Andrew Sullivan wrote:
I will note that I sort of hate the approach outlined in section 3 in
the draft as written. In the current form, we have a table that
either has a value in the subordinate field, _or_ a value in the RR
field; in either case, the other field has to be NULL. If I put on my
data geek hat, this sort of makes me queasy, on normal form grounds.
Given the way different underscore uses have emerged, however, I have
yet to come up with anything that would make this much simpler, and
still actually cover the cases we have.
Andrew,
Thanks for posting this.
As we've discussed privately, this seems to be a relatively straightforward
trade-off between cleanliness of the design versus number of tables that IANA
will have to maintain. But by straightforward, I mean that understanding the
choice is easy, not that making the decision is.
The concern that drove the current document's design was to reduce the number
of tables. However your concern that it causes two columns of the table to
interact is correct and I certainly agree that it has its downsides.
Do others have some thoughts on this?
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
DNSOP mailing list
DNSOP@ietf.org
https://www1.ietf.org/mailman/listinfo/dnsop
