ICU4C, ICU4J API Proposal Universal Time Scale

/*
 *******************************************************************************
 * Copyright (C)  2004, International Business Machines Corporation and        *
 * others. All Rights Reserved.                                                *
 *******************************************************************************
 *
 */

package com.ibm.icu.util;

import java.math.BigDecimal;

/** 
 * There are quite a few different conventions for binary datetime, depending on different
 * platforms and protocols. Some of these have severe drawbacks. For example, people using
 * Unix time (seconds since Jan 1, 1970) think that they are safe until near the year 2038.
 * But cases can and do arise where arithmetic manipulations causes serious problems. Consider
 * the computation of the average of two datetimes, for example: if one calculates them with
 * <code>averageTime = (time1 + time2)/2</code>, there will be overflow even with dates
 * around the present. Moreover, even if these problems don't occur, there is the issue of
 * conversion back and forth between different systems.
 *
 * <p>
 * Binary datetimes differ in a number of ways: the datatype, the unit,
 * and the epoch (origin). We'll refer to these as time scales. For example:
 *
 * <table border="1" cellspacing="0" cellpadding="4">
 *  <caption>
 *    <h3>Table 1: Binary Time Scales</h3>
 *
 *  </caption>
 *  <tr>
 *    <th align="left">Source</th>
 *    <th align="left">Datatype</th>
 *    <th align="left">Unit</th>
 *    <th align="left">Epoch</th>
 *  </tr>
 *
 *  <tr>
 *    <td>JAVA_TIME</td>
 *    <td>long</td>
 *    <td>milliseconds</td>
 *    <td>Jan 1, 1970</td>
 *  </tr>
 *  <tr>
 *
 *    <td>UNIX_TIME</td>
 *    <td>int or long</td>
 *    <td>seconds</td>
 *    <td>Jan 1, 1970</td>
 *  </tr>
 *  <tr>
 *    <td>ICU4C</td>
 *
 *    <td>double</td>
 *    <td>milliseconds</td>
 *    <td>Jan 1, 1970</td>
 *  </tr>
 *  <tr>
 *    <td>WINDOWS_FILE_TIME</td>
 *    <td>long</td>
 *
 *    <td>ticks (100 nanoseconds)</td>
 *    <td>Jan 1, 1601</td>
 *  </tr>
 *  <tr>
 *    <td>WINDOWS_DATE_TIME</td>
 *    <td>long</td>
 *    <td>ticks (100 nanoseconds)</td>
 *
 *    <td>Jan 1, 0001</td>
 *  </tr>
 *  <tr>
 *    <td>MAC_OLD_TIME</td>
 *    <td>int</td>
 *    <td>seconds</td>
 *    <td>Jan 1, 1904</td>
 *
 *  </tr>
 *  <tr>
 *    <td>MAC_TIME</td>
 *    <td>double</td>
 *    <td>seconds</td>
 *    <td>Jan 1, 2001</td>
 *  </tr>
 *
 *  <tr>
 *    <td>EXCEL_TIME</td>
 *    <td>?</td>
 *    <td>days</td>
 *    <td>Dec 31, 1899</td>
 *  </tr>
 *  <tr>
 *
 *    <td>DB2_TIME</td>
 *    <td>?</td>
 *    <td>days</td>
 *    <td>Dec 31, 1899</td>
 *  </tr>
 * </table>
 *
 * <p>
 * All of the epochs start at 00:00 am (the earliest possible time on the day in question),
 * and are assumed to be UTC.
 *
 * <p>
 * The ranges for different datatypes are given in the following table (all values in years).
 * The range of years includes the entire range expressible with positive and negative
 * values of the datatype. The range of years for double is the range that would be allowed
