[WIP] Easy Pack Maker ⚡

I’m working on a doc that should tremendously speed up building new Coda packs.

So far it’s only a small proof of concept: a module that converts a JSON response from an endpoint into schema code:

Ultimately this will be a complete solution that will let you:

  • configure multiple schemas with custom mappings and Coda type parameters
  • specify API endpoints
  • set up those end points to function as formulas, actions, column formats, or sync tables
  • ultimately — generate dynamic sync tables as well

You’d be able to use it to:

  • quickly generate packs for yourself
  • make packs for the upcoming Pack Gallery to share or sell
  • make packs to apply for the Maker Fund :money_mouth_face:

The mission here is to make a pack generator that

  • for complicated APIs — will aid in generating the scaffolding, saving some good hours of time
  • for simple APIs — will be able to produce a pack with no further tinkering required (i.e. you won’t have to learn the SDK at all)

There’s probably some overlap with what @Bill_French is making here. I haven’t seen the actual solution though. Regardless, I have my own vision of how to build this. (You already know how meticulously I craft my docs, right? :slight_smile: )

The tool is going to be paid. I haven’t yet fully decided on the final monetization model — whether it will be a one-time fee, a flat monthly subscription, selling packages of #s of requests to the generator, or a mix of either.

You can purchase early access now though and get access to the tool perpetually with all further updates for $60 with no extra cost. Please :point_right: fill out this form and either:

  • subscribe to my Patreon with a $60 pledge (charged immediately; you can reduce or remove it before the end of the month)
  • specify that you wish to pay with credit card and I’ll send you a payment request through Payoneer shortly.

For everyone who’s been my Patron already: the tool will be available if your cumulative pledge amount over time is >= $60 already — no need to pay more :slight_smile:

Patreon is, of course, cooler because you’ll get access to some extra niceties. But if you want it simple and don’t care about any other content from me, the Payoneer way also works.

Cheers,
Paul

5 Likes

It’s not about how it begins; it’s about the entire journey and I predict your project will be a long and fun journey. I wish my day job afforded me the opportunity to devote significant energies to PackMule, but as the opening dialogue makes clear, I have no clue where my approach will lead me or at what pace. Easy Pack Maker already has a great vision and some very clever ideas - full steam ahead!

The journey you are on is inspirational and not without challenges. However, all that you’ve done with Coda and your long history of deep investigation into the pathways that you have discovered, serve as a powerful prologue to Easy Pack Maker. This makes you an ideal person to paint the vision and execute on it.

Very clever idea. This will save a lot of tedious development work in Pack Studio.

As it should be.

4 Likes

Thanks @Bill_French — you’re too kind as always. I’m still curious of what you would do though.

Be more specific. Concerning the commerce model?

No, I mean how exactly you’d build your Pack Mule. You seemed to have something done already. I submitted the application to “learn more” :wink:

PackMule is still just a few prototype misfits, but it has the basic stuff set up for arbitrary code generation. The intent is to first train it to create Pack-related code, and then maybe other types of code. But the premise is that Coda itself can serve as a formidable code pattern library that is easily extensible. Users would be able to use some initial patterns but then - with experience - create other patterns.

The real meat of PackMule (at the moment) is a Pack that transforms data (variables, schemas, etc) into Pack code that is in some cases, ready to execute. As such, PackMule is fundamentally a database (in Coda) that uses a custom pack to generate Pack code. I use the term “train” in every sense of the term. With some influence from GTP-3, I’m attempting to utilize AI to nudge the ability of PackMule to smartly create code. This is all hypothetical at the moment, but it’s the fun part of this project.

At my current pace, releasing this into the wild is many months away and demoing it to others is perhaps many weeks away. I’m very slammed at work, so the release timing of the Packs beta is coming at a time when I can least afford distractions. I would love to focus on this 200% but duty calls.

1 Like