位掩碼（BitMask）

位運算在實際開發中用得很少，主要原因還是它對於不熟悉的人不好讀不好懂不好計算，如果不經常實踐會生疏。但它的優點自然是計算快，代碼更少。在某些地方它的優勢會更加明顯比如如下代碼：

public class NewPermission {
	// 是否允許查詢，二進制第1位，0表示否，1表示是
	public static final int ALLOW_SELECT = 1 << 0; // 0001
	
	// 是否允許新增，二進制第2位，0表示否，1表示是
	public static final int ALLOW_INSERT = 1 << 1; // 0010
	
	// 是否允許修改，二進制第3位，0表示否，1表示是
	public static final int ALLOW_UPDATE = 1 << 2; // 0100
	
	// 是否允許刪除，二進制第4位，0表示否，1表示是
	public static final int ALLOW_DELETE = 1 << 3; // 1000
	
	// 存儲目前的權限狀態
	private int flag;

	/**
	 *  重新設置權限
	 */
	public void setPermission(int permission) {
		flag = permission;
	}

	/**
	 *  添加一項或多項權限
	 */
	public void enable(int permission) {
		flag |= permission;
	}
	
	/**
	 *  刪除一項或多項權限
	 */
	public void disable(int permission) {
		flag &= ~permission;
	}
	
	/**
	 *  是否擁某些權限
	 */
	public boolean isAllow(int permission) {
		return (flag & permission) == permission;
	}
	
	/**
	 *  是否禁用了某些權限
	 */
	public boolean isNotAllow(int permission) {
		return (flag & permission) == 0;
	}
	
	/**
	 *  是否僅僅擁有某些權限
	 */
	public boolean isOnlyAllow(int permission) {
		return flag == permission;
	}
}

在業務上，權限可以有多個，通過普通方式實現上面的業務功能也不會太麻煩。但上面這種取巧的方式在android的源碼中是不少的，可以說是隨處可見，比如EditText的InputType：
 

/*
 * Copyright (C) 2008 The Android Open Source Project
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *      http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

package android.text;

import android.text.TextUtils;

/**
 * Bit definitions for an integer defining the basic content type of text
 * held in an {@link Editable} object. Supported classes may be combined
 * with variations and flags to indicate desired behaviors.
 *
 * 
Examples

* *
*
A password field with with the password visible to the user: *
inputType = TYPE_CLASS_TEXT | * TYPE_TEXT_VARIATION_VISIBLE_PASSWORD * *
A multi-line postal address with automatic capitalization: *
inputType = TYPE_CLASS_TEXT | * TYPE_TEXT_VARIATION_POSTAL_ADDRESS | * TYPE_TEXT_FLAG_MULTI_LINE * *
A time field: *
inputType = TYPE_CLASS_DATETIME | * TYPE_DATETIME_VARIATION_TIME *
*/ public interface InputType { /** * Mask of bits that determine the overall class * of text being given. Currently supported classes are: * {@link #TYPE_CLASS_TEXT}, {@link #TYPE_CLASS_NUMBER}, * {@link #TYPE_CLASS_PHONE}, {@link #TYPE_CLASS_DATETIME}. *
IME authors: If the class is not one you * understand, assume {@link #TYPE_CLASS_TEXT} with NO variation * or flags.

