Re: [secdir] SECDIR review of draft-kyzivat-case-sensitive-abnf

[Chris Lonvick <[e-mail]>]

Hi Paul,

On 9/6/14, 9:12 AM, Paul Kyzivat wrote:
> Chris,
>
> Thanks for the comments.
>
> On 9/5/14 6:02 PM, Chris Lonvick wrote:
>> Hi,
>>
>> I have reviewed this document as part of the security directorate's
>> ongoing effort to review all IETF documents being processed by the IESG.
>> These comments were written primarily for the benefit of the security
>> area directors. Document editors and WG chairs should treat these
>> comments just like any other last call comments.
>>
>> The abstract is:
>>
>>     This document extends the base definition of ABNF (Augmented Mackus-
>>     Naur Form) to include a way to specify ASCII string literals that 
>> are
>>     matched in a case-sensitive manner.
>>
>>
>> Overall, I don't like the statement in the Security Considerations
>> section, but it is consistent with all other documents related to
>> defining ABNF, and I can't find any noteworthy security issues anyway.
>>  From that, I have no objection to moving this document forward.
>
> As you can see, I just followed precedent since I wasn't doing 
> anything that would alter the security implications in any way.
>
> But I am open to suggestions for something better to say.

Nah, I like the consistency.  The only concern (very minor) that I have 
is about how older ABNF interpreters would react to seeing these new 
literals.  I don't think that they're mission critical in any way and 
the one that I tried (BAP) just gave me an error.
>
>> I did find some nits and have some suggestions for improving 
>> readability.
>>
>> 1 - "Mackus-Naur" is used in two places rather than "Backus-Naur".
>
> Yes. I don't know how that happened.
Kind'a funny - you're not the first to make that error.
http://publib.boulder.ibm.com/infocenter/wtelecom/v6r2m0/index.jsp?topic=/com.ibm.diameter.rf.doc/rf_rawinterface_r.html
>
>> 2 - The last sentence of section 2.1 is:
>>
>>     This mechanism has a clear readability
>>     disadvantage, with respect to using a literal text string with a
>>     prefix, and new the prefix mechanism is preferred.
>>
>>
>> Perhaps you meant:
>>     This mechanism of using a literal text string with a prefix has a 
>> clear
>>     readability disadvantage.  The prefix mechanism described in this
>>     specification can be much more easily read.
>
> No. "This mechanism" refers to "the way that has been used in the 
> past" (specify the individual characters numerically). How about:
>
> "The new way (using a literal text string with a prefix) has a clear 
> readability advantage over the old way."

Works for me.  See the response I sent to Barry.
>
>> 3 - This part of Section 2.1 may be cleared up some:
>>   ---vvv---
>>
>> If no prefix is present then the string is case-insensitive.
>>
>>     Hence:
>>
>>           rulename = %i"aBc"
>>
>>     and:
>>
>>           rulename = "abc"
>>
>>     will both match "abc", "Abc", "aBc", "abC", "ABc", "aBC", "AbC", and
>>     "ABC".
>>
>>
>>   ---^^^---
>>
>>   Suggested:
>>    ---vvv---
>>       To be consistent with current implementations of ABNF, having no
>>       prefix means that the string is case-insensitive, and is 
>> equivalent
>>       to having the "%i" prefix.
>
> This seems good, except for the use of "current". That doesn't age 
> well. I suggest replacing "current" with "prior".
>
Works for me as well.

Best regards,
Chris
>     Thanks,
>     Paul
>
>>     Hence:
>>
>>           rulename = %i"aBc"
>>
>>     and:
>>
>>           rulename = "abc"
>>
>>     are equivalent and both will match "abc", "Abc", "aBc", "abC", 
>> "ABc",
>>     "aBC", "AbC", and "ABC".
>> ---^^^---
>>
>> Best regards,
>> Chris
>>
>