|
305
36.3.21.2
5634.07
|
"The PER shall be less than 10% for a PSDU with the rate-dependent input levels listed in Table 36-67 (Receiver minimum input level sensitivity)." (With PSDU octet lengths as sgiven by the following sentence.) With
every new amndment comes a new list, of ever-goriwng size, of these minimum receiver sensitivity numbers. What purpose do these requirements fill? It's natural to assume that they are an effort to make sure that devices perform at least at some basic level
in all modes. But they don't really do that, at least not directly, because there is no requirement to perform at any level at all for any PSDU length other than one specific one per MCS. Instead, the requirement seems to set an indirect minimum performance
level for all PSDU lengths, since if the device's MCS reception works at the stated single PSDU length, that provides evidence that the consituent blocks--BCC decoder, LDPC decoder, etc.--are correctly implemented. This is indeed useful. However--we have already
ensured all that, by specifying corresponding minmum receiver sensitivity numbers in previous clauses. What * new * information is conveyed by a device's result corresponding to any entry on this table? Compared, in particular, to HT? It's all very marginal,
isn't it? As part of 11mf, we should start the process of pruning out surplus requiremements and simplifying the standard. As with other obsolete text, it is hard to do this all in one go. A useful first step would be to make the numbers in this table recommended
rather than mandatory. Note that the numbers in the corresponding tables in earlier clauses would remain mandatory.
|
Change "shall" to "should".
|