Links and Notes - April 13th 2021

Obsidian plugin idea: Publish to ghost from Obsidian

I'm definitely running into a pain point with my workflow of writing my posts on mobile/desktop Obsidian before publishing it via my ghost blog instance. The experience of crafting the post within Obsidian is excellent. I'm not going to swap that for anything. But publishing? Lawds. Copy. Paste. Add in summary. Select tags. It's a pain when trying to do daily posts. Especially when the finishing touches on the daily posts are so often done via Mobile and the mobile experience of ghost is not great

I could get one of those apps that allow creating content in markdown which can then be published using the ghost API. But another app? With probably differing markdown support? Really?

What I want is the simplicity of an Obsidian plugin. I add the API token to the plug-in settings, and when I'm done creating my post, I just hit publish from within the Obsidian command palette.

This has a lot of potential for me. For one, I use templates heavily when creating stuff in Obsidian. My blog post templates include the title, the summary, the actual post itself, all the stuff that might be needed by a script to generate the values needed for an API.

But if I wanted to take it further, I could also enhance my flow of linking to past flows. Right now, I include the link of each published post within each file as part of my front matter fields. If I used Obsidian's internal linking to link to past posts or even headers within those posts, it would be somewhat trivial to write code to convert those wikilinks to web links before publishing. This would ease my current workflow of copy pasting URLs and grabbing header IDs to link directly to headers.

Add this to another idea I desperately want to try

Gitlet - JavaScript implementation of git with beautiful comments

Via HN

Came across this this morning. It's exactly what it says it is. Git implemented in JS. But to be honest, that's only a small part of what makes this so interesting. It's the beauty in the way in which this code has been presented that makes it a topic of fascination.

It's not just me either. Check the HN thread. Quite a few commenters are amazed by the way the code has been rendered as part of an HTML page here and each block of code has some prose describing its functionality and rationale. It's even more fascinating once I take a look at the underlying source code file itself. Each section of code is followed by several lines of comments (written in markdown style).

According to several HN commenters this is apparently a concept called literate programming described by Donald Knuth. To be honest, the above link has content that's far too heavy for my liking. I really prefer what I'm seeing in the main Gitlet link. Conceptually I really get it seeing it in practice like that. And obviously, this makes me want to try it out too.

I'm not blind to the biggest risk of documentation rot. In fact, the more prose, the harder it might be to "refactor". But the truth is, documentation rot always happens anyways. The alternative to this is self documenting code. Truth is, I've rarely seen that either. I've seen well named variables. But understanding what a random section of code never feels like it's been self documented. In some ways, self documenting code feels like the holy grail. Always sought. Never found.

Therefore, this idea seems like an ok compromise. And one I'm eager to try in my newer projects.

Moana and character animation

In 2014, Dreamworks' How to Train your dragon 2 came out. I remember being blown away watching it in the cinema. Everything in the character animation felt like such a leap forward. The environment was definitely a leap forward as well — the water sprays had me wide eyed at the time — but it was the character animation that got me.

It's not like it was life-like. That would actually be undesirable. Exaggeration is what brings life to animation. But within the exaggeration it felt like someone had found a dial of "micro actions" and spun it up a 100x. The faces and the bodies with all the muscle twitches looked so expressive that I wondered if we'd reached a significant peak in this area.

This week I watched Moana. Moana came out in 2016. This week I was once again blown away with a similar feeling of amazement at how good the character animations looked. I even pulled up 2015's Frozen to see if I was misremembering anything and I wasn't. Moana had that similar leap that HTTYD 2 did. I wonder if Disney/Pixar had to push themselves to match the animation levels after 2014 or if it was just inevitable that the studios would just get this good eventually. Admittedly, Onward was missing that feeling so maybe it's just a directorial choice.

Whatever it is, I'm suddenly feeling excited to catch up on animated films that I haven't seen in the past years. I've not seen Frozen 2. I've not seen Soul. I've missed Toy Story 3. And I've not seen how to train your dragon 3!!! I believe I've missed out on most animated films since 2015. Now all I want to do is sit down and make notes while I compare animation across years, studios, and genres.

This blog doesn't have a comment box. But I'd love to hear any thoughts y'all might have. Send them to [email protected]

Previous links and notes (April 12th)

Next links and note (April 14th)

Posted on April 13 2021 by Adnan Issadeen