[
Lists Home |
Date Index |
Thread Index
]
>>>
>>> worse - the encoder/decoder must therefore be deployed in pairs -
>>> and upgraded in pairs. this may later become the basis of something,
>>> but it is along way from being a binary xml coding that is useful in
>>> general.
>>
>>
>> That's because binary XML makes most sense in tightly coupled systems
>> (and the intended usage is clearly stated as such - there's no
>> ambiguity here).
>
>
> this doesn't answer the question of how to upgrade a million phones
> because a newer, better serialisation has been worked out. in fact
> that in general is the problem with all widely deployed technologies -
> how many websites still consider supporting netscape4?
>
> bnux seems to me to be a useful inhouse protocol, but not a general,
> standardised, industry solution.
It's not a standard, and so obviously that's the case.
The only compelling reason for a binary standard would be if the
cost/benefit ratio is appropriate to begin with. Which boils down to
"how much faster? And perhaps, how much smaller?" vs. "how much
additional complexity and pain does it cause". Until those questions
are answered with prototypes and hard data, argueing remains handwaving
speculative FUD, and no reasonable opinion can be formed either way.
|
