We all need standards. In a car it’s useful to be able to know that everybody is maintianing the same standard as to which side of the road they drive on – mostly. Or that what Shell calls diesel will work in a diesel engined car. If everybody sticks to standards then we all know where we are and everything will work smoothly. And so it is in the world of web page design. Or is it?
There is a body, the W3c, which develops standards on which all web pages should be constructed. They have a page where you can test your website for compliance – here it is: http://validator.w3.org/
As an embryonic web developer back in the miasma that was in the days of Internet Explorer 2, Netscape 3 and AOL’s very own custom browser, standards were a right royal pain. Just getting a page to work and display was success enough. Now things have changed, the world has moved on.
There is a rich satisfaction when the green square of validation reveals itself that perhaps only someone who has struggled with proprietary html tags and attributes can understand. This is often followed by an intense depression as testing reveals that beautifully crafted html, that perfect layout which works in all modern browsers actually chokes Internet Explorer 6. Yes, it’s still out there. Yes we have to cater for it.
But I digress. We can write html which validates. Is there ever a case for allowing non-validating html code? I think the answer has to be a guarded “yes, sometimes”. Why? Well, once upon a time if you had come to jewelion.com you would have seen the proud boast “always bespoke”. Things have changed, the world has moved on. There’s so much that can be done in such a sophisticated manner that it is just not practical to develop it all from the ground up, to engage in needless wheel re-invention. Take the little carousel on the front page of this website – someone has developed that and very kindly made it available. However, they didn’t pay close attention to standards and it doesn’t validate. At first it produced over 20 errors – I have corrected them but there remains one.
As the page works – is the existence of one validation issue important? I think not.
