[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference vaxuum::online_bookbuilding

Title:Online Bookbuilding
Notice:This conference is write-locked: see note 1.3.
Moderator:VAXUUM::UTT
Created:Fri Aug 12 1988
Last Modified:Mon Jul 15 1991
Last Successful Update:Fri Jun 06 1997
Number of topics:440
Total number of notes:2134

281.0. "ONLINE.SHELF Doctype Discussion/Problems" by NAVIGO::GRANT (I've saved $2355.00 since I quit smoking.) Thu Feb 22 1990 10:31

The purpose of this note is to provide a common note to discuss the 
ONLINE.SHELF doctype.  If you have any comments, suggestions or problems,
please enter them as replies to this topic.

Thank you
T.RTitleUserPersonal
Name
DateLines
281.1Suggestions for alternative methodsVAXUUM::DEVRIESBy their notes ye shall know themMon Feb 26 1990 10:3241
    Sinice you opened the door -- is this the best way to build a
    bookshelf?  I think of VAX DOCUMENT as a formatter for large, complex
    books, not as a data processing or database management tool.
    
    I think of the task of creating a bookshelf as similar to that of
    creating a doc$designs.dat or doc$destinations.dat -- currently fragile
    and error-prone, if you just use a text editor, but an awfully simple
    format to require such a complex tool as VAX DOCUMENT.
    
    Current suggested alternatives for building the *.dat files could be
    extended to building bookshelf file as well.  These include:
    
    1) Document the format, provide entry templates, and let the users bang
       away at the file; maybe provide a verify tool they can run
       afterwards to show if the file is proper.
    
    2) Provide a full-blown editing tool to control the whole process. 
       This would hide the data format from the user and guarantee correct
       results, but it would have to be a full-featured editor and require
       the user to learn yet another complex interactive interface.
    
    3) Provide a straight-forward command interface like that in AUTHORIZE
       or (shudder) LICENSE:
    
    	$ EDITSHELF myshelf /ADD_BOOK=mybook/TITLE="This is My Book, So There!"
    	$ EDITSHELF myshelf /MODIFY_BOOK=mybook -
    		/TITLE="This is My Book, But You Can Read It, Too"
    	$ EDITSHELF myshelf DELETE_BOOK mybook
    	$ EDITSHELF myshelf /ADD_SHELF_mydisk:myothershelf -
     		/TITLE="I Made Myshelf Myself!"
    	$ EDITSHELF/LIST myshelf
    
       or words to that effect.  (Syntax is negotiable.)
    
    As you can tell, I'm warming up to alternative #3 these days.  It still
    leaves open the possibility of a binary file format that can be
    optimized for whatever the software wishes, while insuring that only
    legal entries are put in there.  And it would translate easily into a
    menu interface.
    
    Mark
281.2The shortest path is a straight line...AIRBAG::SWATKOElectrons are cheap. Trees are not.Wed Feb 28 1990 11:2332
Ah, here's that note I've been looking for...

If the goal is to create a simple data file (a bookshelf file), then the
route of running a text formatter on a marked up file has to be, in my
opinion, the most round-about and complicated way of doing it that I can
think of! (Actually, I'm restraining myself here and trying to be polite.)

There are lots of alternatives which would be much easier to implement,
easier to understand, and easier to USE.

My two nominations for best way to handle the situation are:

Mark's #3).  Provide a straightforward command line interface with DCL HELP
availible inside and outside the tool (if running on VMS).  At that point,
noone cares whether the actual stored format is binary or ASCII.

Mark's #1, slightly modified).  Document the ASCII shelf format and let the
user edit it with his/her favorite text editor.  Instead of a verification
tool, simply modify the bookreader to make the shelf reading code a little
more robust so it will point out an error in the shelf file instead of
falling on it'd face and dying.  That should not be very hard to do.  A
simple

	Syntax error on line nn of file xxxxxx
	<give actual text of erroneouos line here>

should be more than sufficient.

Solution #3, to me, provides a more "professional" and well designed
solution.

-Mike
281.3credit where credit's dueMARKUP::DEVRIESBy their notes ye shall know themWed Feb 28 1990 11:405
    RE: Mark's #3
    
    Mike, I knew you'd like it, since it was your idea!
    
    Mark