I have been exchanging e-mails with a D.C. principal about the new teacher evaluation system, IMPACT. Our conversation is not on the usual topic, what the new system is doing for her teachers. She tells me about what it is doing to her.

I see the flaws in IMPACT. I have devoted columns to teachers who think it is a mistake. But I consider it worth a try because it is better than what the District was doing before, and than what most suburban schools are still doing.

This principal, whom I have agreed not to name because she feared central office reprisals, saw little to admire in the evaluation system of the suburban district where she once worked. She does not defend the limp D.C. assessment system before former D.C. schools chancellor Michelle A. Rhee introduced IMPACT.

To her, that’s not the point. She is not comparing the way teachers in general were judged before and after IMPACT. She is comparing the way her teachers were and are judged, and how the new system hinders her ability to be a good principal.

She is a devotee of education expert Carolyn J. Downey’s three-minute walk-through method of classroom observation.

“Her basic premise,” the principal said, “is that principals and teachers can work together to change teaching practices for the better when the principals actually get into the classrooms on a regular basis and use these visits for conversations about instruction.”

In the six years she did this before IMPACT took over in 2009, “I rarely spent a long period of time in each room,” she said. “Many days I visited every class on the same grade level. . . . I saw teachers at their best, and needing some help. I was able to see if various classes on the same grade level were being consistent in their pacing of instruction. I knew my students well, especially the ones that were struggling. I was able to give constant feedback to the staff about what was happening in their classrooms and found that these continuous visits removed all levels of defensiveness in our conversations.”

“The students never knew when I was coming around, and my popping in and out of rooms gave them a clear message that I knew what was going on in the building,” she said.

Then IMPACT began. She had to make three 30-minute observations of each teacher each year, write reports and discuss them with the individuals. Outside experts called master educators did two similar observations each year for each teacher.

This has had, for her, a troubling effect. “I rarely am seen walking around the school since I have specific appointments to keep in the classrooms,” she said. “I can rarely compare what is happening from room to room since the visits are on different days at different times. The teachers have started commenting on the fact that the students rarely see me anymore . . . I feel my knowledge of specific students and their needs is slipping away.”

The master educators, whose work was vividly illustrated by my colleague Stephanie McCrummen’s March 18 article, are to the principal and her teachers “just a blip on the radar screen,” she said. Their e-mail communication improved, but mostly they play their parts, “and are gone, never to be seen again,” she said. “The most important element of improving teaching practice, developing a relationship with other colleagues, sharing ideas and techniques, visiting other classrooms and talking about specific students is completely absent from the master educator experience.”

I still think IMPACT has a place in low-performing schools with inexperienced teachers who need that structure while they develop their skills. But when a principal knows her stuff, as this one does, and has teachers well-rated by master educators in a school with good results, couldn’t we let her run it her way? If achievement plummets, she’s out. But if she succeeds, what’s the problem?