SCROLLIN' ALONG...
Overview Article
Part 1 -
About This Study
This study is an investigation into how different Browsers control their Default Scrollbars, by observing their behavior as a function of the overflow property of both the html and the body elements.
This topic is the starting point of what I term as my "Chasing-The-Tail Experiments", since I kept on chasing my tail due to interferences from other factors unrelated to the subject, along the way. As a result, I had to go back to square one many times, and try to identify, isolate and organize these UBO's under separate topics:
-
Privileged Ones
Identifying the Browser given privileges of the body, html and phantom elements, and try to Normalize them, whenever possible, by reseting the default values of their CSS properties. www.web-zonez.com/screen/duo/index.html -
100% ≠ 100% ?
An investigation into the Control Logic (or illogic) used by different Browsers - based on their height/width calculations, to control (i.e. activate or de-activate) the Scrollbars. www.web-zonez.com/screen/100/index.html
Eventually, having identified these UBO's and eliminated them from the equation, or, at times, "discounting" for them, one by one, I could finally arrive at the real experiments presented in this study. I must say, in the process I uncovered some unexpected skeletons in the closet.
Overflow Matrix
A 5x5 matrix was established - setting each possible value of the overflow property of the body element against those of the html element, to formulate the basic test structure:
- Default ("overflow" property not specified)
- Visible
- Hidden
- Auto
- Scroll
Please Note That....
- No assumptions were made as to the 'default' value of the "overflow" property (i.e. overflow not specified), for either element.
- Consequently, default state was included as a possible value in this matrix.
Background Colors
The most effective and powerful tool in these experiments was the use of different background colors (chosen for no other reason than to provide a sharp contrast) for each of three elements that constitute the basis of these experiments (i.e. html, body & #test) for the purposes of positively identifing the limits of each.
During the course of these experiments, it was observed that the background color of the scrollbar(s) (i.e. the color behind scrollbar itself) for Opera consistently coincided with that of the element it belong to (dark red for html, dark green for body). This provided a further and very useful tool in identifying clearly which element any given scrollbar belongs to. Furthermore, a gray background - a background color not used in these files, distinguished - beyond the shadow of doubt, the "Browser Default Scrollbars" from the others, which in turn confirmed the existince of a third layer - referred to as the "Phantom of Opera", - over and above the html element in hierarchy, to which the Default Vertical Scrollbar belongs to.
Elements Used & Settings
Html:
- The background was set to dark red (#a52a2a),
- The height and width were set to 100% for the entire range of experiments. (Reference: Privileged Ones) www.web-zonez/screen/duo/index.html
Body:
- The background was set to dark green (#006700).
- Furthermore, the padding and margin properties were also set "0". (Reference: Privileged Ones) www.web-zonez/screen/duo/index.html
- The height and width setting were determined by the definition of the test criteria for each phase of these tests.
#Test:
- The background was set to light yellow (#fffacd).
-
Inserted inside the body element for the purposes of conducting Vertical and Horizontal testing by varying its height/width settings, as follows:
- Vertical Scroll:
Height >> height of body & Width < width of body - Horizontal Scroll:
Width >> width of body & Height < height of body
- Vertical Scroll:
Tests & Results
Each test page used in these experiments is:
- Stand-alone with an Internal Style Sheet
- In the Standards Mode (XHTML 1.0 Transitional)
The following browsers/versions were used in this study:
- Internet Explorer 6.0 & 7.0
- Mozilla 1.7
- Firefox 1.5
- Netscape 7.1 & 8.1
- Opera 7.54 & 9.1
Within the scope of this study, if the results of two versions of a given the results were same, only the results of the most recent version were presented. If the results were different, then the results of both versions were presented.
Experiment Sets
Using the same test criteria and the matrix established, the tests were conducted in 3 Phases representing three different body sizes (height and width) in relation to those of the Peep-Hole (aka Screen) as follows. It should also be noted that height and width of the html element was set to 100% for the entire range of experiments.
-
Phase I: BODY < PEEP-HOLE:
www.web-zonez/screen/scroll/matrix2/index.html
This set of experiments use the body element is set to be much smaller (both in height & width) than the Peep-Hole. Again, the height/width values were arbitrary, but intentionally selected to be much smaller/larger than most screen resolutions in use and to facilitate evaluation of the raw test files by anyone. The choice of absolute values of height/width was based on the considerations of improved preciseness, and to eliminate any possibility of inaccuracies of in percentage calculations in different browsers. Since the experiments involved testing the Vertical Scroll (i.e. forcing the vertical Scrollbar alone) and Horizontal Scroll (i.e. forcing the Horizontal Scrollbar alone) behavior individually, 50 test files were neded for a complete set of experiments.
-
Phase II: BODY > PEEP-HOLE:
www.web-zonez/screen/scroll/matrix3/index.html
This set of experiments use the body element is set to be much greater (both in height & width) than the Peep-Hole. Again, the height/width values were arbitrary, but intentionally selected to be much smaller/larger than most screen resolutions in use and to facilitate evaluation of the raw test files by anyone. The choice of absolute values of height/width was based on the considerations of improved preciseness, and to eliminate any possibility of inaccuracies of in percentage calculations in different browsers.Since the experiments involved testing the Vertical Scroll (i.e. forcing the vertical Scrollbar alone) and Horizontal Scroll (i.e. forcing the Horizontal Scrollbar alone) behavior individually, 50 test files were neded for a complete set of experiments.
-
Phase III: BODY = PEEP-HOLE:
www.web-zonez/screen/scroll/matrix4/index.html
This set of experiments use the body element is set to be equal (both in height & width) to the Peep-Hole, by setting its height amp; width to 100%. Since the experiments involved testing the Vertical Scroll (i.e. forcing the vertical Scrollbar alone) and Horizontal Scroll (i.e. forcing the Horizontal Scrollbar alone) behavior individually, as well as a Combined Scroll format (i.e. forcing both the scrollbars), 75 test files were neded for a complete set of experiments.
What is Next?
I stongly urge you to review all of the experiments and their results, as listed above. I beleive that they represent quite a comprehenesive study which generated an enormous amount of data - not just data, but confirmed data! Just the sheer numbers will show that: 175 Test Files!
However, if you are impatient, or in a hurry, you can jump the gun and review the overall results and conclusion of these experiments, as presented in Part 2 of this article.
- Feed-back & Comments:
- Please send any comments, opinions, and even contributions on this subject, using the Feed-Back Form.
- Further Reading:
- You can review the list of Related Articles, with brief descriptions and links on the Auxiliary Panel,