 * without losing precision to the corresponding unit.
 *
 * <table border="1" cellspacing="0" cellpadding="4">
 *  <tr>
 *    <th align="left">Units</th>
 *    <th align="left">long</th>
 *    <th align="left">double</th>
 *    <th align="left">int</th>
 *  </tr>
 *
 *  <tr>
 *    <td>1 sec</td>
 *    <td align="right">5.84542×10¹¹</td>
 *    <td align="right">285,420,920.94</td>
 *    <td align="right">136.10</td>
 *  </tr>
 *  <tr>
 *
 *    <td>1 millisecond</td>
 *    <td align="right">584,542,046.09</td>
 *    <td align="right">285,420.92</td>
 *    <td align="right">0.14</td>
 *  </tr>
 *  <tr>
 *    <td>1 microsecond</td>
 *
 *    <td align="right">584,542.05</td>
 *    <td align="right">285.42</td>
 *    <td align="right">0.00</td>
 *  </tr>
 *  <tr>
 *    <td>100 nanoseconds (tick)</td>
 *    <td align="right">58,454.20</td>
 *    <td align="right">28.54</td>
 *    <td align="right">0.00</td>
 *  </tr>
 *  <tr>
 *    <td>1 nanosecond</td>
 *    <td align="right">584.5420461</td>
 *    <td align="right">0.2854</td>
 *    <td align="right">0.00</td>
 *  </tr>
 * </table>
 *
 * <p>
 * These functions implement a universal time scale which can be used as a 'pivot',
 * and provide conversion functions to and from all other major time scales.
 * This datetimes to be converted to the pivot time, safely manipulated,
 * and converted back to any other datetime time scale.
 *
 *<p>
 * So what to use for this pivot? Java time has plenty of range, but cannot represent
 * Windows datetimes without severe loss of precision. ICU4C time addresses this by using a
 * double that is otherwise equivalent to the Java time. However, there are disadvantages
 * with doubles. They provide for much more graceful degradation in arithmetic operations.
 * But they only have 53 bits of accuracy, which means that they will lose precision when
 * converting back and forth to ticks. What would really be nice would be a
 * <code>long double</code> (80 bits -- 64 bit mantissa), but that is not supported on most systems.
 *
 *<p>
 * The Unix extended time uses a structure with two components: time in seconds and a
 * fractional field (microseconds). However, this is clumsy, slow, and
 * prone to error (you always have to keep track of overflow and underflow in the
 * fractional field). <code>BigDecimal</code> would allow for arbitrary precision and arbitrary range,
 * but we would not want to use this as the normal type, because it is slow and does not
 * have a fixed size.
 *
 *<p>
 * Because of these issues, we ended up concluding that the Windows datetime would be the
 * best pivot. However, we use the full range allowed by the datatype, allowing for
 * datetimes back to 29,000 BC and up to 29,000 AD. This time scale is very fine grained,
 * does not lose precision, and covers a range that will meet almost all requirements.
 * It will not handle the range that Java times would, but frankly, being able to handle dates
 * before 29,000 BC or after 29,000 AD is of very limited interest. However, for those cases,
 * we also allow conversion to an optional <code>BigDecimal</code> format that would have arbitrary
 * precision and range.
 *
 */

