As Lloyd pointed out in his article, writing for other systems is great, but if SJ Games doesn't want to publish your GURPS book, tough. Not so with d20. That's one of the main thrusts of the article. It's also one of the reasons there is so much d20 stuff already.True, a pragmatic DM will find a rule screw, crumple up the character sheet, and tell the player "start over" (or "get out"). But it does point out a problem with d20 ... there are a plethora of rules, and "feats" that bend the rules, and feats that interact with other feats ... (Compare to point-buy systems like GURPS 4e or Champions, where most powers are combinations of existing advantages and already fairly balanced, or comparatively lighter systems like Call of Cthulhu and World of Darkness where rules take a back seat to color, atmosphere, and, alas, real-world research.)
A responsible writer will attempt to playtest his work before unleashing it on the public. Of course, playtesting undercuts the argument that "with a little research, you can pump out a lot of articles". An irresponsible writer will simply toss something out there; by the time some poor schmuck actually tries to use it, he'll have already collected his fee. As a previous post notes, eventually such a writer will gain a reputation for doing shoddy work ... but maybe he'll be selling time shares by then.
And as for the "problem" you have identified with d20, please remember that one man's problem is another man's feature. Many d20 fans *love* the plethora of rules, which allow them to pick and choose what they like, rather than what a single publisher deems acceptable.
Speaking as a freelance writer, playtesting is the publisher's job. If a publisher publishes something that was not playtested, blame the publisher. For the meager rates a d20 freelancer earns, he doesn't have the time to playtest things himself. An "irresponsible" writer might throw tons of news ideas out there; it's the publisher, editor and playtesters that seperate the wheat from the chaff. Not every idea a writer has will be good; but if a publisher can't tell a poor idea from a good one, you can hardly blame the writer.