Conference vaxuum::document_ft

Add to that some problems with OVERHEADS.

1.  <author_info> expands to <author_info>(\\\) 
     instead of <author_info>([text]\[text]...) (or whatever the 
     correct notation is)

2.  Many (possibly all) of the OVERHEADS tags cannot be expanded
    by typing a few unique characters then expanding:

     sl^E   for <slide>
     aut^E  for <auto_number>
             .
             .
             .



Aside from overheads, selecting tag-groups from the menu causes 
tag-groups to be inserted into your text rather than simply displaying 
the menu.  The user's first instinct is to expand tag-group, which 
expands to <definition_list> since that is the first item in the menu.

I'm on a roll here.  One more thing: the App. A in Step-By-Step starts 
out ok, but there's a step in there that says to go to BUFFER $PASTE, 
then exit.  It would be very helpful to know the reason for and use of 
$PASTE.  A picture of the different kinds of buffers would be very 
helpful.

The good news: I introduced some students to DOCUMENT this week and LSE 
in particular today.  The response was very positive.  There's a 
movement abrewing out there...

Rose

    Thanks for finding the problems...I will fix these problems
    and get them into the next update.
    
    Julie

The template for <example_sequence> does not have the parameters.  It should be:

<example_sequence>([heading_info][\NONUMBER])
<endexample_sequence>

      ---Karen

ele  and [ele  expand to ELEMENT
                         ELEMENT-OR-TEMPLATE

And the user said, "This is good."



element and [element   expand to <ELEMENT>({element-name})

And the user said, "This is not so good."


Rose

    Re: .4
    
    In the second case, you have typed a complete token name.  LSE
    therefore interprets that as a specific token, and not as an
    abbreviation to be matched.  Thus, it completely expands the ELEMENT
    token, and ignores the fact that the token happens to be a proper
    prefix of another token.
    
    There are arguments in favor of both sorts of behaviors: always
    trying anything that matches, or stopping if we have a complete
    token.  Neither is 100% perfect.  The compelling argument in favor
    of LSE comes from the symbols IF and IFIX in Fortran.  The former
    is a commonly used keyword, while the latter is a rarely used function
    name; users would be really upset if everytime they typed "IF<expand>"
    they got a menu.  (There are examples from other languages as well,
    such as WRITE vs WRITELN in Pascal.)
    
    In the meantime, if you know in advance that you want the menu,
    then expand "ele".  If you know in advance you want
    ELEMENT-OR-TEMPLATE, you may find it easiest to expand "element-".
    
    Finally, if you forget, and accidentally expand "element" when you
    want ELEMENT-OR-TEMPLATE, use the UNEXAND command, which is bound
    to <GOLD><EXPAND>  (i. e. on the EDT keypad, <GOLD><CTRL/E>, and
    on the EVE keypad, <GOLD><CTRL/?>).
    
       Gary