Style -- Templates and Constants

"Jack Waugh" <[email protected]> Thu, 09 Sep 2004 17:19:39 -0000
Newsgroups gmane.education.weblabs.requests
Message-ID <[email protected]>
(The reason I'm asking this here rather than in
groups.yahoo.com/group/ToonTalk is that I don't want to insinuate that
TT is yet available to the public.  I have tried the public version
enough to be convinced that I wouldn't recommend it to anyone; you
can't save anything complex at all.  How many people have bought the
CDs, anyway?  I posted a query over there about what people were doing
with TT and no one has answered.)

==================================================================

Style -- Templates and Constants
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Consider the programming of a robot that wants to request services
from other robots.  It doesn't matter whether the robot is given a
bird via which it can send the request, or is given a team of robots
on a truck to which it can add a box containing the request.  The
point is that the programmer desires that the robot makes a request.

One usually builds up a request in the form of a box, where the holes
have specific meanings to be understood by the robots that are going
to receive and interpret the request.  For example, one hole could
hold a verb to say what general kind of service is being requested,
and subsequent holes could hold arguments that hook the request to the
specific data it's to act on and to a destination where to send a
calculated result.

So the robot's programmer is interested in using the appropriate
_template_ when building the request; this would be a box with the
correct number of holes, maybe some labels to remind what the holes
are respectively for, and perhaps in some low-level cases, some
constants (more templates, etc.) that the receiving robots can use in
their work.

My question is this:  Which style is more desirable, building up these
templates in the robot's programming, or having the robot find each
template that it wants at some fixed index into some box (the box to
be made available at runtime) (e. g., always look in the J-th hole)? 
In one case, the robot's training specifies the template itself; in
the other case, the robot's training specifies an index into something
at which to find the template.

To anticipate that some of you may answer that the place to find
templates is the main book, and that the robot should find them by
dropping a name on the book and causing it to turn to the appropriate
page, I say such a proceedure seems nonportable between the
environments belonging to different players.  What if someone else
uses the same name for a different template, for example?  Would we
have to use qualified names, with a central registry of which ones
were in use?  This sounds awfully involved, even compared to having
each robot programmed to build up copies of the templates its
programmer wants it to use.

As for use of a book other than the main book, what are the
implications for the directories and files of a TT user when books get
massively copied around?  Can their use lead to a mass-storage memory
leak?  Performance penalties compared to the use of boxes?

Opinions on building the templates in the training vs. indexing for them?



------------------------ Yahoo! Groups Sponsor --------------------~--> 
Make a clean sweep of pop-up ads. Yahoo! Companion Toolbar.
Now with Pop-Up Blocker. Get it for free!
http://us.click.yahoo.com/L5YrjA/eSIIAA/yQLSAA/o1XolB/TM
--------------------------------------------------------------------~-> 

 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/WebLabs-requests/

<*> To unsubscribe from this group, send an email to:
    [email protected]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/