Fueling the need for a new Doclet API were several issues with the pre-existing Doclet API:

• It was not ideal for testing or concurrent usage. This stemmed from its implementation of static methods.
• The language model used in the API had several limitations and became more problematic with each successive Java upgrade.
• The API was inefficient, largely due to its heavy use of substring matching.
• There was no reference provided regarding the specific location of any given comment. This made diagnostics and troubleshooting difficult.