Notes on standardization/specification
otaylor at redhat.com
Thu Apr 29 18:15:56 EEST 2004
Just quickly writing up some thoughts I had on the standardization/
specification process from the discussion we just finished at XDevConf:
There are basically three different pieces that we might want to worry
about standardizing or specifying:
- Network protocols
- Inter-application protocols
The last is what has been most discussed/worked on in the context of
freedesktop.org over the last couple of years. This is the one where we
have a goal that really matches the IETF/W3C/etc: interoperability.
But for the other pieces - APIs and network protocols... I'd argue that
we really have different goals in mind:
- Ensuring API and ability stability
- Encouraging formal specification
- Encouraging finishing things. To avoid the 0.9 syndrome.
What I'd really encourage is that we need to make a "standards process"
that tries to bring these goals to the open source development model
rather than imposing some foreign preexisting structure. We can't
let our development be dominated by people who enjoy sitting on
- A lightweight standardization process is definitely need for
the interop specs; XEMBED, Desktop files, wm-spec, etc.
- Specification and test cases have to be cultural rather than imposed.
We need to write down a good best-practices document and ostracize
people who don't follow them.
- I think an architectural review board is essentially useless; review
needs to be done by people interested in the particular issues.
The way to get that review is to encourage that specification
- The goal of finishing off specs and APIs is best done via a "release
process"; have a boat that people have to get on or miss. Force
people to be part of the release to get to a 1.0.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.freedesktop.org/archives/xdg/attachments/20040429/dbcbc2f8/attachment.pgp
More information about the xdg