Re: [netmod] Adam Roach's No Objection on draft-ietf-netmod-rfc6087bis-18: (with COMMENT)

[Juergen Schoenwaelder <[e-mail]>]

On Fri, Mar 09, 2018 at 03:22:26PM +0100, Benoit Claise wrote:
> Hi,
> > On Thu, Mar 08, 2018 at 04:08:32PM -0800, Andy Bierman wrote:
> > > I don't really know what a guideline should say about patterns.
> > > I will try to add something that says to document the pattern limitations
> > > and keep the pattern as simple as possible,
> > > 
> > I object to a statement that "pattern should be as simple as
> > possible".
> The only guideline that makes sense is: "the pattern must be correct" or
> "the pattern must be meaningful". Not even worth mentioning, as this is so
> obvious!
> Whether it's a simple or a complex regex is orthogonal here: the pattern
> must fulfill its job.

Properties such as 'correct' or 'meaningful' are not easy to define
for pattern. Pattern can have different levels of strictness and often
greater strictness comes at the price of increased complexity of a
pattern.  There is no general advice that can be given how to trade
strictness of a pattern against pattern complexity.

That said, pattern can be incorrect (rejecting input that they are
supposed to accept). If pattern become increasingly complex to
increase their strictness, then it is important that they are
sufficiently tested not to be incorrect.

/js

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <https://www.jacobs-university.de/>