Note: My Web pages are best viewed with style sheets enabled.
De gustibus non est disputandum.
My pages have been hand-coded according to HTML 4.01 (using no Web page tool) and should be viewable with any Web browser that supports that HTML version. In accord with the Viewable With Any Browser Campaign, my pages follow these guidelines:
All my files and directories use the standard naming convention for UNIX. That means none of my links require %20 inserted for a blank character within the name. (This does not mean that links from my Web pages to external files — over which I have no control — also meet this guideline.)
Some Web developers encode the date of a Web page into the page's URI, changing the URI whenever the page is updated. Some tools used by Web developers generate dynamic URIs, automatically changing them each time pages are changed. Since this requires changing the links to that page in other pages of that site, those pages also have their URIs changed as they are automatically updated. The result is that bookmarks ("favorites") cannot be set to specific pages within the Web site, and external pages in someone else's Web site cannot have links to those pages. I do not do this. My URIs are constant.
When I update a page, I generally test all the links — to external Web sites and to my own Web site — on that page. If I reorganize pages and thus change their URIs (e.g., when I moved this and other pages into my /internet directory), I leave forwarding pages at the old URIs for at least three months.
*** Begin Right Sidebar ***Contrast between text and background colors is very important. Place your mouse over each of the colored areas below to see the text as black against a white background. (For some browsers, that is the way the text will appear without placing your mouse over the area.)
Even when there is sufficient contrast, light text against a dark background causes a distraction that interferes with gleaning the meaning of the text.
*** End Right Sidebar ***
On occasion, I have a small amount of text that is not the default color (as above). For example, I use red for emphasis. All text colors that I specify provide sufficient contrast with my pale green background and also with a white background. The pale green background is sufficiently different from true green that a person who is red-green color-blind should not have trouble with the content of my pages.
When a user hovers his or her cursor over a link, yellow shading appears in the background. This helps to identify the text as a link. For those who cannot see the yellow shading because of limitations of their browsers or because of a visual handicap, nothing is lost in viewing my pages.
For browser compatibility, I always specify colors with hex codes (e.g., #ff0000 or #f00 are red) instead of color names.
I do nothing to inhibit a user from forcing his or her own colors for text or background.
Most images are under 10K in size. All are under 75K.
Except for some photographs, all graphics are loaded from the same Web server as my Web pages. There is no dependency on other servers of unknown reliability. To conserve space on my Web server, however, some photographs are loaded from Flickr. Links to photos on Flickr are indicated with an icon showing a camera .
My pages have a faint background image (sometimes called a watermark). The colors and intensity have been chosen to minimize interference with reading foreground text. (NOTE: Mozilla bug #124150 may cause scrolling of my pages to be jerky on computers with older, slower memories.)
While some of my oldest pages do not specify any character set and some not-so-old pages specify the WINDOWS-1252 character set, I now use only ISO-8859-1 to ensure that a broad audience — including those using non-PC platforms running non-Windows operating systems — can properly view my pages. With two exceptions, ISO-8859-1 is explicitly specified in all pages displaying in the lower-left corner (as does this page). The two pages that discuss and use escape sequences also must still use WINDOWS-1252.
Please note, however, that photographs and other graphics on my Web pages are sized in pixels. For larger photographs, right-left scrolling might be necessary if your browser window is not maximized. Such scrolling is not needed on a PC with 800x600 resolution even with an Office toolbar at the right edge taking space away from a maximized browser window.
I do launch a new browser window when I think you might want to switch back and forth between the new page and the page from which it was selected (e.g., my list of useful Internet tools launches a new window once for its many links to my glossary of Internet terms). Otherwise, I consider the unnecessary proliferation of windows a damned annoyance almost as bad as popups, and I won't inflict that unnecessarily on those who browse my pages.
*** Begin Right Sidebar ***My counter counts each access to a Web page where it is installed. "Access" includes not only when someone views that page, but also when a search engine bot visits to index it. There are also other non-viewing accesses. These are all counted. My analysis indicates that many accesses are non-viewing. Thus, my counter reports more visits than the actual number of viewers.
*** End Right Sidebar ***
The counter is a server-side include (SSI) script, a feature of Apache Web servers (and others). Other SSI scripts are used by some of my pages. These include displaying the number of days or weeks to some future event or since some past event, displaying the date of a page's latest update, and generating the Index of my Web site.
When using a SSI script, the Web server inserts the displayed information into the affected Web page as if that information already resided in the HTML file. The process changes the timestamp of the file to reflect when the script executed, overriding the timestamp of the file on the Web server (which reflects when the file was last updated). This affects caching since the page from the Web server will always have a later timestamp than any cached copy. (See Mozilla bug #233293 for a suggested browser enhancement to compensate for this effect.)
All of my scripts reside on my Web server, within my domain. There is no dependency on external servers. The results all appear in-line as part of the text of a page, not as graphics.
Because I have seven distinct icons, none of them are named favicon.ICO. Instead, each page has a LINK declaration to identify which icon to use.
used on my Oak Park pages.
used on my copyright page.
used on my 404 page.
used on my PGP pages.
used on pages about Judaism.
used on most of my other pages.
Overall, I have tried to make my Web pages adaptable to the environment where they are viewed — to how you have configured your own computer — without requiring you to adapt to the constraints of my pages.