public final class UniversalTimeScale
{
    /**
     * Used in the JDK. Data is a <code>long</code>. Value
     * is milliseconds since January 1, 1970.
     *
     * @draft ICU 3.2
     */

    static final int JAVA_TIME = 0;
    /**
     * Used in Unix systems. Data is an <code>int> or a <code>long</code>. Value
     * is seconds since January 1, 1970.
     *
     * @draft ICU 3.2
     */
    static final int UNIX_TIME = 1;

    /**
     * Used in the ICU4C. Data is a <code>double</code>. Value
     * is milliseconds since January 1, 1970.
     *
     * @draft ICU 3.2
     */
    static final int ICU4C_TIME = 2;

    /**
     * Used in Windows for file times. Data is a <code>long</code>. Value
     * is ticks (1 tick == 100 nanoseconds) since January 1, 1601.
     *
     * @draft ICU 3.2
     */
    static final int WINDOWS_FILE_TIME = 3;

    /**
     * Used in Windows for date time (?). Data is a <code>long</code>. Value
     * is ticks (1 tick == 100 nanoseconds) since January 1, 0001.
     *
     * @draft ICU 3.2
     */
    static final int WINDOWS_DATE_TIME = 4;

    /**
     * Used in older Macintosh systems. Data is an <code>int</code>. Value
     * is seconds since January 1, 1904.
     *
     * @draft ICU 3.2
     */
    static final int MAC_OLD_TIME = 5;

    /**
     * Used in the JDK. Data is a <code>double</code>. Value
     * is milliseconds since January 1, 2001.
     *
     * @draft ICU 3.2
     */
    static final int MAC_TIME = 6;

    /**
     * Used in Excel. Data is a <code>?unknown?</code>. Value
     * is days since December 31, 1899.
     *
     * @draft ICU 3.2
     */
    static final int EXCEL_TIME = 7;

    /**
     * Used in DB2. Data is a <code>?unknown?</code>. Value
     * is days since December 31, 1899.
     *
     * @draft ICU 3.2
     */
    static final int DB2_TIME = 8;
    
    /* Conversion to 'universal time scale' */

    /**
     * Convert a <code>double</code> datetime from the given time scale to the universal time scale.
     *
     * @param otherTime The <code>double</code> datetime
     * @param timeScale The time scale to convert from
     * 
     * @return The datetime converted to the universal time scale
     *
     * @draft ICU 3.2
     */
    static long from(double otherTime, int timeScale);

    /**
     * Convert a <code>long</code> datetime from the given time scale to the universal time scale.
     *
     * @param otherTime The <code>long</code> datetime
     * @param timeScale The time scale to convert from
     * 
     * @return The datetime converted to the universal time scale
     *
     * @draft ICU 3.2
     */
    static long from(long otherTime, int timeScale);

    /**
     * Convert a <code>double</code> datetime from the given time scale to the universal time scale.
     * All calculations are done using <code>BigDecimal</code> to guarantee that the value
     * does not go out of range.
     *
     * @param otherTime The <code>double</code> datetime
     * @param timeScale The time scale to convert from
     * 
     * @return The datetime converted to the universal time scale
     *
     * @draft ICU 3.2
     */
    static BigDecimal bigDecimalFrom(double otherTime, int timeScale);

    /**
     * Convert a <code>long</code> datetime from the given time scale to the universal time scale.
     * All calculations are done using <code>BigDecimal</code> to guarantee that the value
     * does not go out of range.
     *
     * @param otherTime The <code>long</code> datetime
     * @param timeScale The time scale to convert from
     * 
     * @return The datetime converted to the universal time scale
     *
     * @draft ICU 3.2
     */
    static BigDecimal bigDecimalFrom(long otherTime, int timeScale);

    /* Conversion from 'universal time scale' */

    /**
     * Convert a datetime from the universal time scale to a <code>double</code> in the given time scale.
     *
     * @param universal The datetime in the universal time scale
     * @param timeScale The time scale to convert to
     * 
     * @return The datetime converted to the given time scale
     *
     * @draft ICU 3.2
     */
    static double toDouble(long universalTime, int timeScale);
    
    /**
     * Convert a datetime from the universal time scale stored as a <code>BigDecimal</code> to a
     * <code>double</code> in the given time scale.
     *
     * @param universal The datetime in the universal time scale
     * @param timeScale The time scale to convert to
     * 
     * @return The datetime converted to the given time scale
     *
     * @draft ICU 3.2
     */
    static double toDouble(BigDecimal universalTime, int timeScale);

    /**
     * Convert a datetime from the universal time scale stored as a <code>BigDecimal</code> to a
     * <code>long</code> in the given time scale.
     *
     * @param universal The datetime in the universal time scale
     * @param timeScale The time scale to convert to
     * 
     * @return The datetime converted to the given time scale
     *
     * @draft ICU 3.2
     */
    static long toLong(long universalTime, int timeScale);
    
