Images, audio, and video are memory-intensive items, and many apps have these features. There are tools to help monitor memory use (e.g. Eclipse MAT, Android DDMS), and these tools are good for diagnosing problems, but you still need to understand enough to be able to fix memory leaks.

Here are some ideas for reducing the memory footprint of your application:

Reduce image sizes:  Lower the width, height, pixel bit-depth, and compress your images as much as you can.  Of course when an image is uncompressed and loaded into a Bitmap object then it takes up more memory, but you can reduce the memory footprint by lowering the image quality (for example see: View.setDrawingCacheQuality and View.DRAWING_CACHE_QUALITY_LOW) or even disabling the cache on views containing images.

Lower audio and video bitrates: I don’t know this for a fact, but I would guess that a lower-bitrate audio stream may have a smaller memory requirement during playback.  For example a mono-48 kbit/second audio file would require decoded fewer samples per second than a stereo-192 kbit/second file. (Please comment to this post if you test this theory or know the answer.)

Destroy or reuse objects: When you can, re-use objects and make sure that old objects are fully-destroyed. when they are no used anymore.  Even better, never create objects in the first place if possible.  This is especially true for bitmap objects – be sure to call the Bitmap.recycle() method. Remember to clear callback methods of objects before destroying them, because otherwise an object may not be properly returned to the memory heap during a java garbage collection operation.

Use final and static:  Virtual methods take up more space, and are slower, than static methods.  Final variables and arrays are stored in code space and not the memory heap.  Granted the difference is very small compared with 16MB of space, but every little bit counts!

Separate applications for localization:  If you are developing apps for multiple languages, consider creating separate applications instead of including all the language-specific strings, images, audio files, and videos in a single application.

Rely on external storage:  If you know that there is smart-card memory available, use that to store data instead of in memory.

Revert to an earlier Android SDK: When we reverted our most-memory-challenged Android application from Android 2.2 to Android 1.5, we gained two important things: first, we are now compatible with almost all existing Android phones; and second we reduced our memory footprint by almost a megabyte.  This latter statement is extremely interesting, since it indicates that the Android framework is getting bloated by all the new features added between versions 1.5 and 2.2.

GORGES CEO Chris Grant meets with Congressman Hinchey

When the Congressman asked me about GORGES, I let him know about our rapid business growth over the last 5 years, primarily due to our innovation in Internet technologies, and our ability to help our customers businesses grow.

I also mentioned that we are actively keeping software development work from heading to offshore development companies, as our expert developers, low cost of living, and proven processes compete very well on the global market.  While our hourly rates may be higher than India, for example, our total cost to complete a project is often lower.

There are many reasons for this, foremost among them dedicated project management, a willingness to understand our clients’ business needs to create the right-fitting solution, as well as in-house experienced web developers and designers.

After hearing this, the Congressman remarked ‘Now this is the kind of company that America needs to create jobs and move this economy along’.

Just what we were thinking….

Our servers use utilities such as fail2ban or denyhosts that look for repeated failed login attempts, and once found they direct the firewall service to ban the originating IP addresses.  However this technique fails when the attack is distributed among thousands of compromised “zombie” computers that are doing the bidding of a malicious hacker.

Our log files correctly diagnosed each attempt from an individual IP address, but a new attempt was immediately started from a different IP address.  We were clearly looking at an attack coordinated from a single unknown source.

There are several ways to reduce or thwart these attacks, including:

• never allow remote root logins (but attacks still occur on non-root-user names)
• have a chroot jail shell in case an attack on a non-root account succeeds
• changing SSH service to use a non-obvious port (such as port 5022 instead of 22)
• deactivate password authentication and rely exclusively on authentication keys
• restrict allowed IP address by country of origin
• only allow certain IP addresses or ranges of addresses to have access

We chose to implement more than one of these solutions, and I wanted to share some techniques we used for our implementation.

For the last rule that only allows certain IP addresses, I wanted to start with a list of valid IP addresses used in the last month.  The following script extracts these IP numbers from our SSH log file, sorts them alphabetically, and then removes duplicates.  Note that this server uses Fedora – you may need to tweak it for other linux distributions.

root# fgrep "Accepted" /var/log/secure* | awk '{print $11}' | sort | uniq
166.77.6.4
205.232.34.1
67.255.5.155
...

The IP addresses from the above script should be added to the file /etc/hosts.allow in the following format:

# hosts.allow   This file contains access rules which are used to
#               allow or deny connections to network services that
#               either use the tcp_wrappers library or that have been
#               started through a tcp_wrappers-enabled xinetd.
#
#               See 'man 5 hosts_options' and 'man 5 hosts_access'
#               for information on rule syntax.
#               See 'man tcpd' for information on tcp_wrappers

# allow local addresses
all: 127.0.0.1
all: 192.168.1.*

# valid IP addresses gathered June 2010
all: 166.77.6.4
all: 205.232.34.1
all: 67.255.5.155
...

Now disallow all other IP addresses for SSH by editing the file /etc/hosts.deny:

# hosts.deny    This file contains access rules which are used to
#               deny connections to network services that either use
#               the tcp_wrappers library or that have been
#               started through a tcp_wrappers-enabled xinetd.
#
#               The rules in this file can also be set up in
#               /etc/hosts.allow with a 'deny' option instead.
#
#               See 'man 5 hosts_options' and 'man 5 hosts_access'
#               for information on rule syntax.
#               See 'man tcpd' for information on tcp_wrappers
#
# The portmap line is redundant, but it is left to remind you that
# the new secure portmap uses hosts.deny and hosts.allow.  In particular
# you should know that NFS uses portmap!

# deny SSH service except for IP numbers in /etc/hosts.allow file
sshd: all

Restart your SSH service, and your server should now be a bit more secure against distributed dictionary attacks:

root# service sshd restart

An Internet search using keywords from the other mentioned solutions above will teach you how to change SSH port, disallow password authentication, etc.

Frameworks generally belong to one of two camps, “pure frameworks” and “full stack frameworks”.

Most frameworks don’t make a distinction between “framework” and “application” features, and as a result they become cluttered. They also become more daunting to newcomers, with their sheer volume of classes that attempt to solve every practical problem a web developer might ever encounter – user management, document and image handling, upload and downloading, etc.

My favorite PHP framework, Yii, is a pure framework – free from the burden that many frameworks drag around, namely a full “application stack”. Yii consists almost exclusively of components and features designed to support certain practices or paradigms.

One of my favorite things about Yii, as compared to “full stack” frameworks, such as Zend, is that Yii only comes with components and architecture that can be rightfully said to belong in the “framework” domain, and not the “application” domain.

Of course, it may be hard in some cases to draw an exact line between the two, and I think the first criteria for selecting features that belong in the framework, and not in an application stack, should be:

• Is it absolutely general-purpose?
• If not, is it fully extensible?

Certain features, like the URL manager or Active Record, are not absolutely general purpose, in the sense that they may not satisfy every possible need anyone could ever have. But they are sufficiently general purpose in the sense that almost everybody is going to need at least the core functionality of those components. And because they are fully extensible, developers can build on top of them, rather than having to replace them, if they find that a component does not fully cater to their specific needs.

The reason why I do not want application features in a framework, is because I know from experience that these will not meet my strict requirements. I will eventually end up replacing many of these features and the existing features provided by the framework are reduced to baggage.

Everybody is different – we all have different goals, and even if we share some goals, we usually have different means for reaching those goals.

I believe the reason why we web developers converge around a framework, is because we agree on certain practices. The framework is designed to leverage those practices in a convenient and streamlined way.

The subtle art of the framework is to achieve accepted practice, without getting in the way of individuality – to enable us to adhere to the good practices that we agree upon, while still allowing us to be as different, as versatile and as colorful as we can!

In short, frameworks enable us to focus our efforts on “business logic”, the functionality that has real value to the website users. Thanks to the framework, we can focus on the practical requirements, without sacrificing the values that professional software developers care about: clean, maintainable and extensible code.

