W3C home > Mailing lists > Public > xml-dist-app@w3.org > February 2000

irc chat: low-end and mid-level

From: Ken MacLeod <ken@bitsko.slc.ut.us>
Date: 25 Feb 2000 12:47:42 -0600
To: xml-dist-app@w3.org
Message-ID: <x5og95ruep.fsf@bitsko.slc.ut.us>
In the chat there was some confusion about what I meant by "low-end"
and "mid-level" (could be "high-end", but I think there's more up
there).  Hopefully this will clarify.

"low-end" to me is "simple serialization".  In simple serialization
the goal is to take application data and convert it to an on-the-wire
format with as little infrastructure as necessary (IDLs, data models,
schemas, etc.).

Simple serialization presupposes that either the application at both
ends understand the data completely, check it, and coerce it as
necessary; or that the application at one end (likely a server role)
has more infrastructure available to it.  Simple serialization is
generally "automatic" and requires little or no knowledge of the data
being serialized.  Many common Web and Internet protocols work like
this.

"mid-level" to me is more complex serialization, there is end-to-end
infrastructure that defines the serialization and how it's applied.
Mid-level serialization generally requires some mapping of application
data to the serialized format or strict rules governing the
serialization.  Many common distributed computing protocols work like
this.

In low-end protocols, there is typical a single fixed document type
that all application data is serialized to (XML-RPC, LDO XML, WDDX) or
a simple object-to-XML mapping is done (Coins).  In mid-level
protocols, there is typically a document type or schema associated
with each application data entity and/or message.

Please note, I'm not arguing in favor of one or the other, just noting
the need for both.

  -- Ken
Received on Friday, 25 February 2000 13:47:52 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:56 GMT