This is the second appeal in a class action brought by appellee-taxpayers in 1973 to enjoin appellant-District of Columbia from using unequal levels of assessment1 of estimated market value in determining the assessment for taxation purposes of single-family residential properties. The trial court originally granted an injunction to appellees, prohibiting the District from using an assessment level other than 55% for single-family residential properties. This court sustained that injunction in District of Columbia v. Green; D.C.App., 310 A.2d 848 (1973). The District then undertook to comply with the injunction and mailed out adjusted real property tax bills to single-family residential taxpayers for fiscal year 1974. Appellees, however, filed a “Motion to Compel Respondents [the District of Columbia] to Correct and Reissue Real Property Tax Bills for Fiscal Year 1974, Or in the Alternative for Refunds”. They asserted that the District had not complied with the injunction since the adjusted tax bills were not based on an exact 55% level of assessment. On August 6, 1974, the trial court granted this motion and ordered the District to recompute assessed values, correct the tax roll, and reissue supplementary tax bills or refunds within 90 days, using a method suggested by the trial court and designed to reach greater mathematical exactitude in the computation of real property taxes. After the District failed to comply within 90 days, the trial court again ordered compliance, this time by December 10, 1974. On December 9, 1974, the District complied with the order of August 6, 1974, and sent out supplementary tax bills or refunds based on the trial court’s method of computation. The District of Columbia appeals from the trial court’s order of August 6, 1974. We hold that the trial court correctly required the District of Columbia to use a method of calculation capable of reducing to a greater degree the mathematical error in the recomputation. We also hold it was not error in declining to view the differences in amounts of taxes between the two methods of computation to be de minimis.
A discussion regarding the mathematical process used 2 and the events leading up to the litigation will be helpful. First, a brief explanation of the components of the taxation process and their storage in the computer: The taxes owed to the District of Columbia each year by a real property owner are computed by multiplying the estimated market value (a unique value for each piece of real property, determined periodically by District tax assessors) by the level of assessment (the percentage of estimated market value upon which the tax will be levied). The product is expressed in dollars per hundred dollars of assessed value. The tax owed is then computed by *307multiplying the tax rate by the number of hundreds of dollars in the real property’s assessed value. The computers used by the District in the tax computation process store the following three significant numbers: (1) the building value at the level of assessment rate; (2) the land area; and (3) a number, in dollars and cents at the level of assessment rate, representing the value of the land area per square foot (the “street rate”).
Second, a brief explanation of the background of this litigation: The District of Columbia began a cyclical reassessment program in 1973 to raise by 5%-increments over several years the level of assessment for all single-family residential properties from 55% to 65%, the latter being the level of assessment for all multi-family residential, commercial, and industrial real properties. When the level of assessment for 44,485 single-family residences was changed for fiscal year 1974 from 55% to 60%, leaving nearly 19,000 other single-family residences with a level of assessment of only 55%, some of the group of 44,485 owners brought the original suit to enjoin the District from raising the level of assessment as to only some of the single-family property owners. The trial court then enjoined the District from
using, for purposes of taxation of single-family residential properties within the District of Columbia, unequal levels of assessment ... in determining the assessment ... of such properties, and from assessing such properties at a level of assessment other than 55% of estimated market value unless and until an equal level of assessment is established by the District after full compliance with the . . . District of Columbia Administrative Procedure Act . . [District of Columbia v. Green, supra at 850; footnote omitted.]
See also Record at 1069 (Trial Court’s Order of June 29, 1973). This court then, as previously mentioned, affirmed. District of Columbia v. Green, supra.
In attempting to comply with the original injunction and to recompute the taxes based on an assessment level of 55% rather than 60%, the District used the following four-step method: (1) by multiplying the "street rate” at 60% by .91667 (the decimal equivalent of fifty-five sixtieths (5%oths) rounded to five figures), rounded to the nearest penny, to get the “street rate” at 55%; (2) by multiplying that “street rate” at 55% to the land area (square feet) to get the land value at 55%; (3) by multiplying the assessed building value at 60% by .91667 (again, the decimal equivalent of 5%oths, rounded to five figures) to get the building value at 55%; and (4) by adding the building value at 55% of the land value at 55% to get total assessed value at 55% (Record at 1548, Opinion and Order of Trial Court, August 6, 1974). This method resulted in an overall average level of assessment for each piece of real property of 55.00772%. (Record at 228-29.)
After appellees filed their motion to compel the District to reissue tax bills or refunds, the method which the trial court ordered the District to use consisted of the following:
[1] multiply the street rate by the Respondents’ [District’s] own factor of .91667 and retain that result as a five place figure after the decimal. [2] That five place figure is then multiplied by the number of square feet and this result, rounded to the nearest dollar, gives the land calculation. [3] The assessment of improvements is then multiplied by the factor of .91667 and this result is rounded to the nearest dollar. [4] The two results are added: land and improvements to give the correct answer as to how to adjust from 60% to 55%. [Record at 1552 (Order of Trial Court, August 6, 1974).]
The District used this method on December 9, 1974, and complied with the trial court’s order of August 6, 1974. This method resulted in an overall average level *308of assessment for each piece of real property of 54.99908%. By comparing these methods it is apparent that a difference of potential significance (depending on the area of real property) originates in step (1) where the product of the street rate multiplied by .91667 is either carried to five decimal points or rounded to the nearest penny.