With a good framework, we can deliver value and high code quality, without charging extra for quality!

However there is a glaring limitation:  there are base classes for horizontal scrollviews and vertical scrollviews, but not one where one can scroll in two dimensions at the same time.

It would be a challenge to write a two-dimensional scrollview class from scratch.  Luckily for us, the entire Android platform is open-source, so we had access to both the vertical and horizontal scrollview source code.  After a few hours of work we had created a new TwoDScrollView class.

First, here is a disclaimer: this class has been “munged” together and is not bullet-proof for a true generalized solution.  For example the methods that handle sub-view focusing have not been tested, and there are some nuances about how to prioritize focused sub-views in two-dimensions.  I have a fully-stripped version of this class without sub-view focusing or key events that I’d be happy to share upon request.

Without further ado, here is the new TwoDScrollView class:

/*
 * Copyright (C) 2006 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.
 */
/*
 * Revised 5/19/2010 by GORGES
 * Now supports two-dimensional view scrolling
 * http://GORGES.us
 */

package us.gorges.my_package;

import java.util.List;

import android.content.Context;
import android.graphics.Rect;
import android.util.AttributeSet;
import android.view.FocusFinder;
import android.view.KeyEvent;
import android.view.MotionEvent;
import android.view.VelocityTracker;
import android.view.View;
import android.view.ViewConfiguration;
import android.view.ViewGroup;
import android.view.ViewParent;
import android.view.animation.AnimationUtils;
import android.widget.FrameLayout;
import android.widget.LinearLayout;
import android.widget.Scroller;
import android.widget.TextView;

/**
 * Layout container for a view hierarchy that can be scrolled by the user,
 * allowing it to be larger than the physical display.  A TwoDScrollView
 * is a {@link FrameLayout}, meaning you should place one child in it
 * containing the entire contents to scroll; this child may itself be a layout
 * manager with a complex hierarchy of objects.  A child that is often used
 * is a {@link LinearLayout} in a vertical orientation, presenting a vertical
 * array of top-level items that the user can scroll through.
 *
 * <p>The {@link TextView} class also
 * takes care of its own scrolling, so does not require a TwoDScrollView, but
 * using the two together is possible to achieve the effect of a text view
 * within a larger container.
 */
public class TwoDScrollView extends FrameLayout {
 static final int ANIMATED_SCROLL_GAP = 250;
 static final float MAX_SCROLL_FACTOR = 0.5f;

 private long mLastScroll;

 private final Rect mTempRect = new Rect();
 private Scroller mScroller;

 /**
 * Flag to indicate that we are moving focus ourselves. This is so the
 * code that watches for focus changes initiated outside this TwoDScrollView
 * knows that it does not have to do anything.
 */
 private boolean mTwoDScrollViewMovedFocus;

 /**
 * Position of the last motion event.
 */
 private float mLastMotionY;
 private float mLastMotionX;

 /**
 * True when the layout has changed but the traversal has not come through yet.
 * Ideally the view hierarchy would keep track of this for us.
 */
 private boolean mIsLayoutDirty = true;

 /**
 * The child to give focus to in the event that a child has requested focus while the
 * layout is dirty. This prevents the scroll from being wrong if the child has not been
 * laid out before requesting focus.
 */
 private View mChildToScrollTo = null;

 /**
 * True if the user is currently dragging this TwoDScrollView around. This is
 * not the same as 'is being flinged', which can be checked by
 * mScroller.isFinished() (flinging begins when the user lifts his finger).
 */
 private boolean mIsBeingDragged = false;

 /**
 * Determines speed during touch scrolling
 */
 private VelocityTracker mVelocityTracker;

 /**
 * Whether arrow scrolling is animated.
 */
 private int mTouchSlop;
 private int mMinimumVelocity;
 private int mMaximumVelocity;

 public TwoDScrollView(Context context) {
   super(context);
   initTwoDScrollView();
 }

 public TwoDScrollView(Context context, AttributeSet attrs) {
   super(context, attrs);
   initTwoDScrollView();
 }

 public TwoDScrollView(Context context, AttributeSet attrs, int defStyle) {
   super(context, attrs, defStyle);
   initTwoDScrollView();
 }

 @Override
 protected float getTopFadingEdgeStrength() {
   if (getChildCount() == 0) {
     return 0.0f;
   }
   final int length = getVerticalFadingEdgeLength();
   if (getScrollY() < length) {
     return getScrollY() / (float) length;
   }
   return 1.0f;
 }

 @Override
 protected float getBottomFadingEdgeStrength() {
   if (getChildCount() == 0) {
     return 0.0f;
   }
   final int length = getVerticalFadingEdgeLength();
   final int bottomEdge = getHeight() - getPaddingBottom();
   final int span = getChildAt(0).getBottom() - getScrollY() - bottomEdge;
   if (span < length) {
     return span / (float) length;
   }
   return 1.0f;
 }

 @Override
 protected float getLeftFadingEdgeStrength() {
   if (getChildCount() == 0) {
     return 0.0f;
   }
   final int length = getHorizontalFadingEdgeLength();
   if (getScrollX() < length) {
     return getScrollX() / (float) length;
   }
   return 1.0f;
 }

 @Override
 protected float getRightFadingEdgeStrength() {
   if (getChildCount() == 0) {
     return 0.0f;
   }
   final int length = getHorizontalFadingEdgeLength();
   final int rightEdge = getWidth() - getPaddingRight();
   final int span = getChildAt(0).getRight() - getScrollX() - rightEdge;
   if (span < length) {
     return span / (float) length;
   }
   return 1.0f;
 }

 /**
 * @return The maximum amount this scroll view will scroll in response to
 *   an arrow event.
 */
 public int getMaxScrollAmountVertical() {
   return (int) (MAX_SCROLL_FACTOR * getHeight());
 }
 public int getMaxScrollAmountHorizontal() {
   return (int) (MAX_SCROLL_FACTOR * getWidth());
 }

 private void initTwoDScrollView() {
   mScroller = new Scroller(getContext());
   setFocusable(true);
   setDescendantFocusability(FOCUS_AFTER_DESCENDANTS);
   setWillNotDraw(false);
   final ViewConfiguration configuration = ViewConfiguration.get(getContext());
   mTouchSlop = configuration.getScaledTouchSlop();
   mMinimumVelocity = configuration.getScaledMinimumFlingVelocity();
   mMaximumVelocity = configuration.getScaledMaximumFlingVelocity();
 }

 @Override
 public void addView(View child) {
   if (getChildCount() > 0) {
     throw new IllegalStateException("TwoDScrollView can host only one direct child");
   }
   super.addView(child);
 }

 @Override
 public void addView(View child, int index) {
   if (getChildCount() > 0) {
     throw new IllegalStateException("TwoDScrollView can host only one direct child");
   }
   super.addView(child, index);
 }

 @Override
 public void addView(View child, ViewGroup.LayoutParams params) {
   if (getChildCount() > 0) {
     throw new IllegalStateException("TwoDScrollView can host only one direct child");
   }
   super.addView(child, params);
 }

 @Override
 public void addView(View child, int index, ViewGroup.LayoutParams params) {
   if (getChildCount() > 0) {
     throw new IllegalStateException("TwoDScrollView can host only one direct child");
   }
   super.addView(child, index, params);
 }

 /**
 * @return Returns true this TwoDScrollView can be scrolled
 */
 private boolean canScroll() {
   View child = getChildAt(0);
   if (child != null) {
     int childHeight = child.getHeight();
     int childWidth = child.getWidth();
     return (getHeight() < childHeight + getPaddingTop() + getPaddingBottom()) ||
            (getWidth() < childWidth + getPaddingLeft() + getPaddingRight());
   }
   return false;
 }

 @Override
 public boolean dispatchKeyEvent(KeyEvent event) {
   // Let the focused view and/or our descendants get the key first
   boolean handled = super.dispatchKeyEvent(event);
   if (handled) {
     return true;
   }
   return executeKeyEvent(event);
 }