*/ public static final int TYPE_MASK_CLASS = 0x0000000f; /** * Mask of bits that determine the variation of * the base content class. */ public static final int TYPE_MASK_VARIATION = 0x00000ff0; /** * Mask of bits that provide addition bit flags * of options. */ public static final int TYPE_MASK_FLAGS = 0x00fff000; /** * Special content type for when no explicit type has been specified. * This should be interpreted to mean that the target input connection * is not rich, it can not process and show things like candidate text nor * retrieve the current text, so the input method will need to run in a * limited "generate key events" mode, if it supports it. Note that some * input methods may not support it, for example a voice-based input * method will likely not be able to generate key events even if this * flag is set. */ public static final int TYPE_NULL = 0x00000000; // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- /** * Class for normal text. This class supports the following flags (only * one of which should be set): * {@link #TYPE_TEXT_FLAG_CAP_CHARACTERS}, * {@link #TYPE_TEXT_FLAG_CAP_WORDS}, and. * {@link #TYPE_TEXT_FLAG_CAP_SENTENCES}. It also supports the * following variations: * {@link #TYPE_TEXT_VARIATION_NORMAL}, and * {@link #TYPE_TEXT_VARIATION_URI}. If you do not recognize the * variation, normal should be assumed. */ public static final int TYPE_CLASS_TEXT = 0x00000001; /** * Flag for {@link #TYPE_CLASS_TEXT}: capitalize all characters. Overrides * {@link #TYPE_TEXT_FLAG_CAP_WORDS} and * {@link #TYPE_TEXT_FLAG_CAP_SENTENCES}. This value is explicitly defined * to be the same as {@link TextUtils#CAP_MODE_CHARACTERS}. Of course, * this only affects languages where there are upper-case and lower-case letters. */ public static final int TYPE_TEXT_FLAG_CAP_CHARACTERS = 0x00001000; /** * Flag for {@link #TYPE_CLASS_TEXT}: capitalize the first character of * every word. Overrides {@link #TYPE_TEXT_FLAG_CAP_SENTENCES}. This * value is explicitly defined * to be the same as {@link TextUtils#CAP_MODE_WORDS}. Of course, * this only affects languages where there are upper-case and lower-case letters. */ public static final int TYPE_TEXT_FLAG_CAP_WORDS = 0x00002000; /** * Flag for {@link #TYPE_CLASS_TEXT}: capitalize the first character of * each sentence. This value is explicitly defined * to be the same as {@link TextUtils#CAP_MODE_SENTENCES}. For example * in English it means to capitalize after a period and a space (note that other * languages may have different characters for period, or not use spaces, * or use different grammatical rules). Of course, * this only affects languages where there are upper-case and lower-case letters. */ public static final int TYPE_TEXT_FLAG_CAP_SENTENCES = 0x00004000; /** * Flag for {@link #TYPE_CLASS_TEXT}: the user is entering free-form * text that should have auto-correction applied to it. Without this flag, * the IME will not try to correct typos. You should always set this flag * unless you really expect users to type non-words in this field, for * example to choose a name for a character in a game. * Contrast this with {@link #TYPE_TEXT_FLAG_AUTO_COMPLETE} and * {@link #TYPE_TEXT_FLAG_NO_SUGGESTIONS}: * {@code TYPE_TEXT_FLAG_AUTO_CORRECT} means that the IME will try to * auto-correct typos as the user is typing, but does not define whether * the IME offers an interface to show suggestions. */ public static final int TYPE_TEXT_FLAG_AUTO_CORRECT = 0x00008000; /** * Flag for {@link #TYPE_CLASS_TEXT}: the text editor (which means * the application) is performing auto-completion of the text being entered * based on its own semantics, which it will present to the user as they type. * This generally means that the input method should not be showing * candidates itself, but can expect the editor to supply its own * completions/candidates from * {@link android.view.inputmethod.InputMethodSession#displayCompletions * InputMethodSession.displayCompletions()} as a result of the editor calling * {@link android.view.inputmethod.InputMethodManager#displayCompletions * InputMethodManager.displayCompletions()}. * Note the contrast with {@link #TYPE_TEXT_FLAG_AUTO_CORRECT} and * {@link #TYPE_TEXT_FLAG_NO_SUGGESTIONS}: * {@code TYPE_TEXT_FLAG_AUTO_COMPLETE} means the editor should show an * interface for displaying suggestions, but instead of supplying its own * it will rely on the Editor to pass completions/corrections. */ public static final int TYPE_TEXT_FLAG_AUTO_COMPLETE = 0x00010000; /** * Flag for {@link #TYPE_CLASS_TEXT}: multiple lines of text can be * entered into the field. If this flag is not set, the text field * will be constrained to a single line. The IME may also choose not to * display an enter key when this flag is not set, as there should be no * need to create new lines. */ public static final int TYPE_TEXT_FLAG_MULTI_LINE = 0x00020000; /** * Flag for {@link #TYPE_CLASS_TEXT}: the regular text view associated * with this should