Re: requesting SWAP information

Hi Vijayanand,
  That's great.  You seem like you certainly got off on the right
footing.  SWAP is basically a derivative of both the HTTP/1.1 and
WebDAV protocols.  Keith Swenson's (kswenson@ms2.com) has some powerpoint slides
on Internet scale workflow to provide some examples from a talk he gave
at the work activities conference last year.  I think the
electronic configuration and purchase of some number of computer systems
is a useful example.  

Also, the approach we've been taking is  that we intend to simply use the 
Apache mod_dav and extend its methods to support the new SWAP methods.  Microsoft
also has Web Folders for Windows 98 and 2k.  This is the sort of user interaction
we are imagining wrt initiating, tracking, and terminating ongoing work. 
You might consider playing around with that.  An easier mechanism that I 
tend to use is to run an Apache server with the mod_dav plugin, telnet
to the port (usually 80) and issue the protocol headers and statements by hand,
tweaking the code to support the desired inputs and results.  From a philosophical
standpoint, there has been some discussion about whether SWAP should be a new
protocol or an HTTP/WebDAV derivative.  Our view is that we can justify it
being a derivative, but there are both workflow & protocol developers that
come down on both sides of the issue.  

One other thing that needs to happen with SWAP is that someone needs to go 
and think through all the return and status codes.  This could be done simply
by looking at the HTTP/1.1 and WebDAV status codes and then figuring out
what codes would be relevant or required for workflow-like things.  As you know,
exceptions in workflow happen quite regularly, so SWAP needs to be fairly robust
in that area, i.e. someone has to go through and distinguish protocol type exceptions
from workflow type exceptions and map them to meaningful status codes.  

Finally, the WfMC interfaces are a great source for 
culling requirements.  The strategy I take is, look at their examples and figure
out how exactly the information has to be put on the wire.   One of the problems
with using Web protocols for event notification is that there is no event 
backchannel to the original callee.  One of the tricks we've been using is 
just registering

Surendra Reddy has also recently presented some work on the subject, so you may 
consider sending him email at skreddy@us.oracle.com.  I've also CC'd the 
swap list to see if anyone else has any suggestions. 

Greg

Vijayanand Bharadwaj wrote:
> 
> dear Sir,
> I would like to introduce myself as a graduate student at West Virginia
> University(WVU), Morgantown WV.
> I am also a research assistant at the Concurrent Engineering Research Center
> (WVU) and am in the process of beginning  research work towards my Master's
> degree.
> 
> My work revolves around Workflow and the internet. I was introduced to  the
> University of California Irvine
> website and SWAP and i am very interested and wish to learn more. I have
> read the documents published by
> the SWAP  group which you have displayed on the website and also gone
> through the WFMC's publications.
> 
> I have also read the article that you have written along with Dr.Gail Kaiser
> "Leveraging the Web to manage Workflow"
> and this has provided me an inspiration and a basis for my research work.
> I would like to work with the SWAP and try and adopt it to a workflow
> scenario  using the internet and world wide web.
> My knowledge of XML and the  http ptrotocol is quite rudimentary and my
> efforts are in trying to be profecient in them
> 
> I would be very grateful if you could point me towards some resources that
> the SWAP group and others have been instrumental in creating which
> demonstate the use of the SWAP, especially some samples of code etc.
> which could help me understand the way in which one could use SWAP. I see
> that in your article you showed a few examples.
> 
> I would like to apologize for the inconvinience that is being caused by my
> request but your help and guidance would be very valuable in shedding some
> light on this issue and would
> greatly assist me in my endeavours.
> 
> I would like to with your permission ,continue this interaction and inform
> you of my efforts.
> 
> thanking you
> yours sincerely
> 
> Vijayanand Bharadwaj
> 
> 886 Chestnut Ridge Road,
> Concurrent Engg. Research Center
> West Virginia University
> Morgantown WV 26505

Received on Thursday, 5 August 1999 10:09:51 UTC