Fixing up and validating all of my old sites (whew!)

I've had some time today, and decided that it was finally time to upgrade some of my own web sites, and especially to make sure all the pages validate. The Carolina Web Solutions site only took about 10 minutes; it's a simple site and the code is pretty decent.

Not ancient and laughable, like it was on one of my other sites, The Patty Site. Bringing that site into the modern era took me five or six hours of what I would call hard work. It all validates now, and the CSS is a little better, and the pages are all templated and SSI'd. Unfortunately, the content is still pretty outdated, and the design leaves a lot to be desired! But at least getting things straight under the hood is a good start.

Now I'm working on this site, WebDevBiz.com - not the page you're reading, or any of the Blogger pages; they'll have to wait until I'm ready to delve into the Blogger code, because there's all kinds of scary non-valid code in there that I didn't write. But the main part of the site I'm responsible for, and I'm going to get it all to validate if it kills me! Hopefully it won't kill me.

A couple of things that helped:

• Firefox and the Web Developer Toolbar. One keystroke (Ctrl-Shft-H, if you must know) runs the current page through the W3C validator.
  
• I discovered that it seems to work best to go down the list of errors one by one, starting from the top. For some reasons unknown to me, fixing certain errors makes about 20 other errors below it go away, and I found that often I only needed to fix one or two things.
• Dreamweaver's Find and Replace was useful when the validator only mentioned cryptically that there was something wrong with a tag. I'm afraid there were a lot of tags on one particular page, and I finally found the culprit by using Find Next and scrolling through each and every one.

I'm using an XHTML 1.0 Transitional doctype for most of my sites these days, and often the only thing I needed to do was to find errant break tags that had been pasted in from somewhere else and needed to be changed to XHTML-style, and to find image tags without ALT attributes.

I'm glad that I've been pretty determined to write simple, clean code for quite a few years now, because even the pages I built before I'd ever heard of XHTML or validation weren't too bad.

This process also made me think about my use of tables for layout. More and more, I'm unashamed of this practice; my table-based layouts work fine everywhere they need to work, and it simply isn't true that they're hard to update. At least, not when I use them intelligently. I will admit that a certain site which uses way too many nested tables just gave me a lot of trouble, because the code really was hard to read through and troubleshoot. I haven't built a site with a careless use of tables like that in a few years, but now I'm all the more determined not to.

Anyway, within the next day or so I should have the non-blog pages of www.webdevbiz.com all validating. This is kind of like cleaning out closets or something - exhausting, yet interesting, and ultimately satisfying. :-)