Hello there, thanks for the interest!
I’m quite happy on how the development has progressed so far, but unfortunately it is a bit delayed at the moment due to life events. I would love to have something before eoy, at least a minimal version. But I cannot commit to any date.
I agree on the list view approach. However I wouldn’t be able to provide a full featured task management. I think having a list view complementary to the Kanban view would be quite useful indeed.
Hello, first of all, thanks for taking the time to test and leave very useful feedback! Really appreciated, I’ll try to reply below on some of your topics.
Team License: everyone on the team can try the demo, which is fully functional, for as long as they want without issues until you decide if it fits your workflow or not.
Timelines: Honestly, I’m considering removing timelines for v2 as they are poorly implemented and might be redundant with the node based dialogue tree. (which is far more powerful), alternatively I’m also peeking at some libraries like this or this
Sortable list of projects: Good idea! Added to the backlog.
Backup: You can backup the whole database of a project as a json file with the “Backup database” action, or just the game content with “Export game content”. The whole app data lives inside %appdata%/drafft or ~/.config/drafft folder. Although there is no supported or easy way to use that folder as a backup. Its better to use the previously named options. There is also the “Auto Export Project Backup “ to save a fresh db backup when loading a project. You can also get a db-dump via the basic API
Version control: this is tricky, as the database lives in the ‘browser’ as an in-memory database (not file based). However I have an idea on how to make a ‘semi portable’ version. this will DISABLE the real time editing benefits of course.
Better Task management: Big fan of ClickUp here! As much as I’d love to see a full featured task management with @mentions @assignees, notifications and the usual goodies we should expect from it, I’m afraid it would be a big effort to implement correctly, especially in this “self hosted” approach. What I will consider is to add a “list-based” basic task manager (emphasis on BASIC) to complement the Kanban view.
Streaming/recording: go for it! Also, if you have useful clips to share I would love to watch/link them on the site.
For the import process, so far most of the data works. The main idea on not “supporting” V1 is not to worry about backwards compatibility when adding more features. But I guess we will need some sort of “migration plan”. Unfortunately life got in the way so V2 is taking some time…
About the portable project, this is something I’ve been thinking for some time. The main problem with this is that the current database lives in the browser and we would have to “dump” it to have it in a file. I have A few ideas but not quite solid as a solution.
Maybe… the db gets ‘dumped’ to a file by default on project/app close. And a special flag on each project like “auto import from dump” would do this reimport thing automatically on load. Git operations like pull/push/merge would be responsabily of the user outside of the scope of the app . This would make a project also work on a file share or Dropbox-like app.
Thanks for the feedback!
I agree the software might not be for everyone, that’s why I decided to offer a full featured free demo, a shame you couldn’t try it beforehand. I think you can file a refund request on itch.io : https://itch.io/t/129454/how-are-refunds-handled
About the issues:
There was a bug with import json into new project that it fixed now in v1.4.4. On 1.4.3 you can create an empty project and restore the project into that. (should be available for download soon)
Make sure you are importing the json file, this was an issue for another user here
Thanks for the report, I can confirm about the edit bug in the timeline, it should serialize like the other actions when editing.
Honestly the whole timeline section is quite old, and might need a revamp/sunset, maybe its better/easier to use the branching tree which can have attached scripts and better ux…
About the folder, yes, Also if you have a document (not a folder) selected the new file/folder will be a sibling to that.
Hi! do you export with “Export Game Content” ? If so, you need to use “Backup Database” instead. The Export game content just export a json file intended to be used in the game engine. Let me know!
Hello, hm this is strange, so you have a file locked, and when the app syncs it overrides those changes with old version? does this happen always?
when you say sync, its the initial sync setup you do as a user, or the sync that the app does in the background?
Hi! the implementation is up to the developer, the app will only output a JSON file, or could be queried via a rest API. So the I guess the answer to the question is none, but can potentially work with any engine.
Not really, unfortunately life got in the way. But it is the first thing I’m fixing when I can get back at it.
There is something wrong about the db attachments / electron builder / windows
Hi! Actually I had a regression with that as I’m not very happy with the markdown editor and going back and forth with some tests.
