Boot traps of "Open Standards"
Today I launched Team Space with a larger group of users - and as usual those things went wrong that con go wrong. In my case it was Steven, who kept me from going home on time tonight. All he did is putting his ATOM news feed as he was supposed to do. It turned out that I underestimated the problems with HTML in news feeds. I wonder why a service run by google can't simply provide propper XHTML (as it is recommended for ATOM). Instead, they use HTML which cannot get parsed half of the times by XML processors.
Finally I isolated the problem in my code and corrected the code in a way that the processor tries to parse the HTML as it were XHTML. This works about half the time. For the other half the processor reports problems with some GET parameters in some links. This is because many systems still don't support the propagated semicolon separated GET parameter lists and more people don't know how to put correct strings in the href attribute of the A-tag.
I find it that it is a sad situation that not only private users ignore the open standards of the W3C, but also services provided by internet majors such as google don't pay enough attention to them.
I look forward to the time when the first malformatted RSS feed messes up Team Space. At least RSS2 does not allow encoded information in the summary field.