Optimizing transmission of XML

[Simon]

Is there utility/function that you can use to provide lookup values
for the tags defined in a DTD?

The problem lots of good tag names that are used to describe small
data items is causing some grief by bloating the data to be
transmitted. As GSM is being used at 9.6K it hurts when the tags bloat
the transmitted data

What I am interested in doing is using lookup table for the tags in a
given DTD. The lookup table would be distributed with the DTD. Sending
the documents described by the DTD with lookup values rather than the
tag names. Use a lookup at the receiver to change the lookup value
back to tag names.

I suspect this function is already available and described by a w3c
standard...just can't see it for looking.

 

[Nick Kew]

Is there utility/function that you can use to provide lookup values
for the tags defined in a DTD?

The problem lots of good tag names that are used to describe small
data items is causing some grief by bloating the data to be
transmitted. As GSM is being used at 9.6K it hurts when the tags bloat
the transmitted data

Just use conventional compression - e.g. gzip - and it will tokenise out
repetitive patterns (such as your tags) to the point where they have
a negligible effect.

 

[Ed Day]

I suspect this function is already available and described by a w3c
standard...just can't see it for looking.

The fact is, the W3C currently does not have any standards for
anything like this. The Binary-XML working group is looking into ways
of optimizing XML for binary transmission, but this is still in the
initial stages.

Ed Day
Objective Systems, Inc.

 

[Simon]

The fact is, the W3C currently does not have any standards for
anything like this. The Binary-XML working group is looking into ways
of optimizing XML for binary transmission, but this is still in the
initial stages.

Ed Day
Objective Systems, Inc.

Thanks Nick and Ed.
As it happens I'm already doing an in-store compress using zlib and
sure it helps but there are lots of very different tag strings rather
than lots of repetition.
I'll keep and eye on the Binary-XML work group. I think I can
implement something in the short term using hashing algorithms, but
would have proffered to use a standard.
Simon