I've made a lot of progress. The most significant thing is certainly the demo! (https://www.ditabase.io/demo.html) I only knew vaguely how the actual product would look, and the demo forced me to put it to code.
This resulted in two important changes:
1. No more schemas, only objects
2. Dit is the real product
Objects take the place of schemas in every way. I can no longer picture what exactly you would want out of a schema that you can't get with an object. They cover schema relationships, and because of the way scripts work, every object can use a different language and it doesn't matter. Schemas seemed like a headache, but object orientation is very simple.
Then, the only logical delivery mechanism is a single file type like dit. Whatever you want that's related to DitaBase can be neatly packaged in a dit. That includes all the things I know about now and all the features I might want to add in the future.
The very next thing is certainly the MVP. This will consist of a CLI Linux app that validates dit files in much the same way the demo does. I am optimistic that I can release that by February 1st.
I am looking for feedback almost exclusively on the demo, and whether it makes sense to you as a first product. What projects can you picture using it for? What features seem make or break to you? Do you still think it's a terrible idea? I want to hear it all.
And to be clear, I realize that many of you are probably not my first users, but you are developers. In the future, I hope all of you will be users, so your feedback is still important.
1 comment
[ 1.9 ms ] story [ 12.2 ms ] threadIts been 2 and a half months since I posted https://news.ycombinator.com/item?id=21263839
I've made a lot of progress. The most significant thing is certainly the demo! (https://www.ditabase.io/demo.html) I only knew vaguely how the actual product would look, and the demo forced me to put it to code.
This resulted in two important changes: 1. No more schemas, only objects 2. Dit is the real product
Objects take the place of schemas in every way. I can no longer picture what exactly you would want out of a schema that you can't get with an object. They cover schema relationships, and because of the way scripts work, every object can use a different language and it doesn't matter. Schemas seemed like a headache, but object orientation is very simple.
Then, the only logical delivery mechanism is a single file type like dit. Whatever you want that's related to DitaBase can be neatly packaged in a dit. That includes all the things I know about now and all the features I might want to add in the future.
The very next thing is certainly the MVP. This will consist of a CLI Linux app that validates dit files in much the same way the demo does. I am optimistic that I can release that by February 1st.
I am looking for feedback almost exclusively on the demo, and whether it makes sense to you as a first product. What projects can you picture using it for? What features seem make or break to you? Do you still think it's a terrible idea? I want to hear it all.
And to be clear, I realize that many of you are probably not my first users, but you are developers. In the future, I hope all of you will be users, so your feedback is still important.
Please let me know what you think!