Note on Software Problem and Incorrect Hyperlinks

Many files on this site were developed as Microsoft Word documents and converted to HTML in Word (which means that if they user views the page source, they show up as HTML, and the browser treats them as HTML, not as Word). In cases where I have embedded tables and hyperlinks within the tables, I have occasionally run into problems where the links, when clicked, show the previous link (or perhaps second previous link), instead of what is intended and coded. This problem seems to occur when I have bookmarks on the file, and particularly if I make a mistake inserting an internal hyperlink to a bookmark (for example, click on bookmark instead of "insert hyperlink" -- even if I cancel, the file gets corrupted). The problem may also occur in very large tables.

This happens only with certain versions of Word in Office 2003. When a file is browsed in Mozilla, it also show inappropriate hyperlinks on whole paragraphs of text or other entries. That does not show in Internet Explorer.

To fix the problem, I have to convert the file to Front Page editing. When I go into Front Page "Code" mode, I see extraneous hyperlinks placed there by Word, repeating earlier links, and then intended links are off by one or two. I have to go in and correct everything manually. This includes deleting extraneous href statement and correcting link targets. Sometimes, Front Page will freeze in design view when trying to correct this; I have to go into Code and correct the HTML code manually, which is tedious.

This has happened on a few movie review files and my photo directory. I hope I have them corrected. The visitor may email me when finding other problems. I welcome comments from anyone who may know about this problem. Perhaps I will start a blog on it.

This is a case of a software problem. It is not a virus or worm or malicious code. It is just incorrect code. It may be occurring because of the way Word tries to use the "span" subcommand to tell the browser how to read a file without generating all the href's, but it gets confused and starts generating runaway href's in inappropriate places. I was a mainframe programmer for 30 years and I have seen this sort of problem in IBM mainframe Assembler, where code runs away if an identifier is not on a half-word or full-word boundary (as caused a major production problem with monthly billing statements in the 1980s when I worked for a credit reporting company). 

The <span> element is a feature in XSL stylesheets. (For example, look at p 581 in the Wrox Professional XSL, 2001). Word appears to convert to HTML by going through a scripted procedure that resembles parsing an XML document with an XSD and XML to produce the dynamic HTML, and then saving it. The scripting code (probably VB) in Word could have vulnerabilities where it "gets confused" by certain actions or situations created by the author's Word document when he/she edits it with complicated bookmarks and internal hyperlinks. XSL is designed around Internet Explorer, and other browsers are only beginning to work with it.

I'll add here, that much of my content used to be on a site called hppub , a domain name that I no longer own. The viewer may find a few stray incorrect links to that older domain, which I am fixing when I find them. Most of them have been fixed. 

Contact me

Return to home