In principle, I've always championed the use of 'open' standards in developing web applications. In practical developments, though, I've sometimes strayed from that 'noble' path.
I certainly did initially with this website when I created a custom PHP script to strip special character-pairs coded into plain text to provide simple formating: paragraph breaks, italic & bold text, bullet points and so on. Plain text is wonderful for note-taking and the like, but natively provides zero formating capability. In some implementations (form completion, for example) you're even denied the right to start a new paragraph. My script solved all that, but the problem was that every time I wrote a new post I ended up going back to my crib-sheet to remind myself which character-pair code did what.
Recently I've rewritten the site to use Markdown: an established plain-text variant that employs intuitive layout- and character-triggers to convert plain text to HTML. It's a great improvement. Not in the time it takes to convert plain text to markup - my custom routine was quicker. But the writing experience is improved because I no longer find myself looking up the precise meaning of my style short-codes.
Of course, just to confuse things Markdown itself isn't completely standardised. It can come in several flavours that extend John Gruber's initial concept. Markdown Extra provides access to refinements I rarely need - and if I did would probably just insert as html - such as tables, or complex definition lists. But it's enough of a standard to fit the bill and I know through long use of Markdown how to achieve the effects I want without needing a crib-sheet. Plus, anyone could now read my source code and understand it at a glance. Progress, eh?