> well-documented big catalogs are possible; they're just rare
They’re not that rare, it’s just that flat big catalogs don’t work well.
The most common failure mode for MCP server design is mapping MCP tools to system APIs. This results in a tool to list files, a tool to rename a file, a tool to copy a file, etc.
You can have 100 “tools”, you just need to define 10-12 areas (actual tools in the MCP model) and actions within each one. It creates a hierarchy that models can dive into when needed and avoid context clutter when that area is not needed.
So instead of 15 tools for file operations you have a single “file” tool, with actions like list, copy, rename, etc.
Single biggest arch requirement for any MCP server (well, any that has more than a handful of operations).
JSON and CSV [and a massive amount of custom-made spaghetti code] gave the impression that semantics was totally unnecessary in dataflows. Now that we live in a world where dataflows build themselves, we will need a description language that is much more robust.
Probably generated from chatbot interactions, but formally defining how the consumer can discover content and mix them.
This is stupid analysis.
Complaining without nuance about "D004 - parameter has no description" suggests the article was written by an AI agent, because some parameters simply don't need any description. Prime example is "url" parameter explicitly mentioned in the article - unless there are multiple possible url parameters, adding the description just bloats context.
I'm not going to standardize an MCP server. I come from designing REST API's and the leniency of Python. In comparison to other frameworks, MCP is relatively new and there's valid skepticism because of Anthropic's proprietary schemes. It sounds like AI companies are gatekeeping MCP servers, based on definitions they made up, and can just have their goons go around purity testing them.
> well-documented big catalogs are possible; they're just rare
They’re not that rare, it’s just that flat big catalogs don’t work well.
The most common failure mode for MCP server design is mapping MCP tools to system APIs. This results in a tool to list files, a tool to rename a file, a tool to copy a file, etc.
You can have 100 “tools”, you just need to define 10-12 areas (actual tools in the MCP model) and actions within each one. It creates a hierarchy that models can dive into when needed and avoid context clutter when that area is not needed.
So instead of 15 tools for file operations you have a single “file” tool, with actions like list, copy, rename, etc.
Single biggest arch requirement for any MCP server (well, any that has more than a handful of operations).
JSON and CSV [and a massive amount of custom-made spaghetti code] gave the impression that semantics was totally unnecessary in dataflows. Now that we live in a world where dataflows build themselves, we will need a description language that is much more robust. Probably generated from chatbot interactions, but formally defining how the consumer can discover content and mix them.
Somewhat ironically the link to your results table is broken
Remove the ] from then end
This is stupid analysis. Complaining without nuance about "D004 - parameter has no description" suggests the article was written by an AI agent, because some parameters simply don't need any description. Prime example is "url" parameter explicitly mentioned in the article - unless there are multiple possible url parameters, adding the description just bloats context.
I'm not going to standardize an MCP server. I come from designing REST API's and the leniency of Python. In comparison to other frameworks, MCP is relatively new and there's valid skepticism because of Anthropic's proprietary schemes. It sounds like AI companies are gatekeeping MCP servers, based on definitions they made up, and can just have their goons go around purity testing them.
Be warned, AI Slop ahead
Bro this is just slop.
[dead]
[flagged]
[flagged]
[flagged]
[flagged]
[flagged]