This is all speculation, that is all true. Something that bugs me is
that
the ILS for 31C requires RVR of 4000 or 3/4 of mile visibility. The
METAR
from just before the accident pegged visibility at 1/2 mile. Seems to me
that the approach shouldn't have even started.

Well, for starters, prevailing visibility and a specific runway's RVR
can often differ by quite a bit. That's why RVR is controlling for an
airliner. It also changes minute by minute, and the hourly ATIS isn't
at all a good indicator of what the RVR was when that airplane landed.

Yep, thanks for that. The hourly METAR was:

KMDW 090053Z 10011KT 1/2SM SN FZFG BKN004 OVC014 M03/M05 A3006 RMK AO2
SLP196 R31C/4500FT SNINCR 1/10 P0000 T10331050 $

Showing the field visibilty as 1/2-mile and 31C's RVR as 4,500 feet -
greater than minimum.

The FAA accident report has the following for the weather:

0115 11007KT 1/2SM SN FZFG VV003 M04/M05 A3006 R31C/4500V500

This doesn't appear properly formatted, as with a V separator, it's supposed
to be minimum and maximum. It's possible that it's supposed to be a minimum
of 4,500 and a maximum of 5,000 feet, and the extra zero was dropped off.
Or it could be that visibility ranged from 500 feet to 4,500 feet. That I
find less likely.

It'll be interesting to get the CVR and ATC transcripts and find out what
the pilots were told.

Second, Southwest's 737-700 airplanes have a heads-up display that lets
them use 3000 RVR on that runway at MDW.

Interesting! Do you any additional info on that?