 /**
 * You can call this function yourself to have the scroll view perform
 * scrolling from a key event, just as if the event had been dispatched to
 * it by the view hierarchy.
 *
 * @param event The key event to execute.
 * @return Return true if the event was handled, else false.
 */
 public boolean executeKeyEvent(KeyEvent event) {
   mTempRect.setEmpty();
   if (!canScroll()) {
     if (isFocused()) {
       View currentFocused = findFocus();
       if (currentFocused == this) currentFocused = null;
       View nextFocused = FocusFinder.getInstance().findNextFocus(this, currentFocused, View.FOCUS_DOWN);
       return nextFocused != null && nextFocused != this && nextFocused.requestFocus(View.FOCUS_DOWN);
     }
     return false;
   }
   boolean handled = false;
   if (event.getAction() == KeyEvent.ACTION_DOWN) {
     switch (event.getKeyCode()) {
       case KeyEvent.KEYCODE_DPAD_UP:
         if (!event.isAltPressed()) {
           handled = arrowScroll(View.FOCUS_UP, false);
         } else {
           handled = fullScroll(View.FOCUS_UP, false);
         }
         break;
       case KeyEvent.KEYCODE_DPAD_DOWN:
         if (!event.isAltPressed()) {
           handled = arrowScroll(View.FOCUS_DOWN, false);
         } else {
           handled = fullScroll(View.FOCUS_DOWN, false);
         }
         break;
       case KeyEvent.KEYCODE_DPAD_LEFT:
         if (!event.isAltPressed()) {
           handled = arrowScroll(View.FOCUS_UP, true);
         } else {
           handled = fullScroll(View.FOCUS_UP, true);
         }
         break;
       case KeyEvent.KEYCODE_DPAD_RIGHT:
         if (!event.isAltPressed()) {
           handled = arrowScroll(View.FOCUS_DOWN, true);
         } else {
           handled = fullScroll(View.FOCUS_DOWN, true);
         }
         break;
     }
   }
   return handled;
 }

 @Override
 public boolean onInterceptTouchEvent(MotionEvent ev) {
   /*
   * This method JUST determines whether we want to intercept the motion.
   * If we return true, onMotionEvent will be called and we do the actual
   * scrolling there.
   *
   * Shortcut the most recurring case: the user is in the dragging
   * state and he is moving his finger.  We want to intercept this
   * motion.
   */
   final int action = ev.getAction();
   if ((action == MotionEvent.ACTION_MOVE) && (mIsBeingDragged)) {
     return true;
   }
   if (!canScroll()) {
     mIsBeingDragged = false;
     return false;
   }
   final float y = ev.getY();
   final float x = ev.getX();
   switch (action) {
     case MotionEvent.ACTION_MOVE:
       /*
       * mIsBeingDragged == false, otherwise the shortcut would have caught it. Check
       * whether the user has moved far enough from his original down touch.
       */
       /*
       * Locally do absolute value. mLastMotionY is set to the y value
       * of the down event.
       */
       final int yDiff = (int) Math.abs(y - mLastMotionY);
       final int xDiff = (int) Math.abs(x - mLastMotionX);
       if (yDiff > mTouchSlop || xDiff > mTouchSlop) {
         mIsBeingDragged = true;
       }
       break;

     case MotionEvent.ACTION_DOWN:
       /* Remember location of down touch */
       mLastMotionY = y;
       mLastMotionX = x;

       /*
       * If being flinged and user touches the screen, initiate drag;
       * otherwise don't.  mScroller.isFinished should be false when
       * being flinged.
       */
       mIsBeingDragged = !mScroller.isFinished();
       break;

     case MotionEvent.ACTION_CANCEL:
     case MotionEvent.ACTION_UP:
       /* Release the drag */
       mIsBeingDragged = false;
       break;
   }

   /*
   * The only time we want to intercept motion events is if we are in the
   * drag mode.
   */
   return mIsBeingDragged;
 }

 @Override
 public boolean onTouchEvent(MotionEvent ev) {

   if (ev.getAction() == MotionEvent.ACTION_DOWN && ev.getEdgeFlags() != 0) {
     // Don't handle edge touches immediately -- they may actually belong to one of our
     // descendants.
     return false;
   }

   if (!canScroll()) {
     return false;
   }

   if (mVelocityTracker == null) {
     mVelocityTracker = VelocityTracker.obtain();
   }
   mVelocityTracker.addMovement(ev);

   final int action = ev.getAction();
   final float y = ev.getY();
   final float x = ev.getX();

   switch (action) {
     case MotionEvent.ACTION_DOWN:
       /*
       * If being flinged and user touches, stop the fling. isFinished
       * will be false if being flinged.
       */
       if (!mScroller.isFinished()) {
         mScroller.abortAnimation();
       }

       // Remember where the motion event started
       mLastMotionY = y;
       mLastMotionX = x;
       break;
     case MotionEvent.ACTION_MOVE:
       // Scroll to follow the motion event
       int deltaX = (int) (mLastMotionX - x);
       int deltaY = (int) (mLastMotionY - y);
       mLastMotionX = x;
       mLastMotionY = y;

       if (deltaX < 0) {
         if (getScrollX() < 0) {
           deltaX = 0;
         }
       } else if (deltaX > 0) {
         final int rightEdge = getWidth() - getPaddingRight();
         final int availableToScroll = getChildAt(0).getRight() - getScrollX() - rightEdge;
         if (availableToScroll > 0) {
           deltaX = Math.min(availableToScroll, deltaX);
         } else {
           deltaX = 0;
         }
       }
       if (deltaY < 0) {
         if (getScrollY() < 0) {
           deltaY = 0;
         }
       } else if (deltaY > 0) {
         final int bottomEdge = getHeight() - getPaddingBottom();
         final int availableToScroll = getChildAt(0).getBottom() - getScrollY() - bottomEdge;
         if (availableToScroll > 0) {
           deltaY = Math.min(availableToScroll, deltaY);
         } else {
           deltaY = 0;
         }
       }
       if (deltaY != 0 || deltaX != 0)
         scrollBy(deltaX, deltaY);
       break;
     case MotionEvent.ACTION_UP:
         final VelocityTracker velocityTracker = mVelocityTracker;
         velocityTracker.computeCurrentVelocity(1000, mMaximumVelocity);
         int initialXVelocity = (int) velocityTracker.getXVelocity();
         int initialYVelocity = (int) velocityTracker.getYVelocity();
         if ((Math.abs(initialXVelocity) + Math.abs(initialYVelocity) > mMinimumVelocity) && getChildCount() > 0) {
           fling(-initialXVelocity, -initialYVelocity);
         }
         if (mVelocityTracker != null) {
           mVelocityTracker.recycle();
           mVelocityTracker = null;
         }
   }
   return true;
 }

 /**
  * Finds the next focusable component that fits in this View's bounds
  * (excluding fading edges) pretending that this View's top is located at
  * the parameter top.
  *
  * @param topFocus           look for a candidate is the one at the top of the bounds
  *                           if topFocus is true, or at the bottom of the bounds if topFocus is
  *                           false
  * @param top                the top offset of the bounds in which a focusable must be
  *                           found (the fading edge is assumed to start at this position)
  * @param preferredFocusable the View that has highest priority and will be
  *                           returned if it is within my bounds (null is valid)
  * @return the next focusable component in the bounds or null if none can be
  *         found
  */
 private View findFocusableViewInMyBounds(final boolean topFocus, final int top, final boolean leftFocus, final int left, View preferredFocusable) {
   /*
   * The fading edge's transparent side should be considered for focus
   * since it's mostly visible, so we divide the actual fading edge length
   * by 2.
   */
   final int verticalFadingEdgeLength = getVerticalFadingEdgeLength() / 2;
   final int topWithoutFadingEdge = top + verticalFadingEdgeLength;
   final int bottomWithoutFadingEdge = top + getHeight() - verticalFadingEdgeLength;
   final int horizontalFadingEdgeLength = getHorizontalFadingEdgeLength() / 2;
   final int leftWithoutFadingEdge = left + horizontalFadingEdgeLength;
   final int rightWithoutFadingEdge = left + getWidth() - horizontalFadingEdgeLength;

   if ((preferredFocusable != null)
     && (preferredFocusable.getTop() < bottomWithoutFadingEdge)
     && (preferredFocusable.getBottom() > topWithoutFadingEdge)
     && (preferredFocusable.getLeft() < rightWithoutFadingEdge)
     && (preferredFocusable.getRight() > leftWithoutFadingEdge)) {
     return preferredFocusable;
   }
   return findFocusableViewInBounds(topFocus, topWithoutFadingEdge, bottomWithoutFadingEdge, leftFocus, leftWithoutFadingEdge, rightWithoutFadingEdge);
 }

