juddi-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andreas Wombacher <andreas.wombac...@ipsi.fraunhofer.de>
Subject Re: uddi Browser
Date Thu, 25 Nov 2004 09:05:57 GMT
Harvey,

thanks for your comments. Yes, I plan to continue the work on this
application and really appritiate comments on the application or
requests on extending the application.
see detailed comments below

Harvey Jiang wrote:

> Andreas,
> 
> Many thanks for your help. 
> 
> Now your Cocoon web application works for me under both the server provided
> with the distribution and tomcat.  I like this application! 
> 
> I would like to make a few comments below:
> 
> -Is it possible to have a centralised file to store juddi information such
> as publish URL, inquiry URL, publisher name, etc as default? There are two
> files: index.html and bs_searchresult.xsl to store the juddi information
Yes that is possible and seems to be a good idea. However, storing
passwords in a file in an unencrypted way seems not to be such a good
idea. Another required extension of the current application is to remove
the hard coded user name and password from different places in the code
and keep it in the session information after entering it once.

> -Is it easy to reuse the juddi information in the session? What I mean is
> that when you click "Specify UDDI", the default values are always displayed
> rather than one modified by the user.
Sorry, I do not get the point. The "Specify UDDI" was intented to allow
connecting the GUI to all kinds of different UDDI repositories. In
particular, the original aim has been not to be limited to jUDDI. This
aim has been nearly reached except for the hard coded user name and
password in the different scripts right now. That's the reason why the
values can be changed. Maybe you can try to restate your request. Thanks
a lot.

> -Are you going to enhance this application to include the delete functions?
Deletions are really difficult on such dependent data structures as used
in UDDI. The problem is to keep the data base (UDDI repository)
consistent with the delete operations. In particular, a business (in my
opinion) should only be removed if all services, assertions, etc. have
already been removed, because otherwise there exist data structures
pointing to a business record which has been deleted already. To
guarantee this behaviour via the API is quite complicated and requires
quite some time. Therefore I decided so far not to support deletions of
data. But I am open to discuss this issue and willing to realize it if
the concerns have been discussed.

> 
> Again, thank your very much for your help!
> 
> Steve,
> 
> Are you going to import the application into the juddi cvs repository? When?
> 
> Regards, Harvey


Best regards
Andreas


Mime
View raw message