Lists Home |
Date Index |
Bob Wyman wrote:
> How is writing a SAX interface to ASN.1 defined binary
> encodings any different from writing a SAX interface to CSV
> files? Or, to GEDCOM (the genealogy data format)? I've seen
> both of these as examples in books written by respected XML
> authorities. Neither of these formats is "XML", yet, those
> authors are not condemned for showing or teaching people how
> to use SAX on non-XML data... The ability to strap non-XML
> data into SAX is, in books that teach people how to use SAX,
> often presented as a great *strength* of the interface.
Neither of them need SAX to be optimized on their behalf. Your
What makes you think you're the only one who needs a flag?
> As others have pointed out. SAX2, as it stands today,
> is a perfectly good interface for working with ASN.1 defined
> encodings. It *needs* no changes. This isn't just my
> personal opinion, it is backed up by statements in this list
> from someone working for a vendor that is building a SAX2
> interface for ASN.1 defined encodings. My proposal only
> suggested an optional *extension* to SAX (using the normal
> SAX extension mechanisms) that would make working with SAX
> just a little more comfortable for someone who is used to
> working with typed data. I fail to see the harm in that.
I fail to see the harm in someone who needs that convenience
layering in on top of SAX, perhaps as a filter, and publishing
*that* for everyone else's use. Unicode wonks are happy, people who
need data typing and binary SAX streams are happy.
> On "binary XML". If you read my earlier postings in the
> list, you will see that I have consistently argued that
> there should be NO "binary XML." We already have widely used
> and effective binary encoding standards that are used to
> enable everything from the cell phone system to our network
> management services (SNMP) to our directory services (LDAP),
> etc. We don't need a "new" binary encoding standard even
> though there are a number of folk in the XML community who
> seem to want to create one. Leave XML to textual-encodings.
> ASN.1 has already dealt with the binary problem.
But you need to change a key XML technology to make it work for
non-XML technology? This is where I get lost and I end up saying
your position is dissonanant.
SAX is XML parsing API. If you feel you need to use that XML parsing
API for non-XML, then tune your encoding to do so without impinging
on the API.
Bill de hÓra