RE: [Tools-discuss] Idnits-1.84 feedback

> -----Original Message-----
> From: Henrik Levkowetz [mailto:henrik@levkowetz.com] 
> Sent: Wednesday, January 04, 2006 8:40 AM
> To: Scott Hollenbeck
> Cc: tools-discuss@ietf.org
> Subject: Re: [Tools-discuss] Idnits-1.84 feedback
> 
> Hi Scott,
> 
> on 2006-01-03 20:01 Scott Hollenbeck said the following:
> > I just ran a document through the idnits checker.  It produced some
> > unexpected results:
> 
> [snip]
> 
> > It looks like the tool is checking for 2006 copyright dates 
> now that we're
> > in 2006, but that will be a problem for documents that were 
> written in 2005.
> 
> You're right, it's checking for copyright dates matching the 
> current year.
> I think this is strictly speaking correct, but wouldn't mind 
> adding a grace
> period just after each new year.  (Of course, that conflicts with the
> current policy with respect to boilerplate, which is to 
> require letter-
> perfect adherence, with a male author not even being 
> permitted to change
> "he or she" to "he"; something I personally think is silly, 
> but nevermind.)
> 
> Any suggestion as to the length of such a grace period?
> 
> Or should I instead add a switch to make it possible to say 
> something like
> 
>   $ idnits --year 2005 draft-ietf-sieve-3431bis-04.txt

A switch sounds like a good idea.  Maybe a text box on the web form, too?
Of course, it's not really a big deal if the user recognizes what's going
on.  It's probably OK to just leave things as-is.

I know that the current year needs to be checked when submitting a new
document, but my use case is a little bit different: I also check documents
that are being submitted to the IESG for evaluation.  They might have been
written in the prior year.

> and have idnits accept the given year instead of the current 
> year in copyright
> text, maybe?
> 
> > The BCP 78 error is interesting.  The text is in the 
> document, but it exists
> > after the copyright notice.  Did the tool skip it because 
> of the copyright
> > date error?
> 
> Yes.  If the text had been in a separate paragraph, it would have been
> recognized, but I do some extra work to also recognize 
> multiple boiler-plate
> paragraphs which has been contracted into one.  This is done 
> by concatenating
> potential boilerplate paragraphs and matching the aggregate 
> against the
> current paragraph, and thus fails when one of the parts in 
> the input text
> fails to match.
> 
> A strategy of instead trying to split paragraphs in the draft 
> input into
> potential individual boilerplate paragraphs would not fail in 
> this way, but
> would also (I think) be quite a bit more complex.  I may 
> consider re-writing
> to do in the future, but won't do it now.

Again, it's no big deal.  I just wanted to be sure that I was interpreting
the output correctly.

Thanks,
-Scott-

_______________________________________________
Tools-discuss mailing list
Tools-discuss@ietf.org
https://www1.ietf.org/mailman/listinfo/tools-discuss