 /**
 * Finds the next focusable component that fits in the specified bounds.
 * </p>
 *
 * @param topFocus look for a candidate is the one at the top of the bounds
 *                 if topFocus is true, or at the bottom of the bounds if topFocus is
 *                 false
 * @param top      the top offset of the bounds in which a focusable must be
 *                 found
 * @param bottom   the bottom offset of the bounds in which a focusable must
 *                 be found
 * @return the next focusable component in the bounds or null if none can
 *         be found
 */
 private View findFocusableViewInBounds(boolean topFocus, int top, int bottom, boolean leftFocus, int left, int right) {
   List<View> focusables = getFocusables(View.FOCUS_FORWARD);
   View focusCandidate = null;

   /*
   * A fully contained focusable is one where its top is below the bound's
   * top, and its bottom is above the bound's bottom. A partially
   * contained focusable is one where some part of it is within the
   * bounds, but it also has some part that is not within bounds.  A fully contained
   * focusable is preferred to a partially contained focusable.
   */
   boolean foundFullyContainedFocusable = false;

   int count = focusables.size();
   for (int i = 0; i < count; i++) {
     View view = focusables.get(i);
     int viewTop = view.getTop();
     int viewBottom = view.getBottom();
     int viewLeft = view.getLeft();
     int viewRight = view.getRight();

     if (top < viewBottom && viewTop < bottom && left < viewRight && viewLeft < right) {
       /*
       * the focusable is in the target area, it is a candidate for
       * focusing
       */
       final boolean viewIsFullyContained = (top < viewTop) && (viewBottom < bottom) && (left < viewLeft) && (viewRight < right);
       if (focusCandidate == null) {
         /* No candidate, take this one */
         focusCandidate = view;
         foundFullyContainedFocusable = viewIsFullyContained;
       } else {
         final boolean viewIsCloserToVerticalBoundary =
           (topFocus && viewTop < focusCandidate.getTop()) ||
           (!topFocus && viewBottom > focusCandidate.getBottom());
         final boolean viewIsCloserToHorizontalBoundary =
           (leftFocus && viewLeft < focusCandidate.getLeft()) ||
           (!leftFocus && viewRight > focusCandidate.getRight());
         if (foundFullyContainedFocusable) {
           if (viewIsFullyContained && viewIsCloserToVerticalBoundary && viewIsCloserToHorizontalBoundary) {
             /*
              * We're dealing with only fully contained views, so
              * it has to be closer to the boundary to beat our
              * candidate
              */
             focusCandidate = view;
           }
         } else {
           if (viewIsFullyContained) {
             /* Any fully contained view beats a partially contained view */
             focusCandidate = view;
             foundFullyContainedFocusable = true;
           } else if (viewIsCloserToVerticalBoundary && viewIsCloserToHorizontalBoundary) {
             /*
              * Partially contained view beats another partially
              * contained view if it's closer
              */
             focusCandidate = view;
           }
         }
       }
     }
   }
   return focusCandidate;
 }

 /**
  * <p>Handles scrolling in response to a "home/end" shortcut press. This
  * method will scroll the view to the top or bottom and give the focus
  * to the topmost/bottommost component in the new visible area. If no
  * component is a good candidate for focus, this scrollview reclaims the
  * focus.</p>
  *
  * @param direction the scroll direction: {@link android.view.View#FOCUS_UP}
  *                  to go the top of the view or
  *                  {@link android.view.View#FOCUS_DOWN} to go the bottom
  * @return true if the key event is consumed by this method, false otherwise
  */
 public boolean fullScroll(int direction, boolean horizontal) {
   if (!horizontal) {
     boolean down = direction == View.FOCUS_DOWN;
     int height = getHeight();
     mTempRect.top = 0;
     mTempRect.bottom = height;
     if (down) {
       int count = getChildCount();
       if (count > 0) {
         View view = getChildAt(count - 1);
         mTempRect.bottom = view.getBottom();
         mTempRect.top = mTempRect.bottom - height;
       }
     }
     return scrollAndFocus(direction, mTempRect.top, mTempRect.bottom, 0, 0, 0);
   } else {
     boolean right = direction == View.FOCUS_DOWN;
     int width = getWidth();
     mTempRect.left = 0;
     mTempRect.right = width;
     if (right) {
       int count = getChildCount();
       if (count > 0) {
         View view = getChildAt(count - 1);
         mTempRect.right = view.getBottom();
         mTempRect.left = mTempRect.right - width;
       }
     }
     return scrollAndFocus(0, 0, 0, direction, mTempRect.top, mTempRect.bottom);
   }
 }

 /**
  * <p>Scrolls the view to make the area defined by <code>top</code> and
  * <code>bottom</code> visible. This method attempts to give the focus
  * to a component visible in this area. If no component can be focused in
  * the new visible area, the focus is reclaimed by this scrollview.</p>
  *
  * @param direction the scroll direction: {@link android.view.View#FOCUS_UP}
  *                  to go upward
  *                  {@link android.view.View#FOCUS_DOWN} to downward
  * @param top       the top offset of the new area to be made visible
  * @param bottom    the bottom offset of the new area to be made visible
  * @return true if the key event is consumed by this method, false otherwise
  */
 private boolean scrollAndFocus(int directionY, int top, int bottom, int directionX, int left, int right) {
   boolean handled = true;
   int height = getHeight();
   int containerTop = getScrollY();
   int containerBottom = containerTop + height;
   boolean up = directionY == View.FOCUS_UP;
   int width = getWidth();
   int containerLeft = getScrollX();
   int containerRight = containerLeft + width;
   boolean leftwards = directionX == View.FOCUS_UP;
   View newFocused = findFocusableViewInBounds(up, top, bottom, leftwards, left, right);
   if (newFocused == null) {
     newFocused = this;
   }
   if ((top >= containerTop && bottom <= containerBottom) || (left >= containerLeft && right <= containerRight)) {
     handled = false;
   } else {
     int deltaY = up ? (top - containerTop) : (bottom - containerBottom);
     int deltaX = leftwards ? (left - containerLeft) : (right - containerRight);
     doScroll(deltaX, deltaY);
   }
   if (newFocused != findFocus() && newFocused.requestFocus(directionY)) {
     mTwoDScrollViewMovedFocus = true;
     mTwoDScrollViewMovedFocus = false;
   }
   return handled;
 }

