sgmllib.SGMLParser calls start tag and end_methods 
correctly until it encounters

        <a title="link1" href="url1">One</a>
        <br/><a title="link2" href="someurl2">Two</a>
        <a title="link2" href="url3">Three</a> 

the <br/> seems to cause its parsing to become 
confused and I conly get call backs for tag a twice (link 
1 and 3)
  

Logged In: YES 
user_id=38376

footnote: <br/> is an XML construct, and is not valid HTML.  
In HTML, "<tag/blah/" is short for "<tag>blah</tag>", so the 
BR section is parsed as

START br
DATA ><a title="link2" href="someurl2">Two<
END br
DATA a>

which is 100% correct.  For more on this topic, see:

http://www.cs.tut.fi/~jkorpela/html/empty.html

Logged In: YES 
user_id=38376

footnote 2: if you need to deal with broken HTML, use 
TidyLib:

http://utidylib.berlios.de/
http://effbot.org/zone/element-tidylib.htm

Logged In: YES 
user_id=38376

footnote 3: for the link case, also note that the HTMLParser 
module handles this in a more practical way (that is, it limits 
itself to SGML features that's actually used on the web).

Logged In: YES 
user_id=38376

closing, due to lack of feedback.  using HTMLParser instead
of sgmllib should solve the problem.