| Michael Wojcik 2006-04-28, 6:55 pm |
|
In article <4bc8avF10basaU1@individual.net>, "Frank Swarbrick" <Frank.Swarbrick@efirstbank.com> writes:
>
> I know you were being a bit facetious, but do you honestly think that
> converting, say, a batch transaction posting program to Java is a reasonable
> thing to do?
> And if not Java and not COBOL, then what?
I recommend a combination of OCaml and Ruby, though if Microsoft
comes out with APL.NET that could be a contender, too.
Of course, there's always Eta, which takes the "write it in
something that looks like English" model even a bit further than
COBOL.
In all seriousness: rewriting a program that reifies significant
business logic clearly constitutes substantial risk, so it's only
justifiable if there's substantial average return. The statistics
I've seen set the risk high enough and the average return low enough
to justify relatively few such rewritings. But it's not something
I've spent a lot of effort researching, and I have a vested interest
in COBOL (though we do sell to mixed-language and non-COBOL
environments as well).
See, Frank, if you *had* gone to the MF Dev Forum, you would have
seen my presentation which explained all this in terms of albatross
management and octopus-head masceration...
--
Michael Wojcik michael.wojcik@microfocus.com
Not the author (with K.Ravichandran and T.Rick Fletcher) of "Mode specific
chemistry of HS + N{_2}O(n,1,0) using stimulated Raman excitation".
|