 /**
  * Handle scrolling in response to an up or down arrow click.
  *
  * @param direction The direction corresponding to the arrow key that was
  *                  pressed
  * @return True if we consumed the event, false otherwise
  */
 public boolean arrowScroll(int direction, boolean horizontal) {
   View currentFocused = findFocus();
   if (currentFocused == this) currentFocused = null;
   View nextFocused = FocusFinder.getInstance().findNextFocus(this, currentFocused, direction);
   final int maxJump = horizontal ? getMaxScrollAmountHorizontal() : getMaxScrollAmountVertical();

   if (!horizontal) {
     if (nextFocused != null) {
       nextFocused.getDrawingRect(mTempRect);
       offsetDescendantRectToMyCoords(nextFocused, mTempRect);
       int scrollDelta = computeScrollDeltaToGetChildRectOnScreen(mTempRect);
       doScroll(0, scrollDelta);
       nextFocused.requestFocus(direction);
     } else {
       // no new focus
       int scrollDelta = maxJump;
       if (direction == View.FOCUS_UP && getScrollY() < scrollDelta) {
         scrollDelta = getScrollY();
       } else if (direction == View.FOCUS_DOWN) {
         if (getChildCount() > 0) {
           int daBottom = getChildAt(0).getBottom();
           int screenBottom = getScrollY() + getHeight();
           if (daBottom - screenBottom < maxJump) {
             scrollDelta = daBottom - screenBottom;
           }
         }
       }
       if (scrollDelta == 0) {
         return false;
       }
       doScroll(0, direction == View.FOCUS_DOWN ? scrollDelta : -scrollDelta);
     }
   } else {
     if (nextFocused != null) {
       nextFocused.getDrawingRect(mTempRect);
       offsetDescendantRectToMyCoords(nextFocused, mTempRect);
       int scrollDelta = computeScrollDeltaToGetChildRectOnScreen(mTempRect);
       doScroll(scrollDelta, 0);
       nextFocused.requestFocus(direction);
     } else {
       // no new focus
       int scrollDelta = maxJump;
       if (direction == View.FOCUS_UP && getScrollY() < scrollDelta) {
         scrollDelta = getScrollY();
       } else if (direction == View.FOCUS_DOWN) {
         if (getChildCount() > 0) {
           int daBottom = getChildAt(0).getBottom();
           int screenBottom = getScrollY() + getHeight();
           if (daBottom - screenBottom < maxJump) {
             scrollDelta = daBottom - screenBottom;
           }
         }
       }
       if (scrollDelta == 0) {
         return false;
       }
       doScroll(direction == View.FOCUS_DOWN ? scrollDelta : -scrollDelta, 0);
     }
   }
   return true;
 }

 /**
  * Smooth scroll by a Y delta
  *
  * @param delta the number of pixels to scroll by on the Y axis
  */
 private void doScroll(int deltaX, int deltaY) {
   if (deltaX != 0 || deltaY != 0) {
     smoothScrollBy(deltaX, deltaY);
   }
 }

 /**
  * Like {@link View#scrollBy}, but scroll smoothly instead of immediately.
  *
  * @param dx the number of pixels to scroll by on the X axis
  * @param dy the number of pixels to scroll by on the Y axis
  */
 public final void smoothScrollBy(int dx, int dy) {
   long duration = AnimationUtils.currentAnimationTimeMillis() - mLastScroll;
   if (duration > ANIMATED_SCROLL_GAP) {
     mScroller.startScroll(getScrollX(), getScrollY(), dx, dy);
     awakenScrollBars(mScroller.getDuration());
     invalidate();
   } else {
     if (!mScroller.isFinished()) {
       mScroller.abortAnimation();
     }
     scrollBy(dx, dy);
   }
   mLastScroll = AnimationUtils.currentAnimationTimeMillis();
 }

 /**
  * Like {@link #scrollTo}, but scroll smoothly instead of immediately.
  *
  * @param x the position where to scroll on the X axis
  * @param y the position where to scroll on the Y axis
  */
 public final void smoothScrollTo(int x, int y) {
   smoothScrollBy(x - getScrollX(), y - getScrollY());
 }

 /**
  * <p>The scroll range of a scroll view is the overall height of all of its
  * children.</p>
  */
 @Override
 protected int computeVerticalScrollRange() {
   int count = getChildCount();
   return count == 0 ? getHeight() : (getChildAt(0)).getBottom();
 }
 @Override
 protected int computeHorizontalScrollRange() {
   int count = getChildCount();
   return count == 0 ? getWidth() : (getChildAt(0)).getRight();
 }

 @Override
 protected void measureChild(View child, int parentWidthMeasureSpec, int parentHeightMeasureSpec) {
   ViewGroup.LayoutParams lp = child.getLayoutParams();
   int childWidthMeasureSpec;
   int childHeightMeasureSpec;

   childWidthMeasureSpec = getChildMeasureSpec(parentWidthMeasureSpec, getPaddingLeft() + getPaddingRight(), lp.width);
   childHeightMeasureSpec = MeasureSpec.makeMeasureSpec(0, MeasureSpec.UNSPECIFIED);

   child.measure(childWidthMeasureSpec, childHeightMeasureSpec);
 }

 @Override
 protected void measureChildWithMargins(View child, int parentWidthMeasureSpec, int widthUsed, int parentHeightMeasureSpec, int heightUsed) {
   final MarginLayoutParams lp = (MarginLayoutParams) child.getLayoutParams();
   final int childWidthMeasureSpec = getChildMeasureSpec(parentWidthMeasureSpec,
   getPaddingLeft() + getPaddingRight() + lp.leftMargin + lp.rightMargin + widthUsed, lp.width);
   final int childHeightMeasureSpec = MeasureSpec.makeMeasureSpec(lp.topMargin + lp.bottomMargin, MeasureSpec.UNSPECIFIED);

   child.measure(childWidthMeasureSpec, childHeightMeasureSpec);
 }

 @Override
 public void computeScroll() {
   if (mScroller.computeScrollOffset()) {
     // This is called at drawing time by ViewGroup.  We don't want to
     // re-show the scrollbars at this point, which scrollTo will do,
     // so we replicate most of scrollTo here.
     //
     //         It's a little odd to call onScrollChanged from inside the drawing.
     //
     //         It is, except when you remember that computeScroll() is used to
     //         animate scrolling. So unless we want to defer the onScrollChanged()
     //         until the end of the animated scrolling, we don't really have a
     //         choice here.
     //
     //         I agree.  The alternative, which I think would be worse, is to post
     //         something and tell the subclasses later.  This is bad because there
     //         will be a window where mScrollX/Y is different from what the app
     //         thinks it is.
     //
     int oldX = getScrollX();
     int oldY = getScrollY();
     int x = mScroller.getCurrX();
     int y = mScroller.getCurrY();
     if (getChildCount() > 0) {
       View child = getChildAt(0);
       scrollTo(clamp(x, getWidth() - getPaddingRight() - getPaddingLeft(), child.getWidth()),
       clamp(y, getHeight() - getPaddingBottom() - getPaddingTop(), child.getHeight()));
     } else {
       scrollTo(x, y);
     }
     if (oldX != getScrollX() || oldY != getScrollY()) {
       onScrollChanged(getScrollX(), getScrollY(), oldX, oldY);
     }

     // Keep on drawing until the animation has finished.
     postInvalidate();
   }
 }

 /**
  * Scrolls the view to the given child.
  *
  * @param child the View to scroll to
  */
 private void scrollToChild(View child) {
   child.getDrawingRect(mTempRect);
   /* Offset from child's local coordinates to TwoDScrollView coordinates */
   offsetDescendantRectToMyCoords(child, mTempRect);
   int scrollDelta = computeScrollDeltaToGetChildRectOnScreen(mTempRect);
   if (scrollDelta != 0) {
     scrollBy(0, scrollDelta);
   }
 }

 /**
  * If rect is off screen, scroll just enough to get it (or at least the
  * first screen size chunk of it) on screen.
  *
  * @param rect      The rectangle.
  * @param immediate True to scroll immediately without animation
  * @return true if scrolling was performed
  */
 private boolean scrollToChildRect(Rect rect, boolean immediate) {
   final int delta = computeScrollDeltaToGetChildRectOnScreen(rect);
   final boolean scroll = delta != 0;
   if (scroll) {
     if (immediate) {
       scrollBy(0, delta);
     } else {
       smoothScrollBy(0, delta);
     }
   }
   return scroll;
 }

