Looking fresh as always! Can't wait to see the whole redesign roll out

I think a computer trying to figure out how to align something optically is going to be very difficult until we have AI with context awareness.

There are many methods for finding the centroid of an object, all expressed in simple algorithms. And I don't think people would rely on it as a crutch — it would be an affordance to be able to override the precise position of a shape by being able to express "align this shape 5% south of the centroid" rather than "set this polygon bound by a rectangle to pixel coordinates 400x666".

Most of the time though, it would just look correct, and good defaults are the hallmark of good software.

You don't see this as a deficiency in existing tools though? What's preventing our tools from being more mathematically aware of the relationship between shapes?

I think it would be nice to reify the idea of a weight, to have the program guess where it lies in some geometry, and still give the user the ability to nudge where necessary. But there's a big difference between nudging some pixels, and nudging a useful, higher-level abstraction.

A big consideration is "how much are you duplicating?"

There's a big divide between template-oriented tools and component-oriented tools.

Template-based tools are usually page focused — whether this is Rails or Jekyll or whatever — there's usually a view representing the current page, calls into partials or template helpers for items on that page, and some series of nested layouts that that page is rendered inside of.

Component-based tools focus on the composition of UI elements. Angular element directives, React components, Ember components — apps are build out of more-specific components consuming more general components atomically, until you have a page.

Traditionally the argument has been "well how much interactivity do you need?", but I think the question needs to start being "how much are your styles something you're duck-typing to from a hodgepodge of server side templating and client-side UI frameworks?"

The answer is always "it depends", but there's a lot of appeal to the self-contained component thing as a building block for web publishing, whether it's a static site or a rich app. React is especially suited to this, as a tree of React components can be rendered either to a live DOM, or as a big string.

Lukas Mathis' Designed for Use comes highly recommended for interaction design. It reads like a case study for Design of Everyday Things applied to software.

So what are all the other mediums you'd like to work in? Who are the outside-of-the-tech-industry people you'd like to work with?

You're trying to bridge print and web publishing with Another, the print → digital → print → ∞ thing with the Great Discontent, the web/book/kindle/tote/t-shirt spectrum on Shape of Design.

It's not all about finding common denominators, I'd imagine, a lot of it has to be about finding the right collaborator, the right subject, or the right audience.

Type inference in this case is a compiler concern.

I believe they used to have a great looping video on the homepage showing it off. It's not clear why they redesigned away from that, it easily got the point across.

Another killer feature they removed early on: the ability to share individual apps. You could share your text editor, and the receiver would only see that application's window — not your entire desktop. It'd even get a place in your dock. Not sure why they removed it, maybe something to do with Windows support. It's too bad, it was truly awesome.

It's still the best desktop sharing tool by far, but I think they've made some questionable design choices lately.

If only they'd contain the page to a reasonable width

Hell, let's just have business as a service. Why bother to analyze your market and figure out how to generate revenue if you can just have someone else do it all?

The sentiment toward the end of this article is great though: we need better tools for creating things. We haven't even solved the problem of presenting text. Sure, many might feel enlightened by Markdown or Latex and other plain-text formatting solutions, but look at how much of the internet's net inputs are in Wordpress' TinyMCE or other terrible wysiwyg interfaces? How many meaningful works are copy-and-pasted from Word or my-website-copy-revision-6.pptx?

If, however, you feel that your blog, your words, and your brand are liquid enough assets to farm out to a blackbox "content generator", what are you doing other than making noise? How are you doing anything other than adding to the entropy of the internet? The problem really isn't the tools, but the attention culture, pageview model, and business-as-abstraction motives. They're tells of companies who have either grown too large, or are trying to grow too fast, to understand what their core value is.