Slava Akhmechet had a good ramble around this topic a while ago on his defmacro blog, e.g. converting Ant XML tasks into Lispy looking S-expressions, etc.
I had no real problem with the parens thing - much less typing than HTML/XML - and most editors will keep track of them, indent and even colour-code them for you. Anyway, even HTML etc will try to ditch the closing tag if there are no children e.g. <foo blah="bar"></foo> is the same as <foo blah="bar " />.
I got permission to cross post my classmate's entire post:
I think the typical comparison of Lisp symbolic expressions (s-exprs) is with XML and not with HTML.
There was a pretty big hullabaloo some years back, just when XML was becoming a buzzword technology, comparing XML to s-exprs. Famously, Erik Naggum's anti-XML rant is always fun to read. The guys over at C2 also have some fairly extensive discussions on this matter.
Interestingly, your defense of closing tags echoes Paul Prescod's defense of XML, the only Pro-XML response I've ever been able to find. Unfortunately the link appears to be broken, and I can't find any mirrors.
My own opinion on the matter is this (just scattered thoughts):
As a programmer, both as a reader and as a writer of code, I've never found parentheses to be a problem. By contrast, the verbosity caused by Lisp's lack of syntax does impact the clarity of code.
A good editor will automatically tell you which parenthesis you are closing, and even automatically indent your code. Losing context is not something I have ever had trouble with.
When working with data, it is a very powerful idea to be able to directly write an in-memory data structure into a human readable structured text document, and then to read it back in again into a format that can be directly manipulated by your language. In Lisp, the idea can be even more powerful since your input file can even be directly processed using the full might of the lisp reader/interpreter. By contrast, I've never found an XML library that didn't leave me cursing its inventors and all of their ancestors. Retrieving data by manipulating an XML DOM or XPath is incredibly clunky and ugly.
EDIT: You can read Paul's article here, courtesy the Wayback Machine.
When I was doing that course, I found the parentheses relatively easy to get used to (with a decent editor, which DrRacket is). And the ability to use  as well as () (with the interpreter making sure they match) also helped. A bigger step, I found, was getting used to prefix notation. In particular with non-commutative operators.
So, for instance, (+ (a b)) is OK. But I have a really hard time telling my brain that (- (a b)) doesn't have a -a in it.