 /**
  * Compute the amount to scroll in the Y direction in order to get
  * a rectangle completely on the screen (or, if taller than the screen,
  * at least the first screen size chunk of it).
  *
  * @param rect The rect.
  * @return The scroll delta.
  */
 protected int computeScrollDeltaToGetChildRectOnScreen(Rect rect) {
   if (getChildCount() == 0) return 0;
   int height = getHeight();
   int screenTop = getScrollY();
   int screenBottom = screenTop + height;
   int fadingEdge = getVerticalFadingEdgeLength();
   // leave room for top fading edge as long as rect isn't at very top
   if (rect.top > 0) {
     screenTop += fadingEdge;
   }

   // leave room for bottom fading edge as long as rect isn't at very bottom
   if (rect.bottom < getChildAt(0).getHeight()) {
     screenBottom -= fadingEdge;
   }
   int scrollYDelta = 0;
   if (rect.bottom > screenBottom && rect.top > screenTop) {
     // need to move down to get it in view: move down just enough so
     // that the entire rectangle is in view (or at least the first
     // screen size chunk).
     if (rect.height() > height) {
       // just enough to get screen size chunk on
       scrollYDelta += (rect.top - screenTop);
     } else {
       // get entire rect at bottom of screen
       scrollYDelta += (rect.bottom - screenBottom);
     }

     // make sure we aren't scrolling beyond the end of our content
     int bottom = getChildAt(0).getBottom();
     int distanceToBottom = bottom - screenBottom;
     scrollYDelta = Math.min(scrollYDelta, distanceToBottom);

   } else if (rect.top < screenTop && rect.bottom < screenBottom) {
     // need to move up to get it in view: move up just enough so that
     // entire rectangle is in view (or at least the first screen
     // size chunk of it).

     if (rect.height() > height) {
       // screen size chunk
       scrollYDelta -= (screenBottom - rect.bottom);
     } else {
       // entire rect at top
       scrollYDelta -= (screenTop - rect.top);
     }

     // make sure we aren't scrolling any further than the top our content
     scrollYDelta = Math.max(scrollYDelta, -getScrollY());
   }
   return scrollYDelta;
 }

 @Override
 public void requestChildFocus(View child, View focused) {
   if (!mTwoDScrollViewMovedFocus) {
     if (!mIsLayoutDirty) {
       scrollToChild(focused);
     } else {
       // The child may not be laid out yet, we can't compute the scroll yet
       mChildToScrollTo = focused;
     }
   }
   super.requestChildFocus(child, focused);
 }

 /**
  * When looking for focus in children of a scroll view, need to be a little
  * more careful not to give focus to something that is scrolled off screen.
  *
  * This is more expensive than the default {@link android.view.ViewGroup}
  * implementation, otherwise this behavior might have been made the default.
  */
 @Override
 protected boolean onRequestFocusInDescendants(int direction, Rect previouslyFocusedRect) {
   // convert from forward / backward notation to up / down / left / right
   // (ugh).
   if (direction == View.FOCUS_FORWARD) {
     direction = View.FOCUS_DOWN;
   } else if (direction == View.FOCUS_BACKWARD) {
     direction = View.FOCUS_UP;
   }

   final View nextFocus = previouslyFocusedRect == null ?
   FocusFinder.getInstance().findNextFocus(this, null, direction) :
   FocusFinder.getInstance().findNextFocusFromRect(this,
   previouslyFocusedRect, direction);

   if (nextFocus == null) {
     return false;
   }

   return nextFocus.requestFocus(direction, previouslyFocusedRect);
 }

 @Override
 public boolean requestChildRectangleOnScreen(View child, Rect rectangle, boolean immediate) {
   // offset into coordinate space of this scroll view
   rectangle.offset(child.getLeft() - child.getScrollX(), child.getTop() - child.getScrollY());
   return scrollToChildRect(rectangle, immediate);
 }

 @Override
 public void requestLayout() {
   mIsLayoutDirty = true;
   super.requestLayout();
 }

 @Override
 protected void onLayout(boolean changed, int l, int t, int r, int b) {
   super.onLayout(changed, l, t, r, b);
   mIsLayoutDirty = false;
   // Give a child focus if it needs it
   if (mChildToScrollTo != null && isViewDescendantOf(mChildToScrollTo, this)) {
     scrollToChild(mChildToScrollTo);
   }
   mChildToScrollTo = null;

   // Calling this with the present values causes it to re-clam them
   scrollTo(getScrollX(), getScrollY());
 }

 @Override
 protected void onSizeChanged(int w, int h, int oldw, int oldh) {
   super.onSizeChanged(w, h, oldw, oldh);

   View currentFocused = findFocus();
   if (null == currentFocused || this == currentFocused)
     return;

   // If the currently-focused view was visible on the screen when the
   // screen was at the old height, then scroll the screen to make that
   // view visible with the new screen height.
   currentFocused.getDrawingRect(mTempRect);
   offsetDescendantRectToMyCoords(currentFocused, mTempRect);
   int scrollDeltaX = computeScrollDeltaToGetChildRectOnScreen(mTempRect);
   int scrollDeltaY = computeScrollDeltaToGetChildRectOnScreen(mTempRect);
   doScroll(scrollDeltaX, scrollDeltaY);
 }

 /**
  * Return true if child is an descendant of parent, (or equal to the parent).
  */
 private boolean isViewDescendantOf(View child, View parent) {
   if (child == parent) {
     return true;
   }

   final ViewParent theParent = child.getParent();
   return (theParent instanceof ViewGroup) && isViewDescendantOf((View) theParent, parent);
 }

 /**
  * Fling the scroll view
  *
  * @param velocityY The initial velocity in the Y direction. Positive
  *                  numbers mean that the finger/curor is moving down the screen,
  *                  which means we want to scroll towards the top.
  */
 public void fling(int velocityX, int velocityY) {
   if (getChildCount() > 0) {
     int height = getHeight() - getPaddingBottom() - getPaddingTop();
     int bottom = getChildAt(0).getHeight();
     int width = getWidth() - getPaddingRight() - getPaddingLeft();
     int right = getChildAt(0).getWidth();

     mScroller.fling(getScrollX(), getScrollY(), velocityX, velocityY, 0, right - width, 0, bottom - height);

     final boolean movingDown = velocityY > 0;
     final boolean movingRight = velocityX > 0;

     View newFocused = findFocusableViewInMyBounds(movingRight, mScroller.getFinalX(), movingDown, mScroller.getFinalY(), findFocus());
     if (newFocused == null) {
       newFocused = this;
     }

     if (newFocused != findFocus() && newFocused.requestFocus(movingDown ? View.FOCUS_DOWN : View.FOCUS_UP)) {
       mTwoDScrollViewMovedFocus = true;
       mTwoDScrollViewMovedFocus = false;
     }

     awakenScrollBars(mScroller.getDuration());
     invalidate();
   }
 }

 /**
  * {@inheritDoc}
  *
  * <p>This version also clamps the scrolling to the bounds of our child.
  */
 public void scrollTo(int x, int y) {
   // we rely on the fact the View.scrollBy calls scrollTo.
   if (getChildCount() > 0) {
     View child = getChildAt(0);
     x = clamp(x, getWidth() - getPaddingRight() - getPaddingLeft(), child.getWidth());
     y = clamp(y, getHeight() - getPaddingBottom() - getPaddingTop(), child.getHeight());
     if (x != getScrollX() || y != getScrollY()) {
       super.scrollTo(x, y);
     }
   }
 }

 private int clamp(int n, int my, int child) {
   if (my >= child || n < 0) {
     /* my >= child is this case:
      *                    |--------------- me ---------------|
      *     |------ child ------|
      * or
      *     |--------------- me ---------------|
      *            |------ child ------|
      * or
      *     |--------------- me ---------------|
      *                                  |------ child ------|
      *
      * n < 0 is this case:
      *     |------ me ------|
      *                    |-------- child --------|
      *     |-- mScrollX --|
      */
     return 0;
   }
   if ((my+n) > child) {
     /* this case:
      *                    |------ me ------|
      *     |------ child ------|
      *     |-- mScrollX --|
      */
     return child-my;
   }
   return n;
 }
}
<pre>

In hindsight I think I know why two-dimensioning scrolling is not inherently included:  it is a memory hog.  There is an emphasis for beautiful graphics in the framework to achieve the best user experience; examples of this emphasis are true-color (24-bit) pixels, alpha transparency channel support, smooth scrolling, and display caching.  However the Android system restricts an application to 16MB of heap memory.  This limit is met quickly for a large cached two-dimensional scrolled view – note that 2,000 by 2,000 pixels times 4 bytes/pixel is 16MB.