not be multi-line, but when a fullscreen input method * is providing text it should use multiple lines if it can. */ public static final int TYPE_TEXT_FLAG_IME_MULTI_LINE = 0x00040000; /** * Flag for {@link #TYPE_CLASS_TEXT}: the input method does not need to * display any dictionary-based candidates. This is useful for text views that * do not contain words from the language and do not benefit from any * dictionary-based completions or corrections. It overrides the * {@link #TYPE_TEXT_FLAG_AUTO_CORRECT} value when set. * Please avoid using this unless you are certain this is what you want. * Many input methods need suggestions to work well, for example the ones * based on gesture typing. Consider clearing * {@link #TYPE_TEXT_FLAG_AUTO_CORRECT} instead if you just do not * want the IME to correct typos. * Note the contrast with {@link #TYPE_TEXT_FLAG_AUTO_CORRECT} and * {@link #TYPE_TEXT_FLAG_AUTO_COMPLETE}: * {@code TYPE_TEXT_FLAG_NO_SUGGESTIONS} means the IME should never * show an interface to display suggestions. Most IMEs will also take this to * mean they should not try to auto-correct what the user is typing. */ public static final int TYPE_TEXT_FLAG_NO_SUGGESTIONS = 0x00080000; // ---------------------------------------------------------------------- /** * Default variation of {@link #TYPE_CLASS_TEXT}: plain old normal text. */ public static final int TYPE_TEXT_VARIATION_NORMAL = 0x00000000; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering a URI. */ public static final int TYPE_TEXT_VARIATION_URI = 0x00000010; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering an e-mail address. */ public static final int TYPE_TEXT_VARIATION_EMAIL_ADDRESS = 0x00000020; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering the subject line of * an e-mail. */ public static final int TYPE_TEXT_VARIATION_EMAIL_SUBJECT = 0x00000030; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering a short, possibly informal * message such as an instant message or a text message. */ public static final int TYPE_TEXT_VARIATION_SHORT_MESSAGE = 0x00000040; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering the content of a long, possibly * formal message such as the body of an e-mail. */ public static final int TYPE_TEXT_VARIATION_LONG_MESSAGE = 0x00000050; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering the name of a person. */ public static final int TYPE_TEXT_VARIATION_PERSON_NAME = 0x00000060; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering a postal mailing address. */ public static final int TYPE_TEXT_VARIATION_POSTAL_ADDRESS = 0x00000070; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering a password. */ public static final int TYPE_TEXT_VARIATION_PASSWORD = 0x00000080; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering a password, which should * be visible to the user. */ public static final int TYPE_TEXT_VARIATION_VISIBLE_PASSWORD = 0x00000090; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering text inside of a web form. */ public static final int TYPE_TEXT_VARIATION_WEB_EDIT_TEXT = 0x000000a0; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering text to filter contents * of a list etc. */ public static final int TYPE_TEXT_VARIATION_FILTER = 0x000000b0; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering text for phonetic * pronunciation, such as a phonetic name field in contacts. This is mostly * useful for languages where one spelling may have several phonetic * readings, like Japanese. */ public static final int TYPE_TEXT_VARIATION_PHONETIC = 0x000000c0; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering e-mail address inside * of a web form. This was added in * {@link android.os.Build.VERSION_CODES#HONEYCOMB}. An IME must target * this API version or later to see this input type; if it doesn't, a request * for this type will be seen as {@link #TYPE_TEXT_VARIATION_EMAIL_ADDRESS} * when passed through {@link android.view.inputmethod.EditorInfo#makeCompatible(int) * EditorInfo.makeCompatible(int)}. */ public static final int TYPE_TEXT_VARIATION_WEB_EMAIL_ADDRESS = 0x000000d0; /** * Variation of {@link #TYPE_CLASS_TEXT}: entering password inside * of a web form. This was added in * {@link android.os.Build.VERSION_CODES#HONEYCOMB}. An IME must target * this API version or later to see this input type; if it doesn't, a request * for this type will be seen as {@link #TYPE_TEXT_VARIATION_PASSWORD} * when passed through {@link android.view.inputmethod.EditorInfo#makeCompatible(int) * EditorInfo.makeCompatible(int)}. */ public static final int TYPE_TEXT_VARIATION_WEB_PASSWORD = 0x000000e0; // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- /** * Class for numeric text. This class supports the following flags: * {@link #TYPE_NUMBER_FLAG_SIGNED} and * {@link #TYPE_NUMBER_FLAG_DECIMAL}. It also supports the following * variations: {@link #TYPE_NUMBER_VARIATION_NORMAL} and * {@link #TYPE_NUMBER_VARIATION_PASSWORD}. *

