Date: Mon, 30 Oct 95 15:19:07 CST From: keen@ncrssd.StPaul.NCR.COM (Hal Keen ) To: p8021@nic.hep.net Comment on P802.1p/D0 Source: Hal Keen, AT&T GIS The potential audience for this standard may include some implementors who are not certain how much authority a "recommendation" carries. For them, the recommendation that two traffic classes be implemented (on page 17, in the fifth paragraph of 3.74) is stated a bit too simply. It could lead to the erroneous impression that departing from the recommendation constitutes a failure to follow the standard. What is really meant is that two traffic classes will "cover the majority of installation requirements" with "minimum implementation overhead, while not disallowing the provision of additional traffic classes where operating requirements warrant it." (Quoted material is taken from Annex F.) In other words, the recommendation addresses only a subset--granted, a very large one--of the range of operating requirements encompassed by the standard. This should be made clear in 3.7.4, not just in an informative annex. It would be sufficient to expand the note, following the statement of the recommendation, to inform the reader that Annex F also offers guidance for cases in which departure from the recommendation may be warranted. With this change, the reader is informed on encountering the recommendation that there are circumstances in which it might be best to depart from it. This reduces the chance of misunderstandings while avoiding any modification to the recommendation itself.