Video summary

Выводим Сову на чистую воду

Main summary

Key takeaways

Technology

Summary

The video is a technical rebuttal to criticism of Mol/MAM, the framework and ecosystem developed by Dmitry Korlovsky. Rather than reviewing another video, Korlovsky reads and responds to claims made in a Telegram discussion, especially those attributed to Sergei Sova. The discussion also includes comments from community member Kirill.

Performance and rendering

  • Korlovsky distinguishes raw rendering speed from efficiency and perceived responsiveness. An interface can feel faster by avoiding work the user does not need to see—for example, by rendering only visible rows and adding more as the user scrolls.
  • He says Mol’s benchmark results do not show that it executes faster than native JavaScript. Instead, he argues that Mol can avoid unnecessary work, including through virtualization.
  • He compares this approach to React’s virtual DOM diffing, while arguing that Mol’s reactive system can address some React problems more directly.

Mol/MAM architecture and ecosystem

  • Korlovsky acknowledges that Mol’s ecosystem is much smaller than the combined ecosystems of major frameworks. He argues that its libraries are, on average, more carefully designed, and that many general-purpose JavaScript libraries can still be reused.
  • He describes MAM as a module system that can express dependencies across different kinds of assets and languages, such as JavaScript and CSS. Its aim is to include only modules that are actually used, rather than relying on bundler tree-shaking to remove unused code.
  • He describes VerLess as an approach in which dependencies use the latest version rather than allowing multiple versions of the same library to coexist. He argues that this helps avoid duplicated packages, while acknowledging that compatibility and dependency-management concerns still need to be handled.
  • In contrast to frameworks such as Next.js, Korlovsky characterizes Mol as allowing the frontend and backend to be developed separately while still permitting shared code. He argues that tightly coupling them can complicate support for other clients, such as mobile apps.

Components, customization, and compatibility

  • Korlovsky says Mol supports both composition and inheritance. Components can be customized through named subcomponents and method overrides, rather than relying only on slots or render props.
  • He argues that major changes should generally be introduced as new components or versions, rather than silently changing an existing component in a way that breaks users. Static typing, tests, and continuous integration are presented as ways to detect incompatibilities.
  • The discussion identifies a real limitation: accidental overrides can occur in class-based systems when a new member shares a name with one in a base class. Korlovsky says this is not unique to Mol and discusses possible syntax changes, but says they have not been prioritized.
  • He contrasts this with projects that keep old dependency versions pinned for long periods and then encounter many incompatibilities during a large upgrade. He argues that continuous integration should test current dependencies together.

Tree and View3

  • Korlovsky describes Tree as a format for representing syntax trees, and View3 as a language built on that format for defining components. He says the declarative syntax can generate JavaScript, TypeScript typings, localization data, and other outputs.
  • He presents View3 as supporting concise component declarations and bindings, including two-way binding. He argues that the syntax is unfamiliar to critics who primarily work in React, rather than inherently difficult to read.
  • A key feature he highlights is retaining source coordinates on syntax-tree nodes. This can help trace errors back to their original locations, including errors found after parsing or transformation.
  • He also acknowledges a TypeScript type-inference workaround that sometimes requires generic parameters to be stated explicitly.

Dependency and versioning arguments

  • Korlovsky argues that pinning package versions does not by itself prevent supply-chain attacks: a pinned dependency can still contain malicious code, and harmful packages can arrive transitively.
  • He emphasizes testing, static analysis, and evaluating dependency provenance and trust. He also distinguishes reproducible builds from continuous integration: reproducing a build does not, by itself, test whether the latest compatible dependencies work together.
  • The video refers to Mol’s approaches to SemVer, VerLess, fully qualified names (FQNs), imports, namespaces, and ESM. Korlovsky argues that FQNs and MAM offer different benefits from conventional imports and JavaScript-only module systems.

Code reviews and broader claims

  • Korlovsky briefly inspects React-based project code and Mantine components. He criticizes the complexity of a button implementation, its large number of props and style variables, and a custom local-storage hook that, in his view, does not respond to changes made in another browser tab.
  • He also criticizes the use of Tailwind CSS in an example project. These are Korlovsky’s assessments, not independent audits.
  • The video is partly about the technical disagreement and partly about the conduct of the Telegram discussion. Korlovsky says Sova’s claims were often unsupported or exaggerated and criticizes the community’s moderation. The video does not present a neutral evaluation of both sides.

Related reviews and guides mentioned

Korlovsky points viewers to separate material on:

  • React’s design and problems
  • Effector
  • Feature-Sliced Design
  • Tailwind CSS
  • Reactive programming and object-reactive programming
  • MAM and ESM
  • SemVer and VerLess
  • FQNs and imports
  • Link and routing formats
  • The Tree format

Main speakers and sources

  • Dmitry Korlovsky — The video’s speaker and maintainer/developer of Mol/MAM; he provides the technical rebuttal and code commentary.
  • Sergei Sova — The principal critic whose Telegram comments are discussed; his views are relayed and interpreted by Korlovsky.
  • Kirill — A community member whose discussion with Sova helped prompt the video.
  • Referenced materials — Mol/MAM documentation and articles, plus code from React-based projects and Mantine.

Rate this summary

Your feedback will help improve summaries.

Improve this summary

Reprocess with a stronger model when the summary feels incomplete or inaccurate.

Pro

Translate summary in another language

Pro

Ask questions to this video

Chat for follow-up questions, clarifications, and source-backed answers.

Coming soon

Share this summary

Original video