Dev Notes 23: Slight Additional Progress on Site Meta

Dev Notes (DN) document progress towards other articles. Discussion is preliminary.

These notes document small, but continued progress on site meta stuff, continuing off of DN-22 for a newsletter and comment section.

Newsletter

I'm still waiting on SES approval. However, I did try to figure out how to get listmonk to make multiple levels of lists. Here's my working levels and descriptions.

• Main: Main articles with data-focused insight or research. ~0-3 per month.
• Deci: The main articles + smaller articles with some smaller point; also possibly early drafts of main articles. ~0-5 per month.
• Centi: All articles + small dev notes. ~0-30 times a month. Dev notes are a nitty-gritty details on how the articles come together. Centi is not the recommended list level for most people.

That's the current idea. I'm still conflicted on whether I want to let people be able to subscribe to these notes. These are supposed to stay informal.

I made a web component for this which lets you sign up and select between these. However, to avoid confusion I won't put it here (it currently ju is broken without SES approval).

So checking back in on my todo list here:

• ✅ Decide a stack -> Listmonk/Lightsail/AWS SES
• ✅ Get server running
• ✅ Demonstrate toy sending to myself
• ⏳ Get production approval from AWS SES
• ✅ Experiment with different lists for different amounts of posts (eg only main posts, or more)
• ✅ Make some kind of signup component for putting at the bottom of articles
• Internal self testing with component

Then nonblocking stuff:

• Customize the signup confirm email and templates
• Rendering from article MDX to newsletter HTML (not necessarily blocking given can just hack a notification email with just like "here's a post" even without the full text)
• Maintainability steps (also not blocking. the list can die)
  • Configure db backups to S3
  • Configure auto recover if the instance fails
• Cleaning up the landing signup to redirect directly to the site

So through the MVP here once SES gets approved.

Overscoping on Comments Feature

In DN-22 I discussed a vision for comment section that is sourced from across the internet. The goal is to encourage general discussion on the web rather than in just a walled garden comment section.

I spent some time today probably wildly overscoping this. I think it could be fairly cool, and others might want it on their own site. It would work not just for blogs, but can work on things like product pages or general pages and such where it provides social proof and interaction for the product. If this is the case, it is possibly a small product. Or an open source tool.

I spent a bit of time researching the market some more. Also, what platforms it would actually be embeddable in. There's the standard static site generators like Ghost, Jekyll that can mostly work. Then there's more closed platforms like Substack. Substack doesn't actually support embedding, but it does support interactive visualizations via Datawrapper. I think there would be a way to feed the internet social activity into one of these datawrapper tables which would dynamically update. Would be kind of cool.

I also thought some about pricing if I were to actually make it a product. There's a few options here. I also am interested in a "revving" model which like an engine can go up or down. Basically want a pricing model that can scale down to zero for some blogs. Personally I am pretty hesitant to sign up for subscriptions for kind of side things like a blog. I don't want to accumulate subscriptions that are hard to get off of, and in a cases where I go months without blogging, would just waste money. I suspect many people have similar feelings. So I think kind of tiers of pricing that rev up and down. Like a $0-5/month tier, a$2-20/month tier, etc. Depending on number of page views it revs up or down up to your selected tier. "Don't post anything for two months and only get a few hundred page views? Np, we scale pricing down to zero." I'd personally like the model as a customer. But it is also untraditional and might be too confusing / not the most sustainable pricing.

I'm not really sure. I need to probably just build the hacky version to see how this works first. Not clear I want to deal with a public version anyways, and just need to see how cool it seems personally. I didn't get far, but got a repo started to start hacking.

Conclusion

This meta stuff is not super interesting, but glad to make a little bit of incremental progress. Tomorrow I'll try to mostly set this meta stuff aside for a bit. I hope to revisit the things with Wikidata for one day. Then thursday and friday I'll try to get the next IMDb post before the end of the week.