The District’s contentions on appeal are that (1) there was substantial compliance by the District with the trial court’s original order prohibiting the use of any level of assessment other than 55% where the adjusted tax bills achieved an average level of assessment of 55.00772%; (2) the trial court erred in ordering recomputation of real property tax bills, in view of the fact that the trial court had not specified a method of computation to be used and in view of the fact that the court was told that the method used by the District was the only method compatible with the District’s computerized assessment system; and (3) the trial court erred in not declaring de minimis the amounts by which the tax bills computed by the District’s method differed from those which would have resulted from the methods suggested by ap-pellees.
We conclude that the trial court did not err in holding the District of Columbia to a more exact computation method, both because of the single-year and projected impact of the less exact method, and because insufficient justification for retaining that method was presented. Although the average level of assessment reached by the District using its method (55.00772%) was near the 55% level, and although no method of computation had been specified by the trial court in its original order, the District could nevertheless be required in its recomputation of tax bills to reach an average level of assessment as close as possible to 55%. We are of the view, despite the parties having treated the issue as one of trial court discretion, that what is presented is a question of the legality of a less precise method of tax computation. The judgment of the trial court required an assessment of all single-family property at 55%. When purity in computation was not possible, the trial court rejected the government’s method, and, so long as the method chosen by the trial court does no violence to the law or the language of the judgment, we will not intercede.
As is apparent in the four-step method used by the District, discussed, supra, the District rounded to the nearest penny the result of the first step in its method (“street rate”) rather than carry it out to five figures. This mathematical compromise caused a discrepancy which was perpetuated and compounded in the succeeding steps. The District stated at oral argument in this court that it “could have gotten a better result” had it not rounded off in the first step. In an effort to justify its action, the District’s computer expert testified at the hearings on appellees’ motion that retaining the results of the first step at five figures would not have been “compatible” with the District’s computer system (Record at 192) and would have necessitated the following: “adjustment” of over 100 established programs in the system ; the “designing” of “a new method of valuating land in the District of Columbia”; the changing of the District’s “complete [computer] system”; and the abandonment of “the appraisal method which has been utilized for valuing land” in the District. The trial court could properly have rejected these justifications and excuses as vague, conclusory, and unsupported by concrete evidence or testimony as to the specific reasons for the changes that allegedly would have been necessitated.
Moreover, there were two alternate methods proposed by appellees which the District could have used to achieve an average level of assessment much closer to exactly 55%. First, the District could have raised the assessment to 100% of market value and then debased that market *309value to 55% (i. e., divided the 60% assessment value by .60 and multiplied the result by 55%). Second, the District could have multiplied the tax bill derived from a 60% level of assessment (which value was stored in the computer) by 5%oths (i. e., multiplied by 55 and then divided by 60). Considering the dearth of adequate foundations for the excuses proffered by the District and the availability to the District of other more accurate methods of computation, we conclude that the trial court was not in error in holding that the Distrifct failed to substantially comply with the original order.
The District’s third contention, that the difference between the results of its method and either method suggested by appellees was de minimis, is also unconvincing. According to the appellees’ suggested methods, 37,073 residential taxpayers would have been taxed $13,468.18 less, and 29,788 residential taxpayers would have been taxed $9,431.14 more than they were taxed under the District’s method. The District would have received $4,055.04 less in real estate taxes using appellees’ suggested methods than it received using its own method (Record at 127). Moreover, the District acknowledged at oral argument that its method of computation, when compared with the methods proposed by appellees, resulted in undercharges to some taxpayers of as much as $35.98, and overcharges to others of as much as $72.14. If, as mentioned in discussing the mootness issue, infra, these variances from a level of assessment of exactly 55% will be perpetrated and compounded in the future, we are unable to declare de minimis in aggregate or as to some individuals an annual overcharge or undercharge, particularly when the error is perpetuated over a period of years, e. g., the $72.14 overcharge or the $35.98 undercharge. Courts have refused to apply the doctrine of “de minimis non curat lex” to situations in which an improper municipal tax assessment extracts only a small amount from each taxpayer, Jenney v. Assessors of Mattapoisett, 322 Mass. 76, 76 N.E.2d 126, 129 (1947), or to tax cases or sales, Cohn v. Little, 199 F.2d 28, 30-31 (8th Cir. 1952). In Lumsden v. Erstine, 205 Ark. 1004, 172 S.W.2d 409, 412 (1943), the court stated:
If a citizen’s rights and property are to be safe, then they must be kept safe against little exactions as well as against large encroachments. The constant drip of water will wear away the largest stone; and if the sovereign by constant inroads in small things is allowed to take the citizen’s property, then the rights of private ownership are gone to the realm of Limbo. Courts are to protect the rights of citizens — that is one of the reasons for the existence of judicial tribunals.
Moreover, it has been held that the de minimis “doctrine applies to questions of minimal damage and transactions between one person and another, not to transactions between a person and the sovereign.” United States v. Lamb, 294 F.Supp. 419, 420 (E.D.Tenn.1968). Consequently, we conclude that the trial court did not err in rejecting the assertion that the amounts involved were de minimis.
This decision will be of limited importance since under Pub.L.No. 93-407, §§ 403(4), 421(a), 88 Stat. 1036 (Sept. 3, 1974), the level of assessment has been set at 100% of fair market value beginning with fiscal year 1976. In addition, we are persuaded that sufficient controversy remains to warrant decision of this case. In this way problems incident to mootness, which are particularly troublesome to a case of this kind, are avoided. E. g., United States v. Munsingwear, Inc., 340 U.S. 36, 71 S.Ct. 104, 95 L.Ed. 36 (1950).
Accordingly, for the foregoing reasons, the trial court’s order of August 6, 1974, is hereby
Affirmed.