[
Lists Home |
Date Index |
Thread Index
]
Danny Ayers wrote:
> (Bob Wyman suggests [1] Infoset-based binary
> serialization of Atom syndication format,
> providing a schema in ASN.1 [2])
Actually, the ASN.1 schema I offered (only a draft, it still
needs work) is really FYI only. Any XSD schema for Atom would serve
just as well as an ASN.1 schema to drive ASN.1 encoders/decoders. This
is because the mapping from XML Schema to ASN.1 has been defined as a
standard by ITU-T and ISO. So, providing a schema in the ASN.1
notation isn't necessary to define a binary encoding of Atom. As soon
as there is an XSD defined, the schema-specific binary encodings of
Atom are defined. i.e. "Binary Atom" (or a binary encoding which is
semantically equivelant to Atom) is already defined whether or not
anyone thought they were doing it...
> It's not hard to imagine the uber-aggregators and
> mega-aggregators communicating with each other in
> this way, sharing as much as possible of the bandwidth
> load of spidering.
This is precisely what we've been discussing with a couple of
sites that currently host a large number of feeds that we consume. The
idea is that while they would, of course, continue to provide RSS or
Atom feeds for normal use, they would provide an alternative feed for
us so that we can minimize the impact of our spidering of their sites.
This is good for everyone I think...
> My own opinion is that ... in most [cases] a custom,
> non-Infoset-based serialization would almost certainly
> be better.
It is certainly the case that non-Infoset-based serializations
will, in general, be more compact and faster to parse than
Infoset-based serializations. As mentioned above, it should be noted
that such non-Infoset-based serializations are implicitly defined as
soon as you define an XML schema for Atom or any other XML format
since XSD to ASN.1 conversion is deterministic and standardized.
However, there is value in non-Infoset-based serializations given the
current usage of XML since there is a great deal of "no-schema" XML on
the net and many people write extensions to schema-defined formats
without providing proper schema updates. Infoset-based serializations
support the same flexibility that XML users have become accustomed to
-- but at a cost... Infoset-based serializations are less compact and
slower to parse than schema-based encodings.
> Another side question, assuming that binary Atom was
> available, how would this work with the Atom API?
Clearly, the Atom API as currently defined isn't going to work
with non-XML encodings. To work, there would have to be some mechanism
for the endpoints to negotiate or discover each other's ability and/or
desire to consume alternate encodings. That isn't supported in the
current API.
bob wyman
|
