Re: [dhcwg] extending the DHCPv4 option space, VSO usage -- was: Re: issues with draft-grochla-80211-dhcp-option-00

[Ted Lemon <[e-mail]>]

On Sep 24, 2010, at 6:33 AM, Alfred HÎnes wrote:
> Regarding Ted's previous comments on the difficulty in establishing
> a uniform [sub-]option request format for vendor options, it should
> be recalled that both the original (#63) and the enhanced (#125)
> DHCPv4 vendor information option specifications strongly recommend
> a uniform format for sub-options.  This could be used as a base
> for follow-up work, restricting the scope to that format, and thus
> avoiding complications and discussions.

I'm sorry, this just seems like noise to me.   Who cares what some other IETF protocol does?   This isn't some other IETF protocol--this is a specific protocol, with specific semantics.  The proposal to use vendor-specific options for this option ignores the fact that you can't *request* one of these options.

There seems to be a strong sudden groundswell of support for using vendor options to avoid consuming option codes.   Vendor options were never intended for this--vendor options were intended for cases where the vendor doesn't want to take the trouble to document what they are doing, or for cases where what they are doing would be of no interest to others, and hence not *worth* documenting.   There's an implication that vendor options might be used in some cases to do things the vendor doesn't *want* to document.   This doesn't fall into any of those three cases, does it?

So why the sudden push to use vendor options in this way?