this? Will this be implemented anytime soon?

SamMarrocco

I won’t presume to speak directly for Jay, but if experience with organizer/outliner apps has taught me anything, it would be: Never ask for a release date or use the words “how soon?” Workflowy was talking about an API for several years and people ‘reverse engineered’ their own in frustration of their lack of development speed….which works great, until the developer of the app changes something and breaks your homebrewed api.

APIs are usually the last things to be implemented by small development teams (if ever), and even when they are, you end up waiting for other developers to decide that the user base is worth the development time.

Users like myself that write a lot of code, aren’t usually numerous enough on an app to make it worth developing an API just for us. Ultimately it has to be worth it for Legend developers’ use in that it will help them when working on the app.

……or I could be wrong and Jay may just decide tomorrow to release a full API! 😉

Jay

I haven’t built an API yet partly because of what SamMarrocco said, developers who need an API at least seem to be a small percentage of the user base. Based on the activity on the forum and all the emails I get, the top priorities for people right now are fixing bugs, making email plugins more stable/useful and supporting more services (the Browser pane project), making it easier to use for new users, and improving the mobile app.

The other main blocker is because of Legend’s end-to-end encryption, all of the data is encrypted in the database, so an API client would need to be able to decrypt it. So we can’t just publish an endpoint, we’d also need to provide a library which detects user passwords and document passwords, asks for those passwords, and manages the encryption keys. That’s certainly doable but just a lot more work than a normal API.

There may be some easier alternatives though, like to make it easier to work with the cached files, or have the app open up a local webserver with an API so you could could communicate directly with the app instance?

For those of you who want/need an API, could you tell me more about what you’d want to do with it? It may be that the target use cases could suggest an easier solution…

justyannicc

Developers don’t necessarily need to support it, you can build your own integration, and tools like Zapier and n8n make this even easier. All I want is the opportunity to do that.

I don’t care if nobody ever build anything on top of the API, as long as I have access, I can build my own integration.

And here is a crazy idea. Open source the code, so the people that want it can build it themselves.

I am just saying, to me, the app is unusable without an API.

SamMarrocco

justyannicc

Understood. And I totally understand why you would like it. Please don’t take any offense when I say this, but a person who doesn’t need developers to support integration of an app, and can write their own code for personal use with an API and write their own integration, is in the vast, far minority of whom commercial apps are written for. I’m one of them, and I feel your pain in knowing you could make something better if you only had the ‘hooks’.

I also appreciate your suggestion for open source…..for which the merits of are a far different debate. But I’d speculate that the the odds of that happening are pretty slim. Open source only helps if your interests/needs are in line with what the various developers find interesting enough to be worth their time, unless you are going to fork someone else’s work for yourself, which leads to it’s own issues. I’ve been involved with FFMPEG and Thunderbird and their development for years. Getting bugs fixed or features added to open source can be like trying to get a bill through congress: It’s not easy, fast or pretty.

A few of us on these forums that have used or tried many ‘organizer/outline’ apps have learned this: Development of them is slow (for various valid reasons) and robust APIs are few and far between unless the apps are more expensive and constantly managed by larger companies that are very capital driven.

Maybe someone could stage a gofundme for specific Legend features. In today’s world, maybe that’s a viable option for faster development of features? Everyone pledges money to a feature, and the highest earning stuff gets pushed to the top of the to-do list. It’s an interesting thought.

justyannicc

Jay
I understand what you are saying, but to me I just need to API otherwise I cannot use legend. I have been looking for a notes app that I like for a long time. I hated obsidian because it was too complicated and the UI is bad, notion because it was not end-to-end encrypted, Anytype because it isn’t nearly ready, apple notes well is apple depended, and standard notes is to basic.

Then I discovered outliners and workflowy. I really loved it because that’s exactly how my brain works, but again no encryption, so I can’t use that. Workflowy comes with a Zapier integration, but the lack of encryption sucks, and even with the Chrome extension that can provide encryption, you can’t access encrypted things on mobile. I went looking for an alternative, and it took me a while to find legend, but I really really like it. I want to switch to it for everything because it could do everything I need it to do, and it works exactly like my brain. But I just can’t because it lacks some features that I need.

I currently use todoist for tasks, google calendar as my calendar, and reclaim to schedule things automatically. This setup does have flaws, like not being E2EE, but it comes with features I just cannot give up. I love todoist reminders at certain times and locations or when I leave a location and reclaims auto scheduling which works with todoist. If I were to use legend, its all in or nothing, because I do not want to have to manually transfer things. So an API would allow me to transfer things I would give a # like #todoist. It would transfer it into todoist and then schedule it with reclaim. It would also allow me to set reminders for the task based on time or location.

I am asking for an API because it is likely easier to implement than these other features. Reminders are likely easier, but the auto scheduling probably isn’t, and I do not even know if you have any interest in implementing it. To me the auto scheduling, in particular, is absolutely crucial, and I cannot live without.

TLDR: Time and Location Based Reminders and auto scheduling are missing features which can be bypassed with an API.

SamMarrocco

Jay

I truly believe that the vast majority of people who ‘need an API’ actually ‘want an API’ because of missing features.

To wit, I personally would want an API for only one reason: The ability to send things to Legend from outside the app. So in my case, a (relatively) simple feature would eliminate my need for an API. I can work around it, so I don’t truly need an API. Because Legend lacks iOS Siri support (i.e. Siri, “Legend Remind me to Mow The Lawn”). I used to have a server running a dotnet app that parsed my iOS Reminders app and sent them to a Dynalist/Workflowy “input email address”, because I could send a voice item to iOSReminders. So my app scraped my reminders for items containing the word “Workflowy/Dynalist”, as in “Workflowy Remind me to Mow the Lawn”. It worked, but required me running an app on my server computer 24/7. A lot of work for ‘want’ of a voice-add an item and save me some typing and navigating.

I’ve used many application APIs in the past. Adobe AfterEffects & Premiere, Autodesk Maya & Flame, The Foundry’s Nuke and other visual effects apps come to mind, using DotNet languages, Python, LUA (Ugh) and C++. The worst things about APIs is this:

The application’s API developers must make a full time job out of creating, maintaining, updating and bug fixing the API. This draws an incredible amount of resources away from the primary app development, and this is time better spent on bug fixes and features to the main app.
APIs (like general app development) are a lead weight that adds mass and drags down the app itself. Adding features then requires fixing APIs. Changing features requires changing the API. Otherwise the API follows the app for years with it’s own set of bugs that are never addressed (see any of the above applications for their own individual track records) and those are managed by large teams of developers. APIs don’t sell software; Features sell software. It’s an unfortunate reality to application development.
-Every time I’ve ‘mastered’ an API, the slightest change would break all the applications I had written that accessed it. This was incredibly frustrating and while my first thought is "damn API developers changed things on me’ in reality this has to be expected. Maintaining backward compatibility is impossible in a large developer environment; even less so in a small one.

By no means am I saying that having an API isn’t useful…..but its limited range of effect on the overall applications needs to be carefully weighed, which Jay already mentioned. I’d rather have the app developer handle development if possible.

Eik

Maybe the following approach works as an api alternative. Legend has a good integration with google calendar. So if you can integratie with google calender, you might have a good integration with Legend as well. For example, I’m on the microsoft platform, I created a webhook on power automate to except general ‘content’ in a JSON format. When this endpoint is called, a google calendar all day event is created in a special calender within my google account. This event is visible in Legend and I can drag en drop it to the right place in the outline. On my iphone I created a Shortcut on the Share Sheet which calls the webhook in power automate.

This falls in te category ‘advanced’, but maybe some of you get inspired by this approach.

« Previous Page