diff --git a/design/meps/mep-cangjie/MEP-CANGJIE.md b/design/meps/mep-cangjie/MEP-CANGJIE.md new file mode 100644 index 0000000..73be169 --- /dev/null +++ b/design/meps/mep-cangjie/MEP-CANGJIE.md @@ -0,0 +1,62 @@ +| title | authors | owning-sig | participating-sigs | status | creation-date | reviewers | approvers | stage | milestone | +| ----- | ------- | ---------- | ------------------ | ------ | ------------- |---------- | --------- | ----- | --------- | +| MEP-CANGJIE | @lijuan-hai | sig-cangjie | sig-cangjie | provisional/implementable/implemented/deferred/rejected/withdrawn/replaced | yyyy-mm-dd | TBD | TBD | alpha/beta/stable | alpha: "v0.5", beta: "v0.6", stable: "v1.0" | + +# MEP-CANGJIE: Cangjie with MindSpore + +## Table of Contents + + + +- [MEP-CANGJIE: Cangjie with MindSpore](#) + - [Summary](#summary) + - [Motivation](#motivation) + - [Goals](#goals) + - [Non-Goals](#non-goals) + + + +## Summary + +As reported, Cangjie programming language is Huawei's first self-developed programming language for the application layer. Cangjie language is also one of the front-end languages ​​supported by MindSpore. It has both development efficiency and operational performance, supports multiple devices in all scenarios, and has strong field expansion capabilities. HPC(high performance computing) and AI scenarios are one of the important scenarios supported by Cangjie language. + +The Cangjie SIG aims to establish connections with AI developers through the MindSpore community, help improve the usability and ease of use of the MindSpore front-end language (Cangjie), and continuously improve user satisfaction. The sig will focus primarily on the HPC and AI combined areas, including the following topics: + +- Discussion of issues/requirements and roadmap planning +- New features introduction and showcase +- Technologies sharing of AI development and applying +- Developer cases and practices sharing + +SIG is committed to helping developers easily learn Cangjie programming language and MindSpore, and use Cangjie to efficiently develop HPC and AI programs, so it will include functional optimization of related components, such as: + +- API documentation, user guides, tutorials and information experience on the official website +- Content quality of developer stories, industry best practices and advanced technical articles +- API design for ease of use, multi-OS (Windows/Linux/Mac) installation support +- Ability for bug reporting, self-debugging, programming language usage improvement suggestions, performance/accuracy tuning + +## Motivation + +- Discussion of issues/requirements and roadmap planning + + In order to allow developers to participate in the process of improving the functionality and ease of use of the Cangjie programming language, we need to actively collect first-hand questions and accurate information from developers as input to improve the design, such as API design, syntax support and debugging methods. + +- New features introduction and showcase + + To help developers stay abreast of improvements to the Cangjie programming language, we need to regularly introduce users to newly released features and show how it improves functionality and ease of use. + +- Technologies sharing of AI development and applying + + In order to help AI developers improve the end-to-end efficiency and quality of AI software development, we need to regularly share with them new AI technologies (such as AI engineering technologies) and the latest feature development progress of the Cangjie programming language. + +- Developer cases and practices sharing + + In order to spread the successful experience of developers, we encourage developers to share their technical practices and cases based on the Cangjie programming language, focusing on how to improve the scope of development efficiency and quality. + +### Goals + +- Help Cangjie programming language improve the completeness: establish channels to regularly listen to users' voices, collect users' first-hand questions and accurate information, and continuously improve the completeness and development efficiency of programming languages ​​by requirements-driven approaches. +- To make developers succeed: help Cangjie programming language developers improve the efficiency and quality of the development life cycle by providing comprehensive support in documentation, API and syntax integrity, error reporting accuracy, AI engineering methodologies, and best practices. + +### Non-Goals + +- Bugs and stability problems of each module