Dissertation
Summary and Discussion
As an aid to the reader, this final chapter of the dissertation restates the research problem and reviews the major methods used in the study. The major sections of this chapter summarize the results and discuss their implications.
As explained in chapter one, the study reported here was an empirical study aimed to argue the uses of the Internet compared to other forms of media present and future, coinciding with discussing and testing current theories in relation to communicating on the Internet with text to provide a basis for further development into achieving a standard font size interval system. This study is not set out to achieve the perfect HTML coding for text legibility and accuracy on every single browser upon every different platform across the Internet, but more so to provide a strong platform of information and testing for others to continue to do so.
Ranges of techniques recommended for the addition of text to web pages were introduced in chapter 2, The Literature Review. It was necessary, then, to employ an empirical ‘trial-and-error’ process in order to test the legibility of text coded according to each technique. This would allow the combination of best practice from each method to give guidelines, which offer successful outcomes in the widest range of situations.
Summarising The Results
An adequate amount of research and testing has presented a solid base for continuing work into this subject. The results showed that using the elements ‘em’ and ‘absolute keywords’ relied on the size of the users individual default browser text size, thus allowing for the opportunity of the aesthetic value of the site to be impaired. Using Pixels (‘px’) and ‘pt’ did allow for consistency, as the browser did not change the text size when the default size was adjusted, but they do create issues of accessibility. The use of percentages was not affective by directly specifying them with regards to an attribute such as a header or sub header. They did prove affective when incorporated into the body as a base font size attribute of the CSS and combined with using ‘em’ to specify a header or text size. This method did not over rule the browser’s default text sizing but did allow for more specificity, compared to when ‘absolute keywords’ were incorporated. Initially a reset code was not used in the CSS, this presented problems again for aesthetic value due to each browser automatically applying default padding around the text along with boldness and line height. When the reset code was introduced and compared with the results that did not incorporate it, it was found to give much more control to the designer and removed the default settings that were pre mentioned, meaning the designer is able to set them specifically as required.
Discussing the Results
Some may find it good practice to leave a webpage’s main text alone and allow for the browser to set it dependent on the individuals pre-set default size, but since most browsers size up rather than down we are left with text that is viewed almost un professionally big, thus destroying the aesthetic look of many websites, hence the use of cascading style sheets (CSS). It is in this researchers belief that text should be sized a little smaller than the default size but certainly larger than the unreadable 9 pixels at 72ppi. On the basis of this study alone, it is difficult to be certain at this point which is the most affective method for maintaining accurate text legibility and consistency on the Internet. One thing is certain however incorporating the use of reset codes into the CSS, allows for more control over spacing surrounding the text and also its decoration by the designer rather than the browsers default settings (see table 5). It would seem that the use of ‘absolute keywords’ provides a simple and consistent text sizing relationship, attributing to this if the designer does not size text below the ‘medium’ then at a default browser size of ‘10px’ the writing would still maintain legibility (see table 8). However at the other end of the scale when the browser’s default is set to ‘16px’ anything ser larger than ‘large’ will appear extremely big on the screen, also opera appears to display ‘keywords’ a size larger than other browsers. Pixels used to be a popular choice for designers as they show exactly what a page will look like across all browsers, however they cannot be re-sized, similar to using ‘pt,’ which for users with impaired vision means small text cannot be seen creating a major accessibility issue that is breaking the Internets legislation section 508, an amendment to The Rehabilitation Act of 1973 which focused on making technology accessible to the disabled (http://www.eeoc.gov/policy/rehab.html). The use of ‘em’ for specifying text size is a strongly acceptable means of sizing text as it does allow for browser re-sizing across all browsers and thus is in compliance with accessibility. When used in Internet Explorer however and sized at a measurement of ‘1em,’ which is a common element for designers wanting to achieve good type layout, the text becomes unreadable if the individual has their browser default text sizing set small. Through testing it was clear that if the base font-size is set to 100% in the body then this stops the text displaying illegible when the browser’s default text size is set to small. The use of percentages when incorporated directly with the main attribute rather than used as the base font in the body, appears to revert the sizing, displaying the main header at a smaller size than the body (see table 10). Further more continued testing showed, even when a font size percentage is incorporated into the body as a base font size it displays the text worse (see table 11). While this single study cannot provide a sound basis for professional use when attempting to achieve consistent text legibility, this study would suggest that from some testing, the most appropriate method of setting accurate text legibility and consistency would be to use a base font size percentage in the body between 65% and 100% depending on the required sizing and then specify other elements directly using the ‘em’ measurement, this is similar to Macdonals theory where he states “For example if you want to make sure your text appears normal size use this rule: ‘body{font-size: 100%;}’ or if you want to make sure your text size appears smaller use this rule ‘body{font-size:74%;}”(Macdonald, 2009). Further work by (Schmitt 2006) states that by setting the base font size in the body to 62.5% then setting a font size of 1em in relative attributes such as the sub header means Internet Explorer will not produce text any smaller than 10 pixels. This is due to at 62.5% 1em equals that of 10 pixels, thus as long as the designer does not size his text smaller than 1em the text will be legible. However I would argue that this again is dependant of the users default text sizing on their browser and while the text produced may not be smaller than 10 pixels for a browser with a standard default settings it may be a problem for users with smaller text sizing settings. A major unanticipated finding resulted from the use of percentages when directly incorporated in to an attribute such as the header or sub-header. The results showed that even when the main header was set with the highest percentage sizing (100%) and the body text lower (50%), the body text actually appeared larger than the header across all browsers for unknown reasons, which is obviously against standard presentation layout. Additional testing seems needed in regards to placing text in lists and nesting divs inside other divs to view the outcome. Also only the most recent up to date browsers were used for the testing platforms, some of which certain individuals may not have and thus it seems that more testing is needed to view the display outcome in older browsers. Also further testing and research into using plus and minus elements for setting the text size is needed, when percentages are used as the base font size ( header{ font-size: +3;} ) The theory is that the text will be plus or minus that of the base font size. Schmitt also states that by using the term ‘important’ at the end of a text sizing element will tell the browser to put emphasise on this rule (Schmitt 2006), this researchers theory is that it will still not affect the browsers default settings but it is an opportunity for further testing.
