On 16 Oct 2008, at 16:17, Malcolm Wallace wrote:
Anthony's original message, containing his wish list for Hat:
"A technical document defining the format of Hat traces, so that implementors of Haskell can add native support for Hat to their respective systems."
It also solves the problems of:
* maintaining and distributing a separate tracing version of the Haskell codebase;
* maintaining a separate Haskell source-source transformer and front-end.
Some extra benefits:
* Traces coming from Haskell implementations would be from well-typed programs;
* The somewhat-invasive nature of Hat on a Haskell codebase could be considered a deterrent to its widespread use; this would no longer be the case.
I understand that native support for Hat will be part of yhc; this would be a good opportunity for "Hat insiders" to put together such a document - the implemetation in yhc could then be used to vaildate it.
FYI, My thesis (just submitted) has a decent sized section on this in it. Whether it passes and makes it to general publication is ofc another matter. Bob