> On Jun 29, 2017, at 9:23 AM, Stephen Farrell <[email protected]>
> wrote:
>
> The library point is relevant because some people
> will already have json support, at which point there
> is no additional complexity, for them. (That said, I've
> always been neutral in religious debates about data
> definition language fashions, but do realise that a
> lot of people have much stronger feelings. In any
> case, the use of terms like "bloat" etc doesn't
> convince me of anything:-)
Postfix, et. al. are part of the "base" footprint of a number
of operating system platforms. The release engineering teams
of those operating system platforms very much prefer to keep
the number of dependencies in check, and to not introduce
unnecessary new components into the base build.
If folks want to see STS support in the base Postfix version
in, e.g., NetBSD then it is best to avoid introducing avoidable
dependencies on additional libraries that would otherwise not
be required for the base system.
JSON *is* complex. My processing of the Rapid7 FDNS feed (22GB
compressed) burns hours of CPU since they switched from simple
white-space separated files to JSON, which I parse with "jq".
JSON is not needed for the problem at hand.
Sure, if inertia trumps simplicity of design, we'll stumble
along with JSON, but I am rather surprised that it is getting
such avid support.
--
Viktor.
_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta
