Twenty-five years ago, this newspaper devoted nearly the entire op-ed page to an extraordinary interview with Harvard professor Robert Coles regarding the desegregation of Boston's public schools.

Coles, a strong integrationist, argued that busing working-class whites and blacks within city limits - and exempting the affluent white suburbs - ignored the larger issue of economic status.

"The ultimate reality is the reality of class," Coles said. Working-class whites and black, he said, "are both competing for a very limited piece of pie, the limits of which are being set by the larger limits of class, which allow them damn little, if anything."

A quarter of a century later, these issues of class and race (and the confusion of the two) are at the center of the fight over the admissions policy at Boston Latin School.

The November decision of the First Circuit Court of Appeals to strike down the use of racial preferences opens the door to a far more robust idea: Why not give a leg up to disadvantaged kids, whatever the race? Such "class based" affirmative action offers four major advantages over racial affirmative action:

First, preferences for the disadvantaged are legal. In the early 1970s, progressive lawyers emphasized the racial segregation of Boston schools because it was race that would trigger judicial involvement. Economic segregation - much to the chagrin of liberal attorneys - presented no constitutional violation.

Today, the legal posture has turned 180 degrees. With the Boston public schools declared "unitary," the use of race, even when intended to benefit people of color, has been found unconstitutional. And though the decision is being appealed, the NAACP is right to fear that the US Supreme Court is likely to use the case to kill not only the Boston Latin plan, but affirmative action in higher education nationwide.

By contrast, because liberals lost their earlier effort to make economic class judicially "suspect," today even the most conservative justices are almost certain to support the constitutionality of class-based preferences.

Second, giving a leg up to poor kids is considered to be fair in a way that counting skin color is not. In determining which of two students has the greater long-run potential, surely a white truck driver's kids are more deserving of consideration than Vernon Jordan's children.

Most Americans sense that counting obstacles overcome, along with academic qualifications, better approximates true meritocracy than adding points for racial background - or looking exclusively at test scores.

A December 1997 New York Times poll found that while Americans reject racial preferences 52 percent to 35 percent, they support replacing such policies with preferences for the poor by 53 percent to 37 percent.

Third, class-based affirmative action will promote the educational goal of having a racially and economically diverse student body. Because people of color are disproportionately poor, they benefit disproportionately from economic preferences.

At UCLA's law school, using an economic-based preference has boosted black, Latino, and Native American enrollment to a level more than five times what it would have been had students been admitted purely on the basis of academic record.

Contrary to conventional wisdom, using a sophisticated definition of economic disadvantage - one that looks not only at income but at such factors as parents' education, net worth, and neighborhood poverty - has produced a positive racial dividend. Of course, class-based affirmative action would also make Boston Latin much more economically diverse than it is today. While Boston Latin has historically been seen as an engine for upward mobility for the poor - to Harvard and beyond - today only 30 percent of Boston Latin students are poor or working class (eligible to receive free or reduced-price lunches) compared with 62 percent of Boston middle and high school students systemwide.

The underrepresentation of disadvantaged students at Boston Latin mirrors in more moderate form the experience of elite universities. Derek Bok and William Bowen's book, "The Shape of the River," found that only 3 percent of students at the elite universities they studied came from the bottom 28 percent socioeconomically. Yet from an educational standpoint, a longshoreman's child (of whatever race) is likely to enrich classroom discussion at least as much as another lawyer's kid (of any race).

Fourth, for liberals there's a powerful political spillover to using class rather than race. Where an exclusive focus on racial preferences and racial desegregation tends to divide working-class whites and blacks, a policy focused on economic status reminds these groups of their common interests.

The late J. Anthony Lukas, author of "Common Ground," told me in 1995 that although the Boston busing crisis pit the affluent white liberal Colin Diver and the underclass African-American Rachel Twymon against the working-class Irish-American Alice McGoff, in fact "the McGoffs and the Twymons have far more in common than either of them does with the Divers."

The potential for an alliance of working-class whites and blacks was tremendous, Lukas said, adding, "I would love to see it happen in this country before I die."

This, then, is the ultimate irony of the First Circuit decision: Conservative judges, by striking down the more modest use of race in admissions, may help bring about a much more enduring and far-reaching policy.

Boston Latin, America's oldest public school, is in a position to lead the nation in something very new: consideration, at long last, of America's "ultimate reality," the division between have and have-not.

Richard Kahlenberg is Education Fellow at The Century Foundation.