W3C home > Mailing lists > Public > www-jigsaw@w3.org > July to August 1996

Reasons against unravelling recursion in lookup.

From: Alexandre Rafalovitch <alex@access.com.au>
Date: Fri, 9 Aug 1996 20:24:01 +1000
Message-Id: <v02140b06ae30bf2b610a@[203.18.229.116]>
To: www-jigsaw@w3.org
Working on a new resource I found a situation where I would preferred to
have recursive lookup instead of loop in a Client. Which does step by step
lookup.

What I needed to do is for parent during lookup stage to know if a
descendent resource does not exist. The reason I need to do that is because
in situation when child does not exist, I want parent to take some evasive
actions (basically return redirect instead of non-found).  It can be done,
if the child of DirectoryResource does not exist, but if it is further
descended, then it is impossible, because lookup of a child is done not
recursively from directory resource, but from a Client. Therefore Parent
never knows if a child is in trouble.

Also, current filter model does not let me catch errors either, because it
is not called on error. I know that next filter model should be able to
work differently and it would solve my problems (hopefully), but  until
then (RSN) I can't do anything.

This letter is not a request for solutions, even though such would be
welcomed, but a food for thought about what was the price of unravelling
recursive lookup call.

Still digging,
Alex.


--------| I feel as confused as a baby in a topless bar. |--------
Received on Friday, 9 August 1996 06:24:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 9 April 2012 12:13:25 GMT