In essence I’m stuck now in deciding if the attachment should be serialized in base64 inside the doc as before, or do a proper attachment database, which will allow the users to sync any kind of binaries, even outside of the scope of the GDD text (a video file could be a GDD entry for example,,)
This however, brings the question if it is worth the effort, or the potential performance implications (using a db as a file repository ins not the best idea) and if it is, maybe it would make sense to migrate to a file based repository instead of the database approach…
With this idea, drafft could support any “file” and a simple file/folder hierarchy. Multi user sync would be done with third party apps like dropbox, git or similar… Integration with the game engine would be easier, even the game assets could be centralized and taken directly from the workspace…
Sorry for the brain dump!
As for reference, I have a version in dev which allow to reference any drafft object with a internal link (drafft://actor/actorname drafft://script/id_or_alias) etc, but I got a bit stuck with the previous topic that is related
Sorry for the inconvenience, and thanks for the tests!!!
I will try to reproduce in a local environment and find a fix. It is definitely related to the electron build/packaging process and how it is retrieving the files.
Hello! Not much really, just simple linear timelines/events/etc it is a really old feature but might be useful for mapping out a story. But honestly maybe the node tree editor is more versatile.
Thanks for the follow up!
Hmmmm this is strange, are you using the portable windows version right?
It seems that it is failing to get to the example jpg from the filesystem for a the new actor, for some reason, probably related to the Windows path filesystem and how it is resolved when packaged.
I have a theory, on what could be causing it, does it happen as well in the installer version?
Hello! Best I can offer right now is a link to the docs repo. You can download a zip and inspect the .md files. Not what you are asking, but better than nothing I suppose.
Hello Thanks for the feedback!
Unfortunately I cannot reproduce #1 and #3, I’m in Linux. Are you in the latest version?
If so, can you check if you get an error in the console when adding the actor and when auto-detecting? (ctrl + shift + I). I suspect is something related to the blob attachment.
About the export, I found a bug and fixed in 1.4.1, thanks! The quick export will start to working after an export has been made. (its the same command without the file picker)
Hello! Some updates about this feature, which is available in v1.4.0
It’s not exactly what we had in mind at first but might be helpful.
The problem of running headless is that the local pouchdb is stored in the browser. Also electron requires a display to run headless so it a bit tricky.
So for now, we have a REST API that runs when a project is loaded (if enabled in the options) that can get you a game export and a database dump.
I’m planning to make a little cli app does the fetching but any standard GET request method would work. This tool could probably have some mechanism to dump from a remote couchdb instance as well.
Well the ::ActorName:: tag Is supposed to be unique, meaning, the app would autodetect and group the same string into the same actor. It is, in practice, an ID.
I think the “real name” should be inside the properties, but again, is up to your implementation.
Hope it makes sense!
Fixed in v1.3.0 live next week. Only the max-size, getting it to use the whole space with empty rows is more challenging than expected!!. Thanks for reporting!
Hello there!
The implementation is really up to you! I would parse the graph.nodes array, locate the “start” node (or any other useful node) and start crawling from there.
Some useful props:
In the extras property you have the type, and the scriptID associated with that node.
Inside each node you have a ports array which you can use to get all in and “out” ports (1 for script nodes) so you can build your flow with that.
Once you reach a choice node, you can query its “out” ports to show the different options to the player, the link array associated with that port will give you the id of each connection the port has, so you can navigate to the next node.
Basically what you have there is the whole graph serialized.
The dialogue player is a (super simple) example on how to implement.
Hey, just to let you know that this on my radar, but I was thinking, instead of having a cli command (which may be really helpful with CI/CD and scripts, but may require extra work to run headless)
What about having a very simple API that returns the JSON, (and maybe some endpoint to force a sync)
The caveat is that in it will require that the user have the app open.
Also, this would mean that this API could be potentially hit from any computer (with network access)
Also, this would mean that, it could be possible in theory, and some imagination… to hook up the game engine to use the API directly for some live preview goodness.
Is not a headless/“offline” solution but its definitely simpler to implement!
Hi, thanks for the suggestion! Including TOML will be pretty easy as it is already a mode in ACE editor (which I’m using) I can add it in next release!
Next release its taking some time though as I’m doing some refactoring (mainly trying to add a more performant MD editor)
