Why self describing data formats:

[James A. Donald]

James A. Donald wrote:
>> Many protocols use some form of self describing data format, for example
>> ASN.1, XML, S expressions, and bencoding.
>>
>> Why?
>>
>> Presumably both ends of the conversation have negotiated what protocol
>> version they are using (and if they have not, you have big problems) and
>> when they receive data, they need to get the data they expect.  If they
>> are looking for list of integer pairs, and they get an integer string
>> pair, then having the string correctly identified as a string is not going to
>> help much.

Charlie Kaufman wrote:
> You are correct that such encodings don't help with any interoperability
> issues. Sometimes, they make reading and writing the spec easier, since
> silly issues like big endian vs. little endian encoding of integers gets
> specified elsewhere. 

They also make it easy to write specs that do not in fact work.  The 
spec writers do not in fact agree, and then leave the problem of 
implementing an under defined spec to the engineer.

> More rarely, it makes coding easier (if there is some
> parsing and encoding engine readily available to the implementers). If
> the protocol is being designed by a committee, it can reduce the number of
> debates over minutia.

In the case of XML, yes there is a parsing engine, and if the structure 
of the DTD reflects the structure of the algorithm, then indeed it makes 
things much easier.  But usually the committee have not thought about 
the algorithm, or have unresolved disagreements about what the algorithm 
should be, leaving the engineer with problems that are at best extremely 
difficult to solve, and are at worst impossible to solve.  Ideally the 
DTD should be developed in parallel with the program that processes the 
XML.  In that case, you get the parsing engine doing a lot of work for 
free, so the engineers do not have to reinvent the wheel.  But if the 
DTD is written first by one group, and the program second, by another 
group, the second group is usually hosed good.

> But the main motivation (imho) is that it's trendy. And once anyone
> proposes a heavyweight "standard" encoding, anyone who opposes it is
> labeled a Luddite.

Sounds true.

---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at metzdowd.com