IBM Home Source Code Excerpt
Http://www.viphot.com
First, browser compatibility as an international company's site, ensuring that the browser can be considered in different browsers, this IBM's approach has to be admired. It cleverly uses a JavaScript (IBMCSS.js) to select a different style sheet according to different browsers, which is why we don't see the style table definition from the IBM page why we have directly opened the IBM page saved on the hard drive. Open IBMCSS.JS, analyze the code, we can find that IBM has designed four styles: IE1.css, IE6.css, R1.CSS, NS1.CSS. Most of these four style sheets are the same, the main difference is the unit of font-size: IE6 with%, R1 in PX, and PT in NS1. The following is the definition of the same style in different CSS: IE6.CSS: body {font-family: Arial, Sans-Serif; font-size: 78%;} r1.css: body, p, b, i, strong , EM, DT, DD, DL, SL, CAPTION, TD, TD, TR, U, BLINK, SELECT, OPTION, FORM, DIV, FONT, LI {Font-Family: Arial, Sans-Serif; Font- Size: 12px } ns1.css: Body, DT, DD, DL, SL, CAPTION, TH, TD, TR, SELECT, OPTION, FORM, FONT {Font-Family: Arial, Sans-Serif; Font-size: 9pt;} From the analysis of IBMCSS.JS logic, you can get the following style sheet and browser correspondence: r1.css: Macintosh IE4.0 or more, Macintosh Netscape 4.0 or more, Netscape 5.0 or more, X11 Netscape5.0 or more (XWindows For Linux or UNIX, I Guest) IE6.CSS: Windows IE4.0 or above, NS1.CSS: Other operating system Netscape4.0 or more IE1.CSS: Other operating system IE4.0 or above I am Windows XP IE6.0 and The four style sheets were tested in two environments in Linux Rat Hat 7.1 NS4.7, and four style sheets in IE6 can be displayed normally. It is only slightly different; in NS4.7, IE1.css And the page font defined by ns1.css is unclear, and the page displayed with R1.CSS and IE6.CSS can be displayed normally. At the same time, IBM also considers that some browsers do not support JavaScript, and tighten a sentence
Second, the standardized comprehensive meta content as a standard page, the detailed meta content is essential, this IBM has been doing very well, we can review the meta content in its 99 page:
Epc@www.ibm.com
"; for" http://www.ibm.com/ "; on" 1997.07.05t21: 46-0500 "R (N 0 S 0 V 0 L 0))" HTTP-Equiv = "PICS-Label"> < Meta Content = "(PICS-1.0" http://www.classify.org/safesurf/ "; l Gen true for" http://www.ibm.com/ "; by"
Epc@www.ibm.com
"R (ss ~~ 000 1 ss ~~ 100 1)" http-equiv = "pics-label"> Today's IBM page Meta content is more rich, from some of the following examples, we can see: 1, basics Information 2, Scheme 3, Description and Keywords The content is completely comprehensive, fully compliant with a regulatory page, I think it is quite worth our study.
Third, Web data stream analysis As an international company's site, only the simple click count is far from the detailed level of the marketing purpose, from the following code we can roughly guess the IBM site uses Keylime Software. Web Data Stream Analysis Scheme (in subsequently confirmed this in Keylime Software website):
