Re: [ontolog-forum] Amazon vs. IBM: Big Blue meets match in battle for t

[Doug McDavid <dougmcdavid@xxxxxxxxx>]

I may have missed mention of this, but in addition to all the meta-data and code discussed in this thread, one of the best sources of fossilized meaning can be found in the actual values stored in tables, files, dbmses.  I have done this work in the past (reverse-engineering insurance systems in mid-'80s, e.g.), but my rate for such work today would be prohibitive :-)

On Sat, Jul 27, 2013 at 8:05 AM, Paul Tyson <phtyson@xxxxxxxxxxxxx> wrote:

On Sat, 2013-07-27 at 10:06 -0400, deddy wrote:
> Michael -
>
> >
> > your example becomes more and more specific and challenging :-) Yes, if all
> > you have is the code, you are in trouble and SW technologies are not a magic
> > bullet for solving it.
> >
>
> Welcome to the world of legacy systems.
>
> I hope against hope that somewhere in the SW stack of tools there just might be something to help with
> understanding legacy systems.

It's all there now. See, for your purposes, RDF, RDB2RDF, SKOS, and
Common Logic.

It sounds like you are dealing with hierarchical data (IMS) and COBOL
structure layouts, so RDB2RDF might not be directly applicable, but you
can perhaps get ideas for mapping features IMS and COBOL data layouts to
URIs. I've seen implementations where IMS data is rehosted in RDBMS, but
I don't know if there are any conventions around that.

>
> It's a very SMALL hope.

It's a real and current opportunity.

>
> What I do see is SW creating yet another tangled layer of undocumented,
>  poorly understood systems.
>

This would be a consequence of poor design or implementation of a
particular system, not anything inherent in the underlying standards.

>
> I've been attending monthly MIT SW meetings for 4+ years.  Once, by
>  chance, I did catch TBL himself saying to the audience that "Semantic
>  Web" was a clever marketing label, but that in reality, "linked data"
>  would be a more appropriate description since there really isn't
>  anything special about semantics in the SW.
>

I prefer "linked data" as well. Semantics, strictly speaking, is all
wetware.

>
> As you may have noticed, my passion is for a process—tool assisted, but
>  requiring human knowledge at the beginnings —to extract & make formal
>  the MENSA_FL --> MEssage Notify Stop Action Flag --> "a collection of
>  dunning flags" process.

I have built part of such a system, using above named standards. I am
not so interested in the automatic discovery of correlated fields--but I
have found that just having everything out in RDF lets you add the
appropriate links (via rules or RDF statements) once you discover them,
and start to provide business value in unexpected ways. You can also
enrich your RDF repository with documentation directly tied to the
objects and fields of interest, so you and your successors only have to
do the analysis once.

The sad thing is, my system was built not primarily to adapt legacy
systems to modern needs, but to fill in the gaps between 3 brand
spanking new enterprise systems costing 100s of millions of US$. It can
of course be extended to any legacy data still of interest to the
business.

Regards,
--Paul

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
 

--

Skype: dougmcdavid

Mobile: 916-549-4600
Second Life: Doug McDavid on Artropolis
Web: enterprisology.com

_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)