Google has kind of answered that question too; you google it. But, what do you get in the data return. “Mary Kay Smith” returns 2,160,000 records, most of the top ones being from services offering to help you find your own Mary Kay. Maybe this is a good start when on the open Internet, but what if Mary Kay Smith works in the Kansas City branch of your company, even with some filtering words like, Kansas City and your company’s name, the return is still going to be front-loaded with helpful services.

This type of search is called a “Free Word” search. It has exception value. Most of us use it more than daily. But, it is not the only way to search and it is often not the appropriate way.

Assuming your company has an employee database you could use an in-house search tool, maybe still a Free Word search tool accessing just that one database. The return would show you all the Mary Kays in your company. Assuming only a few, you quickly click the one showing the KC address and there is her phone and fax number.

We could save a click. Why not have a drop-down list of all the employees? Let’s assume a few thousand. Yeah, the drop-down is too long. So maybe we have instead two “controls”. One is a radio button set asking “First Name” or “Last Name”. The other is an alphabet list. So now we get all the Matts, Marys, Marthas, and so forth. Right, what if we add one more control to limit us to Kansas City. Now we have a “Structured Search”.

You have probably thought of a still better or different way to find Mary Kay using a Structured Search, or maybe even, a combination of Structured Search and Free Search. This is Information Design – the search aspect of Information Design.

Let’s ask a polar question, what is the essential difference between Free Search and Structured Search? Answer is Capital Cost. It will always cost less, even much less, to setup a Free Search. But what about Operational Cost. If employees are searching each other out all day, how much is saved by providing a quick and certain search method? How much is saved by shipping the printer to the right Mary Kay?

Structured Searches cost more upfront because someone has to imagine them, organize a database, setup strict methods for adding, editing, and deleting from the database, and provide a set of controls for making the search. Not all of these costs are avoided by choosing Free Search, but many are.

It’s finally a User Interface decision involving cost, scalability, and accuracy. This is a favorite subject of mine. Call me for quick discussion or to arrange analysis of your search needs.

So why am I inviting you into these inner circles? An information system known as the Internet has pushed its way into your space. You have a website and you may or may not be prepared to deal with anything called Information Design. Here is some help.

Tufte alone has written seven books touching this subject, so I will not presume to encapsulate it here. My goal is to alert you to the existence and depth of this field, while also providing the rudiments for evaluating Information Design in your business application, database application, or website.

Here are two websites by GORGES that show just what Information Design deals with. The first site, http://birdsandbayous.org, is so simple that information design is little more than not making basic mistakes or omissions. There are only a few “information points” and only a basic navigation is needed to reach these points and return from them. This information design is analogous to the chapter ordering in a very small book.

The Martha Stewart site, http://www.marthastewart.com, is on a different planet. Every graphic and text object on this home page is a link to something, and amazingly, it all makes sense. It is visually organized to lead the visitor comfortably through dense information.

Note further, that much of the home page information is layered on top of other information and the visitor is invited to interact with it – to engage.

So it is dense, multi-layered, diverse, and interactive.

Your website’s complexity probably lies somewhere between the Birds and Bayous website and Martha Stewart’s. What can you learn from these two sites to help you evaluate your website or work with your website designers? Here’s one thing:

Insist that when presenting a design to you the designer makes at least three systematic passes across the demo pages in some carefully ordered, slow-paced way.
• In the first pass you should be told both “how” and “why” the information design suits your information set.
• In the second pass you should be told both “how” and “why” the graphic design suits you information set, as well as “how” and “why” it is appropriate to your brand and your communication intention for each of the audiences you have identified.
• In the third pass the designer needs to explain how the graphic design and information designs reinforce each other.

For the best result, allow the designers to make the points above without interruption and then go through it all again in a question and answer format. Good designers seek good criticism.

Here’s one more information design evaluation thing:

Ask two or three innocents to do the “What do you see?” test. With the page hidden, tell them you are going to ask them what they see first. Show the page. Then ask what they see second. You will get variation from person to person, but not much. Do this for the home page and for several subordinate pages. Does the visual prioritizing match your message plan?

My next posting will be also about Information Design – searching for information.

This set of guidelines may be used for evaluating websites as they develop or for examining sites by others. It addresses Usability, User Interface, and Communication. There is more to UX, please see other posting by me and watch for new posting on the subject in the next few months.

While this article is not explicitly about SEO, applying the standards here will only be good for SEO.

Usability

Do the graphics communication or supplement the message and purpose of the page? What is the desired balance between the graphical message and the text message?

People scan web pages. They do not read them. So, layout should be rhythmic and language should be concise. Break up traditional paragraphs by using subheadings, bullet lists, and short paragraphs.

Do not over do the breaking up of the page text. If it is a bullet item, then it does not need to be indented. Use one method or the other.

Font styles, sizes, and alignment should be consistent, and generally controlled by the style sheet. Avoid in-line styles. Using more than three different fonts in one page requires very high design skill. Keep it simple.

Link text should contain clear, action words that indicate what the link points to. Avoid using “Click Here” and other test that refers to the document. Remember, the user should be aware of the message, not the website.

Links to non-web formats, such as PDFs, should be identified by noting the format in the link text.

Users associate underlined text with links. So, do not underline text that is not links, this includes headings, page and section identification, and titles in general.

Color for emphasis and differentiation is under used in web pages. The twentieth century used italics and bold to differentiate. The twenty-first century uses color.

Communication

The purpose of every page should be clear to the user. Good questions to ask when creating and reviewing pages include:
• Is it clear what this page is for?
• How is it different from other pages?
• Is this message necessary?
• What is the first thing I notice on this page and does that match what should be the first thing a user notices?
• Is there anything distracting me from the main purpose of this page?

The language of the site and any system messages should be written in clear terms familiar to most users. Avoid jargon and overly technical terms.

Do not invent syntax. Do you mean “or” or do you mean “and”? Do not write “and/or”. It is one or the other. More distracting yet is “inside/outside”.

Do not over use punctuation. Users are not generally computer programmers. The comma is understandable. “Jim ate rice, the brown kind.” Parenthesis have a purpose, but can be just distracting. “Jim ate rice (the brown kind).”

Strunk and Whites. The Elements of Style belongs on every proofreader’s desk.

All text should be proofread for spelling and grammar mistakes.

User Interface

Link text should contain clear, action words that indicate what the link points to. Avoid using “Click Here” and other test that refers to the document. Remember, the user should be aware of the message, not the website.

Links to non-web formats, such as PDFs, should be identified by noting the format in the link text.

Users associate underlined text with links. So, do not underline text that is not links, this includes headings, page and section identification, and titles in general.

Help user to avoid errors. For example, if data need to be entered in a specific format, provide clear instructions for doing so. Do this consistently. Examples:

Home Phone _________________ +1 (607) 555-1212
URL ________________________ http://www.mydomain.com

Does the text wrap around images correctly? Is it spaced away from borders and other graphical edges?

Did you use H1’s for page titles, H2’s for sub-headings, etc.?

Does each page have a unique <title> ? Does the title match the page content and purpose. Remember, these are the words displayed by search engine returns and the default words stored in bookmarks.

Strong passwords are essential to data security. Passwords are easy enough to produce, but they are not as effective as they should be because human foibles lead to compromises. Humans find strong passwords hard to remember so they write them down. They find them hard to say so they send them in email to coworkers.