    /**
     * Convert a datetime from the universal time scale to a <code>long</code> in the given time scale.
     *
     * @param universal The datetime in the universal time scale
     * @param timeScale The time scale to convert to
     * 
     * @return The datetime converted to the given time scale
     *
     * @draft ICU 3.2
     */
    static long toLong(BigDecimal universalTime, int timeScale);
}

/*
*******************************************************************************
* Copyright (C) 2004, International Business Machines Corporation and
* others. All Rights Reserved.
*******************************************************************************
*/

#ifndef UTMSCALE_H
#define UTMSCALE_H

#include "unicode/utypes.h"

/** 
 * \file
 * \brief C API: Universal Time Scale
 *
 * There are quite a few different conventions for binary datetime, depending on different
 * platforms and protocols. Some of these have severe drawbacks. For example, people using
 * Unix time (seconds since Jan 1, 1970) think that they are safe until near the year 2038.
 * But cases can and do arise where arithmetic manipulations causes serious problems. Consider
 * the computation of the average of two datetimes, for example: if one calculates them with
 * <code>averageTime = (time1 + time2)/2</code>, there will be overflow even with dates
 * around the present. Moreover, even if these problems don't occur, there is the issue of
 * conversion back and forth between different systems.
 *
 * <p>
 * Binary datetimes differ in a number of ways: the datatype, the unit,
 * and the epoch (origin). We'll refer to these as time scales. For example:
 *
 * <table border="1" cellspacing="0" cellpadding="4">
 *  <caption>
 *    <h3>Table 1: Binary Time Scales</h3>
 *
 *  </caption>
 *  <tr>
 *    <th align="left">Source</th>
 *    <th align="left">Datatype</th>
 *    <th align="left">Unit</th>
 *    <th align="left">Epoch</th>
 *  </tr>
 *
 *  <tr>
 *    <td>JAVA_TIME</td>
 *    <td>int64_t</td>
 *    <td>milliseconds</td>
 *    <td>Jan 1, 1970</td>
 *  </tr>
 *  <tr>
 *
 *    <td>UNIX_TIME</td>
 *    <td>int32_t or int64_t</td>
 *    <td>seconds</td>
 *    <td>Jan 1, 1970</td>
 *  </tr>
 *  <tr>
 *    <td>ICU4C_TIME</td>
 *
 *    <td>double</td>
 *    <td>milliseconds</td>
 *    <td>Jan 1, 1970</td>
 *  </tr>
 *  <tr>
 *    <td>WINDOWS_FILE_TIME</td>
 *    <td>int64_t</td>
 *
 *    <td>ticks (100 nanoseconds)</td>
 *    <td>Jan 1, 1601</td>
 *  </tr>
 *  <tr>
 *    <td>WINDOWS_DATE_TIME</td>
 *    <td>int64_t</td>
 *    <td>ticks (100 nanoseconds)</td>
 *
 *    <td>Jan 1, 0001</td>
 *  </tr>
 *  <tr>
 *    <td>MAC_OLD_TIME</td>
 *    <td>int32_t</td>
 *    <td>seconds</td>
 *    <td>Jan 1, 1904</td>
 *
 *  </tr>
 *  <tr>
 *    <td>MAC_TIME</td>
 *    <td>double</td>
 *    <td>seconds</td>
 *    <td>Jan 1, 2001</td>
 *  </tr>
 *
 *  <tr>
 *    <td>EXCEL_TIME</td>
 *    <td>?</td>
 *    <td>days</td>
 *    <td>Dec 31, 1899</td>
 *  </tr>
 *  <tr>
 *
 *    <td>DB2_TIME</td>
 *    <td>?</td>
 *    <td>days</td>
 *    <td>Dec 31, 1899</td>
 *  </tr>
 * </table>
 *
 * <p>
 * All of the epochs start at 00:00 am (the earliest possible time on the day in question),
 * and are assumed to be UTC.
 *
 * <p>
