Re: [babel] Mirja Kühlewind's No Objection on draft-ietf-babel-applicability-08: (with COMMENT)

[Juliusz Chroboczek <[e-mail]>]

The changes below are in git, I'm not submitting a new revision yet.

> 15 minutes to explain really depends on the two people involved.

I'm speaking about giving a talk.  I've done it repeatedly, last time was
just two weeks ago.  I'll keep this bit, if that's okay with you.

> Also implementing “over two nights” really doesn’t tell me much. I think
> it would be more useful to e.g. talk about the lines of code or number
> of components or whatever.

I most humbly disagree.  Time required to do an independent reimplementation
is a much better indicator than mere lines of code.

> I think the “more often than not” is again very blurry. E.g. I would recommend to say instead: 
> "In addition to the above, implementation experience indicates
>   that Babel tends to be robust with respect to bugs: in various 
>   cases implementation bugs did not violate the properties on which
>   Babel relies, and therefore slowed down convergence or caused sub-
>   optimal routing but did not cause the network to collapse."

Agreed.  I've put "in many cases".

>>> "No other routing protocol known to us is similarly robust and
>>> efficient in this particular kind of topology."

>> Again, your objection is not clear to me.  Are you disagreeing with the
>> claim (e.g., because you know otherwise), or do you believe that this
>> statement doesn't belong in this document, and if so, why?

> I would rather suggest:

> “Other widely deployed routing protocol are less robust and efficient in
> this particular kind of topology.”

I've removed this sentence completely.

> I believe there are still more cases in the draft which could be toned
> down a bit to avoid overstating but make a more objective statement
> instead.

Since I'm the original designer, I'm naturally biased.  I'd appreciate it
if you could specify exactly which bits you find objectionable.

Thanks,

-- Juliusz