HTML validation

Question

It doesn't appear that HTML validation has gotten much attention here on the Apex forum. I did a search here and in the OTN forums at large and found no threads with the phrase "HTML validation" in them. In any case I'm trying to go down that path, meaning that I'm trying to create an Apex application wherein the HTML produced passes validation using the W3C validation tool.
Interestingly enough, I always thought Apex did a great job of producing valid HTML. Turns out it's not so good when you check it using the W3C validator. There are all sorts of quirks that show up that don't make a huge difference in display or function but they don't pass muster in terms of standards based HTML or XHTML.
One of the things I found that's quite pervasive, and probably a relatively easy fix, is that when ampersands are inserted into URLs (using things like the #WORKSPACE_IMAGES# substitution string) that those ampersands are not properly escaped and this causes a bunch of cascading errors as a result.
So my question is this: is anyone out there from the Apex development team at Oracle looking at this? Or is it not even on the radar?
As far as I'm concerned, it would be really nice to be able to run an Apex app through the W3C validator and have it pass with flying colors. Does any one else care about this? Or is this thread just going to die a quick, quiet and painless death?
Where do you stand on HTML validation? Is it a moot point if your application looks/behaves the way you want it to?
In case you want to try this out for yourself take a look here: http://validator.w3.org/
Earl
Edited by: EarlL on Dec 9, 2009 3:15 PM

Answer 1

Earl
It doesn't appear that HTML validation has gotten much attention here on the Apex forum. I did a search here and in the OTN forums at large and found no threads with the phrase "HTML validation" in them.[You did not find the right incantation|http://forums.oracle.com/forums/search.jspa?threadID=&q=html+AND+%28validator+OR+standards+OR+compliant%29&objID=f137&dateRange=all&userID=&numResults=15&rankBy=10001]—there have been more than a few, [going back quite a while|http://forums.oracle.com/forums/thread.jspa?messageID=951069] and [right up to today|http://forums.oracle.com/forums/thread.jspa?messageID=3956045].
One of the things I found that's quite pervasive, and probably a relatively easy fix, is that when ampersands are inserted into URLs (using things like the #WORKSPACE_IMAGES# substitution string) that those ampersands are not properly escaped and this causes a bunch of cascading errors as a result.[I pointed this out long ago|http://forums.oracle.com/forums/thread.jspa?messageID=1244310] for exactly the same reasons. If this was fixed it would be a lot easier to identify more significant problems.
So my question is this: is anyone out there from the Apex development team at Oracle looking at this? Or is it not even on the radar?Although the ampersand issue has not been fully resolved, there has been resolution or improvement of a lot of the areas that caused validation errors at the time of that post. However (to the indignation of the majority I suspect) I would have preferred more progress in this dull and invisible area than on more obvious enhancements, on the grounds that getting the basics sorted would provide a better platform to develop these new features on, with less possibility of [introducing further non-compliance in those features|http://forums.oracle.com/forums/thread.jspa?messageID=3228587&#3228587].
Does any one else care about this?Yes. See links above. I was trying to find the time to jump into that current thread.
Where do you stand on HTML validation?Coincidentally, I came across a link to this earlier tonight, which as it turns out is close to my own thoughts on XHTML and validation. Browsers are more forgiving than compilers (as a programmer I wish that they weren't), so (X)HTML and CSS validation plays the same role in my XHTML/CSS workflow as the first pass through the PL/SQL compiler does with database code, or lint-type tools do in other languages: it generates errors and warnings that inform me that there is something wrong with that code. If I can fix it I do, but if it comes out of the APEX engine rather than a template...
Is it a moot point if your application looks/behaves the way you want it to?Looking at the way HTML and CSS are used in the built-in templates, the solutions provided in this forum, and some of the projects I've worked on, that seems to be the majority view. Little or no thought is given to using semantic HTML to provide structure, or CSS that uses anything other than basic class selectors (even [<font> elements usually put in a Forum appearance every month or two|http://forums.oracle.com/forums/thread.jspa?threadID=998662]), and standards-based development in APEX has suffered because of it. Arguments have been made that it was/is necessary either due to lack of (cross-) browser support, and/or to reduce the learning curve for those starting with web technologies. The former is (thankfully) becoming less of an issue, and I have disagreed with the latter point since it was first put forward: that was the recipe for [tag soup|http://en.wikipedia.org/wiki/Tag_soup] in the first place.
It would be preferable to be in the position where the level of standards compliance in our applications was completely in our own hands through having control of all the outputs generated by APEX, either through some kind of exposure of the engine, or a move to entirely template-based output. Those who have a preference or need for standards compliance would be able to make the necessary efforts and the rest could ignore it. I'd like to have that option.