The BBC site has to be ranked among one of the largest in existence, so many people must work on it there's no real way to keep track of each part of it.
If you do a "site:www.bbc.co.uk" in Google it reports 15,200,000 pages!
As for HTML errors, I imagine they use some sort of centralized WYSIWYG to handle it all, so blame that.
As for the "an overly long list of variables - could well be reduced", while amusing, it's easy to understand that whoever originally wrote or controls that list now doesn't know which variables are or aren't used somewhere so can risk removing anything without breaking somebody else's code.
I think that's the unfortunate inevitability with a site that's just sooo big.
As for 83k, it would take about 300ms to download on most UK broadband connections, the entire homepage loaded in under 1 second on my laptop here.
I think what's important to note is their good use of web technology, mainly HTML with small inserts of DHTML, AJAX and flash. Just noted a nice little flash clock on the main page. The site is structured well, loads fast, flows smoothly, and is compatible (unable to test this assumtion) with all platforms and mobile devices.
That said, I think that makes it an extremely good site.
No, I don't work for the BBC! Nor have I studied the code structure, I was just trying to view the site as an end user would.
[edited by: Dabrowski at 2:24 pm (utc) on April 27, 2008]