Remember: The above applies only to those pages of mine that are marked . If you encounter a Web page of mine that is not thus marked, please let me know.
Also, just because a marked page should indeed be viewable with any browser does not mean it will always appear as I intended. That is a function of the browser you use, the platform on which it is executing, and how you have tailored your user settings and options.
Not all browsers support style sheets or all of the HTML 4.01 features (e.g., tables, images). I have tried to format my pages so that they can still be viewed without style sheets or images. Where I use tables, however, tabular presentations are indeed intended; I don't know how my pages with tables appear in a browser without the capability to display them.
Jakob Nielsen is a principal in a consulting firm that specializes in training Webmasters, testing Web sites, and performing usability studies on Web-based services. Nielsen's Web Style can be quite useful to those interested in Web design for broad audiences.
Every so often, Nielsen publishes a list of "Top Ten Web Design Mistakes". Here, I analyze my own Web site in terms of Nielsen's list for 2007.
This refers to search engines that operate on a specific Web site. I use FreeFind, an outside search engine capability. FreeFind controls how my site is crawled and indexed.
I disagree with Nielsen's assertion that the search interface should be on my home page. Different search engines use different queries. Too often, when I search on someone else's Web site, I have to guess at how to create a query. For my own Web site, the search page has some simple guidelines for searching, along with a link to more detailed information on how to create a FreeFind query. All this is just too specialized for a home page.
I use PDF only when working from hard-copy, which I scan into a PDF file. Against over 300 HTML files (Web pages), I have only two PDF files on my Web site.
I don't set the color of links. I leave them at the default colors set by the users.
Here, Nielsen is referring to how humans scan text with their eyes, not to electronic scanning. He advocates the use of heads and subheads, bulleted lists (as seen above in this page), and other writing techniques. However, he also advocates short paragraphs and a simple style of writing; for my response to that, see Content That's Not Written for the Web (below).
I don't do this. I use only relative sizing of fonts, leaving the main content of each page at the user's default size. (See also Legibility Problems below.)
This is primarily a concern for commercial pages. However, I too want search engines to find my pages. Occasionally, I use Metacrawler to search for the topics covered by my pages, to review both the page titles and page descriptions. I also review the search reports produced by the FreeFind search engine used by this Web site. (See also Bad Search above.)
I carry no advertisements in this Web site. I hope none of my pages look like advertisements.
This is the point of the "Viewable With Any Browser Campaign". While previously assuming that Internet Explorer was the only browser to consider, Nielsen now recognizes that well-run businesses " … don't turn away customers just because they prefer a different platform." After all, over half of those who surf the Web use browsers that are not Internet Explorer.
However, Nielsen also includes creating a consistency both within a Web site and with other Web sites. My pages are indeed consistent with each other, a result of using templates that are based on a single general template. Consistency with other Web sites is nearly impossible since other Web sites are rarely consistent with each other. I am not consistent with other sites in the way I present links. I am also not consistent with Web pages that violate the design criteria presented on this page.
See New Windows (above).
Actually, I have been accused of over-answering, giving far more details than anyone really wants.
Prior versions of Nielsen's "Top Ten Web Design Mistakes" included the following:
When Nielsen advocates short and scannable text, I object. Obviously, my style of writing reflects the concept (said only partially in jest), "Why use 10 words when 1,000 will suffice?" Actually, Nielsen implies that content be "dumbed down". That insults the reader, and I won't do it. (See also Writing Style above.)
This primarily involves fonts problems. Some Web sites use fonts that are poorly designed for viewing on a computer monitor, no matter what resolution is available. Then there are fonts that have fixed sizes, are too small, or are colored with too little contrast with the background color. I don't do any of this.
I take some issue with "Explain what users will find at the other end of the link". Since I don't override the user's own color scheme for links and I do highlight links in yellow when a cursor is over them, I expect the user to see the link and understand from the context that there is some information about the linked term. Nevertheless, I do occasionally display a "tooltip" with an explanation of a link, especially when the link is in a graphic image and not text.
While Nielsen seems to advocate lists of links, my links are very often words or phrases within a sentence. However, some of my pages (including this one) also have lists of links for those who wish to pursue a topic further.
" … Flash collected the bronze medal for annoyance. … most of the Flash that Web users encounter each day is bad Flash with no purpose beyond annoying people." I don't use Flash.
Nielsen's focus is on commercial Web sites. Thus, he urges easy access to phone number, E-mail address, and postal address. Mine is a personal site, and I want to protect my privacy. Thus, I only provide my E-mail address; many of my Web pages (including this one) contain an envelope graphic at the bottom, which is a link to my "Send Me E-Mail" page.
Although not really part of the design of Web pages, one very serious complaint is that businesses do put their E-mail addresses on their Web sites or have Web-mail forms. However, they never answer messages submitted via the Internet. Except for the most abusive messages, I always answer my E-mail. I even answer some of the abusive messages, by putting them and my answers on my Web pages to show what jerks the senders are.
Others have also addressed user-oriented design of Web pages. Among them are the following:
14 December 2004
Updated 19 October 2010
"Internet" Table of Contents
David Ross home
The OP and DR digraphs are my original creations and are covered by my copyright. The other icons are either in the public domain or are owned by others under copyrights or trademarks. In the latter case, I have permission for their use.