If all goes well, sometime next year construction will begin upon an ingenious, urbane set of buildings in downtown Washington. This of course is welcome news. But the real surprise is the client: George Washington University.

Time was when the university, buying up properties next to its city campus as fast as they came on the market as part of its long-range plan for financial stability, could do as it pleased within the huge chunk of territory, roughly 14 city blocks, it controlled or coveted.

The trouble was that what pleased the university -- big, hard, boringly tasteful office buildings along with equally boring large-scale academic structures -- had the opposite effect upon many people who had to live near them or pass by or through them daily.

If the string of buildings designed for GWU by the Washington firm of Keyes, Condon and Florance is any indication, however, the university has changed its ways for the better. To be sure, the signs are mixed. In some respects the university still behaves as a shrewd and fast-moving real-estate development corporation that thinks its word is law in a large part of Foggy Bottom.

Nonetheless, to look at the new design for the GW law school is definitely to look on the bright side.

As a step up in quality, the law school project is more important than the huge, highly publicized new building the university will soon start to build behind the "Red Lion Row" of 19th-century houses and shops facing the 2000 block of Pennsylvania Avenue. In deciding to save that pleasant stretch of history by massing a mixed-use office tower behind the Red Lion row, the university was doing its civic duty, responding to changes in the District law governing historic structures as well as a prolonged public hue and cry.

In deciding on the expansion of its law school, the university was acting on its own. The new scheme is not pace-setting architecture -- not design "at the cutting edge," as they say -- but it is intelligent, sensitive, clever, complex and intriguing, qualities that in the long run may be more important to the city than sheer architectural innovation.

In fact a good case can be made that, as important as this design may be as a bellwether of George Washington University's intentions, it is more significant as a sign that a set of fresh ideas concerning urban design has begun to take some hold among Washington developers and architects.

The Keyes, Condon and Florance scheme for the GW law school is not alone. There are plans on the drawing boards of several firms around town that approach similar problems in equally healthy and perhaps even more imaginative ways. The principal issues are two: the urban setting itself and a style -- or, more precisely, a set of invigorating stylistic options -- that architects can use to solve the vexing, ubiquitous problem of fitting the new with the old, doing justice to both without spending a mountain of money in the process.

By any reckoning, the problem of context presented by the law school expansion is a tough nut. At present the school is squeezed into a curious assortment of buildings along the western side of 20th Street NW (between H and I Streets). This stretch includes five distinct, although connected, buildings: a turreted, turn-of-the-century apartment house that quaintly but effectively commands the north corner of the row; the Stockton building, a semi-colonial-style hall designed during the eclectic 1920s; the law library, a strong, red brick structure designed with a bit of mid-'60s sculptural pizazz; and two nice little 19th-century row houses that bring this oddball collection to a frumpy, anticlimatic end.

None of this is tied together, either functionally or esthetically, inside or outside, top or bottom. It is a clear case of a setting where pure restoration alone won't do the job that the university wanted done -- namely, expanding the law school space by about 60 percent -- nor the one that we might desire in the public interest -- a cohesive, visually interesting, texturally pleasing parade of buildings.

Consequently, the Keyes, Condon and Florance solution involves a mixture of preservation, alteration, demolition and new construction. To say that the design works, by-and-large, is no small compliment given the complexity of problems presented by the existing potpourri.

The present law library facade, with its irregular, deeply recessed entrance bay, is a building that wants to be heroic but does so in the wrong place. For the architects it was a design headache that just wouldn't go away -- so they did the next-best thing. They turned its too-strong contrasts of mass and negative space to good account, first cleaning them up somewhat by removing high vertical columns and a ceremonial stairwell, and then by extending this play of spaces and textures onto an adjoining, new facade. The transition is odd but clever and it has the crucial merit of focusing attention upon the strongest part of the ensemble: the corner building that will replace those two outgunned Victorian row houses.

This new building is a delight from many points of view: the incisive way the roof line cuts down in several stages, the tastefully rugged surge of bays, the Victorian echoes in the rounded upper windows and false gables, the archway that, in addition to providing an inviting covered path to the main university quadrangle (in place of a ridiculous little alley), brings the row to a fine, surprising conclusion. Moving stylistically from mid-'60s muscle to early-'80s Victoriana is indeed a surprise, but Keyes, Condon and Florance here provide a pretty good lesson in how to go about it.

The architects didn't do badly at the other end, either, in the clean-lined, neo-Victorian building they designed to replace the turreted apartment building at 20th and H streets, where the play of volumes and recesses, if perhaps a bit too crisp, does indeed provide strength and visual interest where it is needed most.

Stockton Hall, incidentally, will be ingeniously redone on the inside so that it connects with the other buildings, but on the outside it will remain basically the same: a neat, unprepossessing and thoroughly improbable centerpiece that confirms the wisdom -- actually, the necessity -- of concentrating visual energy at either end of the row.

Committed don't-tear-it-downers may object, but the net result is a tremendous improvement in the streetscape and in the architectural interest and readability of the block as a whole. And then the law school design is obviously, even blatantly, the most progressive piece of architecture and urban design commissioned by George Washington University since the building boom began there a decade or so ago. A few more as good as this, and even skeptics might be convinced.