I guess at the moment Jennifer is actively doing some implementation
work on the Hospitality and Expenses project.
I wonder if it is worth it for interested parties to get together for
a brainstorm on long-term architecture for the VG systems? In
particular, I am not sure that the best approach is to design lots of
small independent systems for parsing different sites, each doing a
direct load from a source site and directly inputting the data into a
database.
In the medium term I see at least the following issues as requiring
attention:
(1) because we are fundamentally dealing with services and formats
that we don't control I believe we need to carefully design our
systems to separate the data input from all of the other aspects;
(2) because there will be many cross-cutting concerns affecting all of
the eventual services VG sponsors (i.e. scheduling, data conversion,
syndication, annotation, tagging, visualization, discussion, etc.) we
will reduce our "time-to-market" for individual services and stretch
our donor dollars further if we invest in establishing shared
implementations of these concerns.
Of course it might still be expedient to do the simplest thing
possible for the pilot projects (in order to attract attention and
funding) but having a roadmap is the first step in a longer
development process.
I will be happy to host such a session at my offices on St Laurent at
Mont-Royal.
Cheers,
Erik
For those of use in other jurisdictions (Ottawa, in my case), would
the results from this this be posted anywhere?
Don/
--
karfai [AT] gmail.com
http://www.strangeware.ca
http://blog.strangeware.ca