> (2) However, some authors feel that we should not enforce too strong a > requirement to make deployment harder. Hence, instead of changing the > aforementioned MAY to MUST, these authors feel that it is better to make > the pid endpoint property optional. Hence, we make the two changes of E1 > and E6. I agree with this assessment and am therefore in favour of the changes from "MUST" to "MAY".
- Jan > -----Original Message----- > From: alto [mailto:alto-boun...@ietf.org] On Behalf Of Y. Richard Yang > Sent: Friday, August 08, 2014 6:05 AM > To: IETF ALTO > Cc: Megan Ferguson > Subject: [alto] Request for comments on two changes for RFC7285-to-be > > Dear ALTOnians, > > The authors of the ALTO Protocol are finalizing the final version, to be > published as RFC 7285. We seek your comments/feedback on two non- > editorial changes. > > But first, the text, XML, and comprehensive diff files are available at: > http://www.rfc-editor.org/authors/rfc7285.txt > http://www.rfc-editor.org/authors/rfc7285.xml > http://www.rfc-editor.org/authors/rfc7285-diff.html > > A summary of the key changes is listed below: > E1: the change from MUST to MAY in Section 7.1.1 > E2: update from "misses" to "omits" in Section 8.5.2 > E3: the addition of MUST and text addition to Section 10.8.1 > E4: update from "both" to "either...or" in Section 11.2.3.6 > E5: text update to Section 11.4.1.3, paragraph 1 > E6: text deletion from Section 11.4.1.4. The deleted text is "In particular, > the > information resource closure MUST provide the look up of pid for every ALTO > network map defined." > > The AD has approved all changes except E1 and E6, which are non-editorial. > > There are two reasons for E1 and E6: > > (1) Make the document consistent. Specifically, the first paragraph of Section > 11.4.1 states that "An endpoint property resource provides information > about properties for individual endpoints. It MAY be provided by an ALTO > server." Without E1 and E6, the document will imply that at least one > endpoint properties service (i.e., one to provide pid) must be provided. > Hence, one way to achieve consistency is to change the MAY in the first para > of 11.4.1 to be MUST. > > (2) However, some authors feel that we should not enforce too strong a > requirement to make deployment harder. Hence, instead of changing the > aforementioned MAY to MUST, these authors feel that it is better to make > the pid endpoint property optional. Hence, we make the two changes of E1 > and E6. > > Your comments and feedback will be greatly appreciated. We will wait for > one week for any feedback. > > Thanks a lot. > > Richard _______________________________________________ alto mailing list alto@ietf.org https://www.ietf.org/mailman/listinfo/alto