IME authors: If you do not recognize * the variation, normal should be assumed.

*/ public static final int TYPE_CLASS_NUMBER = 0x00000002; /** * Flag of {@link #TYPE_CLASS_NUMBER}: the number is signed, allowing * a positive or negative sign at the start. */ public static final int TYPE_NUMBER_FLAG_SIGNED = 0x00001000; /** * Flag of {@link #TYPE_CLASS_NUMBER}: the number is decimal, allowing * a decimal point to provide fractional values. */ public static final int TYPE_NUMBER_FLAG_DECIMAL = 0x00002000; // ---------------------------------------------------------------------- /** * Default variation of {@link #TYPE_CLASS_NUMBER}: plain normal * numeric text. This was added in * {@link android.os.Build.VERSION_CODES#HONEYCOMB}. An IME must target * this API version or later to see this input type; if it doesn't, a request * for this type will be dropped when passed through * {@link android.view.inputmethod.EditorInfo#makeCompatible(int) * EditorInfo.makeCompatible(int)}. */ public static final int TYPE_NUMBER_VARIATION_NORMAL = 0x00000000; /** * Variation of {@link #TYPE_CLASS_NUMBER}: entering a numeric password. * This was added in {@link android.os.Build.VERSION_CODES#HONEYCOMB}. An * IME must target this API version or later to see this input type; if it * doesn't, a request for this type will be dropped when passed * through {@link android.view.inputmethod.EditorInfo#makeCompatible(int) * EditorInfo.makeCompatible(int)}. */ public static final int TYPE_NUMBER_VARIATION_PASSWORD = 0x00000010; // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- /** * Class for a phone number. This class currently supports no variations * or flags. */ public static final int TYPE_CLASS_PHONE = 0x00000003; // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- // ---------------------------------------------------------------------- /** * Class for dates and times. It supports the * following variations: * {@link #TYPE_DATETIME_VARIATION_NORMAL} * {@link #TYPE_DATETIME_VARIATION_DATE}, and * {@link #TYPE_DATETIME_VARIATION_TIME}. */ public static final int TYPE_CLASS_DATETIME = 0x00000004; /** * Default variation of {@link #TYPE_CLASS_DATETIME}: allows entering * both a date and time. */ public static final int TYPE_DATETIME_VARIATION_NORMAL = 0x00000000; /** * Default variation of {@link #TYPE_CLASS_DATETIME}: allows entering * only a date. */ public static final int TYPE_DATETIME_VARIATION_DATE = 0x00000010; /** * Default variation of {@link #TYPE_CLASS_DATETIME}: allows entering * only a time. */ public static final int TYPE_DATETIME_VARIATION_TIME = 0x00000020; }

由於位掩碼方式又一個很大的問題，就是類型安全，如果不對類型安全控制後果會難以想象，因此枚舉被用進去，以對參數類型進行控制，源碼當中使用的是@IntDef，而EnumSet作為新的實現形式，在任何方面都比位掩碼方式差不多或者更好，可讀性，性能，以及類型安全等。權限邏輯如下：

public class NewPermission2 {

  private EnumSet permissions = EnumSet.noneOf(Permission.class);

  private enum Permission {
    SELECT, INSERT, UPDATE, DELETE;

    public static final EnumSet ALL_PERMISSIONS = EnumSet.allOf(Permission.class);
    public static final EnumSet VIP_PERMISSIONS =
        EnumSet.range(Permission.SELECT, Permission.UPDATE);
  }

  /**
   * 重新設置權限
   */
  public void setPermission(EnumSet s) {
    this.permissions.clear();
    this.permissions = s;
  }

  /**
   * 刪除一項或多項權限
   */
  public void disable(Permission permission) {
    permissions.remove(permission);
  }

  /**
   * 是否擁某些權限
   */
  public boolean isAllow(Permission s) {
    return permissions.contains(s);
  }

  /**
   * 是否是游客
   */
  public boolean isVip(Permission s) {
    return Permission.VIP_PERMISSIONS.contains(s);
  }
}

EnumSet具體實現可參看源碼。個人覺得這種方式最大的好處，在於可控，無論是對業務邏輯很好理解，還是對代碼實現的方便，各個方面來看，都是要比位掩碼這種取巧方式的官方化。可是對於android開發的性能方面：

1 存在個別反射的身影

2 EnumSet是實現Set的集合類

基於以上兩點，那麼在設計用到Enumset的類的時候，還是需要謹慎一些，盡管官方給出了很可靠的保證。