Re: My Migrations


Lee Unterreiner <lee@xxxxxxxxxxxxxxxxx> wrote in message
news:LIadnU0MEIKYQ3zVnZ2dnUVZ_ovinZ2d@xxxxxxxxxxxxxx
Fellow Dinosuars:

It's been a while.

I've been coding in C# for the last six years and I must admit I have
swallowed all of the OO Kool-Aid.

Note: (When I say 'I', I really mean 'We'. My friend and co-developer is
Rick Hardman. When I use 'we' I'm referring to y'all).

I used to say all my dreams were in COBOL. Now I have a new dream that
some
in this group may share.

In this dream I work for a company called "MyMigrations". Their mission
is
to move EVERYONE to .Net. Their product is called "MyMigrator". It is
to
migrations like MySql is to databases.

MyMigrator sucks in a client's entire software inventory: code, data,
schemas, procs, screens, jcl, etc., etc..
It reads from any IBM, any UNIX, any HP, any Windows, but only one target
platform: .Net.
Development and deployment platform based on SOA and Smart Client
architectures using WCF and Framework 3.5
All data converted to SQL Server. All screens or Accept/Display converted
to Winforms
All source languages (all dialects) are translated to C#. The generated
code is indistinguishable from the code you or I would write.



Oooooh, ambitious! Noble, too. But I don't think one person's lifetime is
sufficient to convert EVERYTHING. You must have some starting point. (F'r
instance, the Prime computers are long since gone, but there is a
considerable inventory of the Prime InfoBasic programs running all over the
place: I know of seven such sites in Winnipeg alone. They all run in an
emulated environment on bread-and-butter Unix or Windows systems. Bet you
hadn't considered them!).

More to the point: all the target facilities are Microsoft products (I
think - correct me if I'm wrong). Is it really a good diea to convert all
the world to a proprietary product? Particularly Microsoft! For all the
virtues of their products, they owe their dominance to superior marketing
and the bandwagon effect, rather than to technical superioirty. And I DO
believe there have been problems reported with at least some of their
products. I don't need to spell out all the disadvantages of a monopoly.

Nothing lasts forever. C# and .NET will someday be superceded. Not that
that matters - presumably Microsoft will provide migration tools - but WHAT
IF the next best paradigm comes from IBM or Univac or even from some outfit
that doesn't exist yet?

Would it be possible to make it a two-step conversion: first step creates an
agnostic, platform-independent, language-independent "specification" member;
second step would convert that to C#/.NET etc. If it's possible, as you
seem to have assumed or proved, that every facility in every language
environment on every systyem can be converted to C# etc., then it's
possible to create a set of specs that could be used in any language
environment to create the necessary programs & forms, etc. That way you
aren't tied to a particular platform and are at least relatively immune to
paradigm shifts. Maintenance of the specification members, for corrections
& upgrades, would then be the same as for any other language.

FWIW
PL


.

Quantcast