 * The ranges for different datatypes are given in the following table (all values in years).
 * The range of years includes the entire range expressible with positive and negative
 * values of the datatype. The range of years for double is the range that would be allowed
 * without losing precision to the corresponding unit.
 *
 * <table border="1" cellspacing="0" cellpadding="4">
 *  <tr>
 *    <th align="left">Units</th>
 *    <th align="left">int64_t</th>
 *    <th align="left">double</th>
 *    <th align="left">int32_t</th>
 *  </tr>
 *
 *  <tr>
 *    <td>1 sec</td>
 *    <td align="right">5.84542×10¹¹</td>
 *    <td align="right">285,420,920.94</td>
 *    <td align="right">136.10</td>
 *  </tr>
 *  <tr>
 *
 *    <td>1 millisecond</td>
 *    <td align="right">584,542,046.09</td>
 *    <td align="right">285,420.92</td>
 *    <td align="right">0.14</td>
 *  </tr>
 *  <tr>
 *    <td>1 microsecond</td>
 *
 *    <td align="right">584,542.05</td>
 *    <td align="right">285.42</td>
 *    <td align="right">0.00</td>
 *  </tr>
 *  <tr>
 *    <td>100 nanoseconds (tick)</td>
 *    <td align="right">58,454.20</td>
 *    <td align="right">28.54</td>
 *    <td align="right">0.00</td>
 *  </tr>
 *  <tr>
 *    <td>1 nanosecond</td>
 *    <td align="right">584.5420461</td>
 *    <td align="right">0.2854</td>
 *    <td align="right">0.00</td>
 *  </tr>
 * </table>
 *
 * <p>
 * These functions implement a universal time scale which can be used as a 'pivot',
 * and provide conversion functions to and from all other major time scales.
 * This datetimes to be converted to the pivot time, safely manipulated,
 * and converted back to any other datetime time scale.
 *
 *<p>
 * So what to use for this pivot? Java time has plenty of range, but cannot represent
 * Windows datetimes without severe loss of precision. ICU4C time addresses this by using a
 * double that is otherwise equivalent to the Java time. However, there are disadvantages
 * with doubles. They provide for much more graceful degradation in arithmetic operations.
 * But they only have 53 bits of accuracy, which means that they will lose precision when
 * converting back and forth to ticks. What would really be nice would be a
 * long double (80 bits -- 64 bit mantissa), but that is not supported on most systems.
 *
 *<p>
 * The Unix extended time uses a structure with two components: time in seconds and a
 * fractional field (microseconds). However, this is clumsy, slow, and
 * prone to error (you always have to keep track of overflow and underflow in the
 * fractional field). BigDecimal would allow for arbitrary precision and arbitrary range,
 * but we do not want to use this as the normal type, because it is slow and does not
 * have a fixed size.
 *
 *<p>
 * Because of these issues, we ended up concluding that the Windows datetime would be the
 * best pivot. However, we use the full range allowed by the datatype, allowing for
 * datetimes back to 29,000 BC and up to 29,000 AD. This time scale is very fine grained,
 * does not lose precision, and covers a range that will meet almost all requirements.
 * It will not handle the range that Java times do, but frankly, being able to handle dates
 * before 29,000 BC or after 29,000 AD is of very limited interest.
 *
 */