These problems can be overcome and it can even be fun.

Make the strong password by making it along with a mnemonic. Like this:

Which bear ate up my donut holes
Wb8^mdOs

Now we can refer to this as the “donut” password without giving much away and it is memorable.

We can also make a family of passwords:

Bears do not care for those expensive holes
B!c4t$Os

Is it useful to have two related passwords that may be called “the bear question” and “the bear answer”? Say, for the admin password and the staff password?

Assigning words to the shift characters on the number keys is the trick here, but not too tricky. “^” is top or up or above. “@” is at or each. “$” is money or cash or dollars ……
You get it.

Now we have pass words that can be remembered without using word combinations, i.e., thefudge, or word bastardizations, i.e., es$enti@l, either of which is easily cracked.

Now we also have passwords that can be named even in the presence of strangers without giving them away, you could just say, “Try the bear answer.”

You may also use these ideas to make your own private passwords.

While you are thinking about this, it is also a great way to generate “family passwords.”

Jane never cared for asparagus
jnc4asparagus or jnc4agus

Have fun!

After perhaps too much futzing, we ended up with a small business phone system without any monthly PBX charges other than the analog phone lines from the phone company.  We had an automated phone directory, the phones worked as intercoms, voice mail turned into attached WAV sound files send as e-mail.  I was able to add phone extensions easily.  I love Asterisk!

Perhaps the success went to my head, since I set my sights on a fancy conference room phone for our growing business.  I bought a used Cisco Polycom CP-7935 phone for 1/3rd the price of a new device.  I’ll admit it – I’m an amateur at telephony and didn’t know the difference between SIP and SCCP.  It turns out that Trixbox only supports SIP extensions by default, and this conference room phone requires SCCP channel protocol.

So finally I am coming to the purpose of this post – simple instructions on to hook up a Cisco VOIP phone (that only uses SCCP protocol) to Trixbox version 2.8.  The following instructions were gleaned from some Google searches, and I am summarizing them since no one had simple steps online for the most-recent Trixbox version.

First we need to install some packages and retrieve the most-recent version of an SCCP solution from SourceForge.  Do the following commands signed in as root user on your Trixbox server:

# yum install asterisk16-devel gcc subversion
# cd /usr/local/
# wget http://sourceforge.net/projects/chan-sccp-b/files/V2/Chan_SCCP-2.0_Final.tar.gz/download
# tar xvfz Chan_SCCP-2.0_Final.tar.gz/download
# cd /usr/local/chan_sccp-b_20090602
# make

If you perform the above steps, you will see that the make operation will not work without modifying some source code.  Thanks to a posting (http://lostentropy.com/2009/09/28/making-chan_sccp-build-with-asterisk-1-6/) I learned that I have to change a reference to a constant from CS_AST_CONTROL_T38 to CS_AST_CONTROL_T38_PARAMETERS.  Make this change to the file /usr/local/chan_sccp-b_20090602/sccp_pbx.c on line 587 using your favorite text editor.

# make
# make install
# amportal restart

Now comes the hard part.  The CP-7935 gets its provisioning file from a TFTP server.  Reset the CP-7935 to its default factor settings.  Next use the device menu to set the TFTP server to your Trixbox IP address (follow steps in the manual from the Cisco website).  The Trixbox should have its TFTP service activated by default; for my server the TFTP directory on the server is /tftpboot/.

Monitor the TFTP log file (/var/log/atftp.log) while you reboot the CP-7935.  You should see a request in the format “SEP#.cnf.xml” in the log file, where the “#” is the MAC address of the CP-7935.  Now create the following file /tftpboot/SEP#.cnf.xml (mine is /tftpboot/SEP00e0752442c5.cnf.xml) with this content and replacing the TRIXBOX_IP_ADDRESS with your Trixbox server address:

<Default>
<callManagerGroup>
<members>
<member priority="0">
<callManager>
<ports>
<ethernetPhonePort>2000</ethernetPhonePort>
<mgcpPorts>
<listen>2427</listen>
<keepAlive>2428</keepAlive>
</mgcpPorts>
</ports>
<processNodeName>TRIXBOX_IP_ADDRESS</processNodeName>
</callManager>
</member>
</members>
</callManagerGroup>
<authenticationURL></authenticationURL>
<loadInformation</loadInformation>
<directoryURL></directoryURL>
<idleURL></idleURL>
<informationURL></informationURL>
<messagesURL></messagesURL>
<servicesURL></servicesURL>
<versionStamp>{Apr 03 2010 12:00:00}</versionStamp>
</Default>

The last configuration entry tag for <versionStamp> is important since it is used by the device to determine if the settings have changed.  Update this versionStamp value to a later date to force the device to reload the settings in the file.

Next we need to write the SCCP configuration file that Asterisk reads. First make a backup of the existing file, and then we will replace it with one tailored for our solution.

# mv /etc/asterisk/sccp.conf /etc/asterisk/sccp.conf.bak

Here are the new contents for the file /etc/asterisk/sccp.conf, and remember to replace the all-capital letter phrases but the specifics of your setup, for example TRIXBOX_SERVER_IP_ADDRESS is replace by the IP address of your trixbox, and the SEP00e0752442c5 with the string “SEP” and the MAC address of your Cisco phone.  Our phone model is 7935, so you will also need to change this to your phone type.

[general]
servername = trixbox
keepalive = 60
debug = 1
context = from-internal
dateFormat = M/D/YA
bindaddr = TRIXBOX_SERVER_IP_ADDRESS
port = 2000
disallow=all
;allow=alaw
allow=ulaw
firstdigittimeout = 16
digittimeout = 8
digittimeoutchar = #
autoanswer_ring_time = 1
autoanswer_tone = 0x32
remotehangup_tone = 0x32
transfer_tone = 0
callwaiting_tone = 0x2d
musicclass=default
language=en
deny=0.0.0.0/0.0.0.0
permit=TRIXBOX_SERVER_IP_ADDRESS/255.255.255.0
localnet = 192.168.93.0/255.255.255.0
dnd = on
rtptos = 184
echocancel = on
silencesuppression = off
trustphoneip = no
tos = 0x68
private = on
mwilamp = on
mwioncall = on
blindtransferindication = ring
cfwdall = on
cfwdbusy = on
[devices]
type = 7935
autologin = CONFERENCE_PHONE_EXTENSION
description = Phone7935
keepalive = 60
transfer = on
park = on
cfwdall = on
cfwdbusy = on
dtmfmode = outband
imageversion = P00308000100
deny=0.0.0.0/0.0.0.0
permit=192.168.93.3/255.255.255.255
dnd = on
trustphoneip = no
private = on
mwilamp = on
mwioncall = on
device => SEP00e0752442c5
[lines]
id = CONFERENCE_PHONE_EXTENSION
pin = 1234
label = CONFERENCE_PHONE_EXTENSION
description = Conference Room
context = from-internal
incominglimit = 3
transfer = on
cid_name = Conference Room
cid_num = CONFERENCE_PHONE_EXTENSION
trnsfvm = 1
secondary_dialtone_digits = 9
secondary_dialtone_tone = 0x22
musicclass=default
language=en
rtptos = 184
echocancel = on
silencesuppression = on
line => CONFERENCE_PHONE_EXTENSION

We’re almost done.  Now restart the asterisk service:

# amportal restart

Finally create an extension in the Trixbox administrator interface, and make sure it matches the value of the CONFERENCE_ROOM_EXTENSION in the sccp.conf file above (we used “30″).  Do a hardware reboot of the conference room phone.

Missing from this posting is any explanation about firmware and provisioning.  Cisco sells firmware upgrades to their devices, and we bet that the existing firmware on the used conference phone device was sufficient.  Our bet paid off.

In summary, this solution may still take a few hours of work for your particular Cisco phone.  The payoff is grand – our office Trixbox solution saves us money daily by not having to lease or maintain an expensive private PBX system.