[
Lists Home |
Date Index |
Thread Index
]
>
>
> What is holding Atom up? Where is in its growth to final (too early
> <--> very close)?
About all I know about Atom is what I read on xml-dev, so take this
with a huge dollop of salt: I'd say that anyone reading this thread
should get an appreciation for how *hard* it is to develop good
standards. It seems so easy from the outside, and after a few weeks
your friends stop finding the obvious bugs ... but then you run across
use cases that seem reasonable but are at odds with your original
assumptions, you realize that prose that seems crystal clear the
original group of people is interpreted very differently by outsiders,
you wallow in the infinite complexities of error handling in a world
that doesn't tend to accept the draconian principle, and ultimately you
have to deal with the slimeballs who find it amusing to leverage your
spec to raise havoc (or sell their fake Viagra, or whatever).
There are several real dilemmas -- standards should describe best
practice, but best practice is intimately tied up in adherence to
standards; standards are intrinsically about interoperability, but the
goal of any business is to differentiate itself from competitors; the
process of building a good standard requires lots of creativity by
individuals, but also a willingness to form a group consensus.
Finally, standards need be stable to be acceptd, but also need to
evolve as the world changes -- gracefully when possible, but by
extinction of a spec that has outlived its usefulness and ability to
accommodate change when necessary.
I have no idea when Atom 1.0 will be "final". I doubt if anyone else
does either. The main thing I like about it compared to RSS 2.0 is
that it doesn't *claim* to be frozen; its owners implicitly acknowledge
how hard this process is and that all versions are interim versions,
subject to revision on the basis of whether they really work for real
problems faced by real people.
|
