On 9/18/13 4:51 PM, John F Sowa wrote:
> Kingsley, David, and Melvin,
>> There are different kinds of silos. When it comes to data, I am
>> yet to find any justification for a silo.
> The distinction between modularity (good) and silos (bad) depends
> on who controls the access to what kind of data for what purpose.
>> Ora Lassila explains the kind of data-de-silo-fication I seek...
>>  http://bit.ly/T2aTNi -- loosely coupling applications and data
>>  http://slidesha.re/TBMT0Q -- size doesn't matter (if your data is in a
> I agree with many of the points that Ora L. makes, but I also have
> many "yes, but..." reactions to others.
> For an OS, I like the Unix design of having open formats for data and
> a collection of general utilities for processing them. (01)
Yes, so do I. (02)
> I prefer that
> to the idea of a dedicated program that controls each data type. (03)
> *BUT* there are good arguments for "object-oriented" languages and
> systems that provide an access-control program for each resource
> and a standardized set of "methods" for accessing it. (05)
Yes, which is where authentication protocols come into play, especially
when the protocol is capable of understanding the entity relationships
and relation semantics expressed through the data. In a nutshell, that's
where attribute based access controls come into play as one can verify
identity principals (using a variety of authentication protocols) and
test said principal against data access policies and acls. (06)
>>> I am yet to find any justification for a silo.
>> There's a long list of justifications...
> I agree that they are all important.
>>> - senior management can assign responsibility (single neck to wring)
>> No, not a justification in my eyes at all.
> I'm as suspicious of senior management as Dilbert is. But I recognize
> the need to have a single person who is responsible for each resource. (07)
Yes, and that's covered via attribute based access controls and the
fidelity of discernible relation semantics baked into the data. (08)
> Whenever there is divided responsibility for anything (in business,
> government, or just planning a party) everyone will let "the other guy"
> take care of it. The result is that nobody does it. *And* you also
> need somebody to check up on the responsible person. It doesn't have
> to be intrusive. An occasional question is enough: "How's it coming?" (09)
Yes, but these are issues that are covered by the fusion of structured
data, relation semantics, and identifiers etc.. (010)
>> Linux is really 3 projects IMHO...
> I agree. I just used it as an example because it was in the news.
> In other notes, I mentioned the SABRE system for airline reservations,
> which IBM developed for American Airlines in the early 1960s. That
> design with its implicit ontology is still embedded in the latest
> and greatest reservation systems for airlines and related services. (011)
Yes, but how do we migrate from that system without taking the
disastrous route of "rip and replace" i.e., attempting to move to a new
system wholesale? A shim over the current data would enable gradual
migration over time. The aforementioned shim could take the form of a
Linked Data view over the existing data. (012)
>> All 3 phases are works in progress ...
> Absolutely! All legacy systems are are works in progress until they
> are "functionally stabilized". That's IBM's euphemism for "dead end".
> The people who need them can use them, but they won't get any updates. (013)
I have no problem with that as a goal. Pragmatically though, the
unpejorative "legacy" espoused above isn't possible if the system data
exists in a silo that doesn't support contemporary data access
mechanisms, for a given point in time. (014)
> 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
Founder & CEO
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen (018)
Description: S/MIME Cryptographic Signature
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
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)