Hi all,

Been a while since I've circulated in Freenet circles - largely due to
the fact that the local telco's stranglehold on the local loop meant
that it was costing me $80/month to run a working freenet node.

Anyway - to the matter at hand. I want to propose the inclusion of an
XML-RPC server within the core node code, like back in the 0.3 days.

My arguments in favour of this:

 1. small amount of code needed, can be a thin wrapper around the
    internal Java client API, java's already got good xml-rpc libs

 2. relatively easy to maintain, only needs touch-ups here and there
    if the client API changes

 3. it's much less work for a client writer to code to an XML-RPC
    interface than to FCP v2

 4. it's less work to write one xml-rpc layer than to write n
    FCP v2 libs, one for each language

 5. any changes in the java client API can be largely absorbed into
    the exposed XML-RPC objects, unless there are major new features

 6. presence of a client xml-rpc interface will reduce the collective
    coding burden shared between core devs, lib authors and client
    authors

I've got enough java knowledge to be able to write an xml-rpc layer
myself (though I'm mainly a Python coder), so would be willing to build
something once 0.7 and its client API stabilise somewhat.

Lastly - would people see value in a (long overdue) freeweb rewrite? I'd
be looking to implement in wx/Python this time. (recall that the old
version was windows/mfc/C++).

-- 
Cheers
David


Reply via email to