> From: Dawn M. Wolthuis
> I'm looking for information related to companies that moved 
> or attempted to move from applications based on MV to Oracle 
> or other relational databases and went belly up in the process.


I don't know that this is so much a "MV-to-Other" conversion phenomenon as a general 
problem with any replacement of any legacy system.
 
New systems will be buggy. Period. There will be everything from requirement to design 
to programming errors.  Everything from fundamental misunderstandings and naove 
simplification of business rules, to the silliest of coding bugs.

Legacy systems have bug worked out.  Legacy systems have business rules buried in them.
A lot of those bug fixes and adaptations to new business demands make the code 
convoluted.  It looks like a building that started out as a one room hunting cabin, 
then by adding one room at a time, it ends up a 3-story 10-bedroom home, with attached 
2 car garage (The garage may be the original hunting shack).

Very often when converting to a new system, that system will have to go through that 
same cycle before it finally reaches the functionality of the system it replaces.   If 
the new system is built in-house, they can hope to make a coherent whole, but they 
_will_ miss things.  If there is continuity of personnel there is some hope.  But if 
they bring in new technologists for the new technology, forgetting that the old 
technologists possess invaluable business knowledge and the ability to read the old 
code and ferret things out (the legacy system's documentation is its source code), 
they are asking for disaster.

If a company buys supposed state-of-the-art software to run its core business 
function, replacing a highly customized system that encapsulates whatever it is that 
differentiates that company from its competitors, then G-d help them! (That's a 
prayer; I am not swearing.)

Read "Things You Should Never Do, Part I", by Joel Spolsky, 
http://www.joelonsoftware.com/articles/fog0000000069.html, for an example from 
Netscape.  Search for "rewrite" in that site's archives for articulate apologies for 
favoring old code.

I don't think anything I have said is peculiar to MV, but I do think much of what has 
been posted to this thread is consistent with it.   MV-based systems may have some 
additional virtues that imply additional costs when abandoned.  Comparing rates of 
success & failure of projects converting _TO_ as well as _FROM_ MV-based systems might 
say something about relative strengths/weaknesses of MV-technology. 

Chuck Stevenson
-------
u2-users mailing list
[EMAIL PROTECTED]
http://www.u2ug.org/listinfo/u2-users

Reply via email to