A couple good reads from Ajaxian recently on new Google features:
- Speed Tracer: Google announces new tools and more thanks to the GWT team
- A busy Chrome week: Extensions, Web Sockets, and more
This post seemed intriguing, until I actually started reading it. Then it got wildly confusing, so I started digging into the supporting links within the post, only to get even further confused… So I clicked further into the supporting links of that post, and the supporting links for that post, before I finally came across this post, which does a really great job of explaining exactly what a data URI is, how it differs from a URL, how to use them, their benefits and their restrictions. In the end, I decided I would have been better off having never clicked the very first link to begin with… Maybe you’ll feel differently… :-)
A subject that seems to re-rear its ugly head every several months or so, it is usually one worth taking a quick glean through, just to refresh any rusty cranium nodes… And Rock Solid HTML Emails, from 24Ways, is another really good wrap-up of everything developers need to remember when they have to build HTML emails.
Another offering from 24Ways is Self-Testing Pages with JavaScript which, on the surface seems pretty cool, but also seems kind of rigid, requiring you, the front-end developer, to catalog all of the elements you think the back-end developer might forget or mess-up. Well, that seems a little tough to me, but not totally useless. I mean, in this page, for example, it would be really easy for me to look through and create an array of elements, IDs, and classes that are pretty important to the page’s integrity, but it would also be easy for me to forget one element, or for my fictitious back-end developer to err in some way that my array would not catch… More useful to me would be a test that compared the back-end developers output to either an XML or JSON tree for comparison, something like:
{
'body' : [{
'header' : ['h1', 'p.tagline', 'form'],
'section' : ['article', ...],
...
'footer' : ['nav']
}]
}…that the JS could then loop through and log any missing elements, IDs or classes. If implemented correctly, the above format could also catch things like missing closing tags, because a missing </header> tag would then make header > section invalid… You could also easily omit elements that are not imperative to a proper layout, such as Microformat classes, etc.
I also think an easy implementation method would be a “teststructure” function in your site-wide JS that would allow you to append a querystring that would tell teststructure to append the self-test script for “this” page. Just a couple thoughts…
Okay, back to football!
Happy reading,
atg