Buck,
Picking up on your last point which is a very important one.
With reference to the data model and business rules extraction done by
X-Analysis, you mentioned that Pathfinder, Abstract, and Total/400 do
something a lot like that. That's true, but not entirely true. What those
products do (good products all of them) is help YOU find or derive what the
business rules and relational data model rules are. What newer versions of
X-Analysis do (the actual new modules are called X-Rules and X-Rev
respectively) is automatically and explicitly extract the business rules and
data model foreign key relationships FOR YOU. X-Analysis then presents these
designs to you in a number of different interactive visual formats, as well
as allowing you to download them for reuse as DDL, XML, Excel Spreadsheets,
PDF, MS Word, UML. So in explicitly isolating and documenting these two
types of design constructs on top of the stuff you have probably seen in
earlier or entry level versions of X-Analysis, it makes processes such
understanding, redesign, reengineering and rewrites much quicker and easier.
The business rules and data model can also be used programmatically to
modernize or reengineer an application or parts of it auto-magically. The
reason this is possible is because the business rules and data model are
explicitly derived and stored in a structured easily accessible way in the
repository, so programs can be written to use them. The alternative is to
use a more traditional approach with any of the products mentioned and do
the work manually thereafter. Examples of this at work today are warming up
IBM's DB2 Web Query with an applications entire meta-data layer derived by
X-Analysis, converting an applications database from DDS to SQL -
automatically with no recompiles, and even expanding a database field and
all its manifestations in programs throughout an entire system -
automatically. These aren't things one does every day, but when they do come
along they can cripple a company's development budgets and momentum.
Why I tend to get a little excited about this (and possibly a little smug -
apologies Terry and David) is that it goes way beyond basic but important
analysis and cross referencing. In a world where there is an ever
diminishing work force of RPG skills and a code base that is increasing,
even one or two developers can quickly and completely take intellectual
control over large systems, even if they haven't seen them before. The same
developers can take on HUGE reengineering or modernisation projects
themselves. See our case studies for real and recent evidence of this. This
naturally plays into the hands of off-shore companies being able to become
conversant very quickly with legacy applications, but it also plays equally
well to having someone local or at least onshore who can do the work of many
lower skilled offshore resources.
Here is an article on the subject of writing programs to change programs on
System i Network you might find interesting:
http://systeminetwork.com/article/application-mapping-ibm-i-writing-programs
-update-your-programs
Cheers.
Stuart Milligan
Tel: 917 267 7523
Toll Free: 800 605 5023 ext 81
Fax: 866 266 3165
web: www.databorough.com
midrange.com
