It is a bit surprising that venerable old pundit workhorse David Broder doesn't know how gerrymandering works, and what it is supposed to accomplish. Gerrymandering is a technique whereby politicians draw political boundaries in a way best suited to help their party - and it doesn't quite operate the way Broder thinks it does.
Broder sees gerrymandering as a huge problem, maybe even bigger than campaign financing itself:
As a number of scholars have pointed out, the scarcity of real competition in nearly all districts has many consequences -- all bad. It makes legislators less responsive to public opinion, since they are in effect safe from challenge in November. It shifts the competition from the general election to the primary, where candidates of more extreme views can hope to attract support from passionately ideological voters and exploit the low turnouts typical of those primaries.
Gerrymandered, one-party districts tend to send highly partisan representatives to the House or the legislature, contributing to the gridlock in government that is so distasteful to voters.
The problem with Broder's article is that the idea behind gerrymandering is not to build safe districts for the party. In fact, the point is to do nearly the exact opposite: build slim majorities in as many districts as possible, so as to maximize the party's number of representatives in Washington D.C.
Read the article and see if that doesn't completely change and obliterate Broder's analysis.