Re: [dispatch] SIP-CLF: text versus binary

The reason was not to carry it in ISUP or the BICC protocol (I'm not 
aware of any spec doing this). Companies wanted to use it in gateways 
where ISUP, BICC and H.323 were the call control protocols. All these 
were binary. That's why ITU-T and 3GPP adopted binary for those 
applications. However as H.248/Megaco evolved many applications were 
more closely tied to the work of the IETF and the usage has tended to be 
toward text implementations. Both the text and binary had tons of 
intellectual effort poured into them.
The text version is certainly easier for humans to read but slightly 
more problematic because a human must understand the grammar.
In terms of the technical performance like the tests here have shown the 
results are very dependent who has done the encoding.

My advice would be to pick one (if you use the coin method make sure it 
doesn't roll under a chair :-) ) and then document an optimised encode / 
decode code as part of the spec to try to minimise performance variability.

Good luck,
 Christian

PS: And no I wasn't rapporteur at the time.

Tom Taylor wrote:
> Just for fun, I'll remind you all that we tried to decide the H.248 
> text vs. binary question by the flip of a coin. The coin came up 
> "text" while the H.248 rapporteur sat there with his jaw touching the 
> floor, but then Pete Cordell came up with the idea of text-encoded 
> ASN.1 and that kept the issue alive. It turned out the real reason the 
> binary wouldn't die was that some people wanted to carry it in ISUP 
> (BICC, actually).
>
> Tom Taylor
>
> Eric Burger wrote:
> ...
>>
>> Why is H.248 a text protocol in the wild, even though the binary 
>> format had tons of intellectual effort poured into it?
>>
>
> _______________________________________________
> dispatch mailing list
> dispatch@ietf.org
> https://www.ietf.org/mailman/listinfo/dispatch
>