/**
 * UDateTimeScale values are used to specify the time scale used for
 * conversion into or out if the universal time scale.
 *
 * @draft ICU 3.2
 */
typedef enum UDateTimeScale {
    /**
     * Used in the JDK. Data is a Java <code>long</code> (<code>int64_t</code>). Value
     * is milliseconds since January 1, 1970.
     *
     * @draft ICU 3.2
     */
    JAVA_TIME = 0,

    /**
     * Used on Unix systems. Data is <code>int32_t</code> or <code>int64_t</code>. Value
     * is seconds since January 1, 1970.
     *
     * @draft ICU 3.2
     */
    UNIX_TIME,
    
    /**
     * Used in IUC4C. Data is a <code>double</code>. Value
     * is milliseconds since January 1, 1970.
     *
     * @draft ICU 3.2
     */
    ICU4C_TIME,
    
    /**
     * Used in Windows for file times. Data is an <code>int64_t</code>. Value
     * is ticks (1 tick == 100 nanoseconds) since January 1, 1601.
     *
     * @draft ICU 3.2
     */
    WINDOWS_FILE_TIME,
    
    /**
     * Used in Windows for dates and times (?). Data is an <code>int64_t</code>. Value
     * is ticks (1 tick == 100 nanoseconds) since January 1, 0001.
     *
     * @draft ICU 3.2
     */
    WINDOWS_DATE_TIME,
    
    /**
     * Used in older Macintosh systems. Data is an <code>int32_t</code>. Value
     * is seconds since January 1, 1904.
     *
     * @draft ICU 3.2
     */
    MAC_OLD_TIME,
    
    /**
     * Used in newer Macintosh systems. Data is a <code>double</code>. Value
     * is seconds since January 1, 2001.
     *
     * @draft ICU 3.2
     */
    MAC_TIME,
    
    /**
     * Used in Excel. Data is an <code>?unknown?</code>. Value
     * is days since December 31, 1899.
     *
     * @draft ICU 3.2
     */
    EXCEL_TIME,
    
    /**
     * Used in DB2. Data is an <code>?unknown?</code>. Value
     * is days since December 31, 1899.
     *
     * @draft ICU 3.2
     */
    DB2_TIME
} UDateTimeScale;

/* Conversion to 'universal time scale' */

/**
 * Convert a <code>double</code> datetime from the given time scale to the universal time scale.
 *
 * @param otherTime The <code>double</code> datetime
 * @param timeScale The time scale to convert from
 * @param status The status code. Set to <code>U_ILLEGAL_ARGUMENT_ERROR</code> if the conversion is out of range.
 * 
 * @return The datetime converted to the universal time scale
 *
 * @draft ICU 3.2
 */
U_DRAFT int64_t  U_EXPORT2
    utmscale_fromDouble(double otherTime, UDateTimeScale timeScale, UErrorCode *status);

/**
 * Convert a <code>int64_t</code> datetime from the given time scale to the universal time scale.
 *
 * @param otherTime The <code>int64_t</code> datetime
 * @param timeScale The time scale to convert from
 * @param status The status code. Set to <code>U_ILLEGAL_ARGUMENT_ERROR</code> if the conversion is out of range.
 * 
 * @return The datetime converted to the universal time scale
 *
 * @draft ICU 3.2
 */
U_DRAFT int64_t U_EXPORT2
    utmscale_fromInt64(int64_t otherTime, UDateTimeScale timeScale, UErrorCode *status);

/* Conversion from 'universal time scale' */

/**
 * Convert a datetime from the universal time scale to a <code>double</code> in the given time scale.
 *
 * @param universal The datetime in the universal time scale
 * @param timeScale The time scale to convert to
 * @param status The status code. Set to <code>U_ILLEGAL_ARGUMENT_ERROR</code> if the conversion is out of range.
 * 
 * @return The datetime converted to the given time scale
 *
 * @draft ICU 3.2
 */
U_DRAFT double U_EXPORT2
    utmscale_toDouble(int64_t universalTime, UDateTimeScale timeScale, UErrorCode *status);

/**
 * Convert a datetime from the universal time scale to a <code>int64_t</code> in the given time scale.
 *
 * @param universal The datetime in the universal time scale
 * @param timeScale The time scale to convert to
 * @param status The status code. Set to <code>U_ILLEGAL_ARGUMENT_ERROR</code> if the conversion is out of range.
 * 
 * @return The datetime converted to the given time scale
 *
 * @draft ICU 3.2
 */
U_DRAFT int64_t U_EXPORT2
    utmscale_toInt64(int64_t universalTime, UDateTimeScale timeScale, UErrorCode *status);

#endif