I don't love that, for Romanov and his collaborators (and for obvious reasons), an actual digital text of one work is still just an avatar – with the status of "digital metaobject" not being attained until the corpus level.

Every college member is equipped with advanced training in the Art of the Metaobject Protocol as a matter of basic course, making even their weakest student mortally dangerous to the average person

I wonder if I can dig up the stupid metaobject protocol thing I did in Clojure

I basically share that thesis after using llms for both corporate clojure and my own custom metaobject protocol-based js framework. The main caveat being, there’s a good deal more friction than just letting it rip with bog standard typescript/go/whatever.

Branding api changes, sometimes metafield / metaobject definition creation on stores (getting ai to generate the graphql for this, then can replicate on multiple stores). Some extensions that still need enabling via graphql

Looking forward to the updates. By custom fields, I mean Shopify catalog fields, specifically list-type metafields with metaobject references like fabric, care, and instructions

（个人见解） 我想咱们都知道js使用prototype来定义或改变实例的行为，就像我们常用的给一个类增加方法的做法是this.constructor.prototype.<name> = function(){...}。我很喜欢这一点，因为这是元对象协议（MOP）的一种典型实现。 元对象协议的核心是：提供一套接口，允许程序员控制一个对象从创建到销毁的整个生命周期的完整过程；其副产物是动态派生新类、反射、对属性读写的完全控制等。 js有能力实现完整的元对象协议（即使在不使用eval和Function的条件下），但它没有提供标准化的接口，令人惋惜。比如咱们知道，类本身也是对象，在js里它是我们的constructor函数。这个函数对象的属性和方法就是类的静态属性和静态方法。constructor prototype给我们提供了控制对象的属性和方法的办法，但是距离“整个生命周期”还有两个关键步骤没覆盖到：新实例的初始化，和类对象的派生。它们没有公开的原语。 要在js里实现控制实例初始化，我用过Proxy，可以实现所有需要的功能（填值、挂钩子、链接外部环境之类的）。派生实在是没招了，当时用了工厂模式，和原生语法说再见。现在想想有点偷懒了，Proxy 反射应该是能做到的，就是工作量有亿点大（ 有了元对象协议（哪怕是部分的），就可以实现一些有趣的功能，比如改变一个实例的类型（change class），或重新初始化一个对象（reinitialize instance）。可惜js也没有给这些过程提供公开原语，可能确实觉得自家程序员用不到这些。。。 元对象协议（MOP）最初就是由《The Art of the Metaobject Protocol》（AMOP）这本书提出的，可惜那帮研究员写书的水平真的是稀烂，配合ADHD根本读不进一点，我最后是干啃spec学会的，，，MOP为高级面向对象开发提供了至关重要的抽象层，几乎所有依赖OOP的场景都需要一定深度的元对象协议，例如所有大型GUI框架都实现了自己的阉割版MOP（包括objc、gtk、qt等）。Java、Python、perl等解释型语言也都有较为完善的MOP实现。 我很喜欢MOP，所以会提到这些东西。个人见解或有疏漏，写这么多也算是安利自己喜欢的东西了qwq

唉唉The Art of the Metaobject Protocol 操作class prototype控制所有instance的行为也是MOP的一部分，和操作元类来控制类对象（js的constructor，可惜在js里我们只能proxy）的行为一样。可惜在prototype的基础上有些东西已经唾手可得，例如change class和reinitialize instance，js却止步于此了

See what else shipped in March: • Faster Admin API queries • Simpler metaobject scopes • Admin intents for Settings, and more community.shopify.dev/t/what…

This week we're shipping: - Live previews for merchandising. - Infinite nested metaobject support. - Change featured media in merchandising. - MCP server. Then we begin rebuilding our dashboard from the ground up with beautiful designs from our product designer who joined in Feb.

you can also import/export metaobject definitions (and entries) to copy to another store, easily replicate your most common usecases, copy to/from development stores

I wish we could edit Shopifys official metaobjects for color. Merchant has a use case where they want to attach supplier information to colors that are shared across products/variants. Current solution is variant metafield that references a Metaobject w that info. But it’d be SO nice to add a “Supplier Info” reference to the color Metaobject

That's big accusation towards Python, you know ;-) Part of why it works better in Ruby is that you get exposed a proper metaobject system, not accidentally leaked internals. Similar to how Smalltalk and Common Lisp MOP did it.

Yes, it would be great to have a native picker. I am using Metaobjects Admin edit intents to allow the user to pick an image from the metaobject edit screen. This is currently the only way to avoid building a custom picker.

Is it possible to compare a metaobject end date to today in Shopify Flow? Sidekick is just making shit up and can't seem to generate it.

Is anyone else experiencing this UX problem with Shopify admin intents - for metaobject edit?

Using @benjaminsehl's Liquid Skills I just one-shot the rewrite of a nasty metaobject menu to Nested Blocks with Caude Code (opus 4.6)

1950s Homoiconicity, REPLs 1960s Incremental Compilation? 1970s Call/cc Actors 1990s The Art of the Metaobject Protocol 2000s Language-oriented programming (Racket)

Ok, so say I had a list of 250 things in JSON today. These things are related to that metaobject. How do I structure this in the JSON-less world with metaobjects so that I can retrieve them on the storefront?

I think what we need, to make the transition from JSON to metaobjects: 1.) A better way to create nested metaobjects via the admin API. 2.) AND/OR - a way to batch metaobject creates into a transaction, such that things changes be rolled back if anything fails.