Re: 15 Ways to Think About Data Quality (Just for a Start)

On 4/13/11 11:41 AM, glenn mcdonald wrote:
> I don't think your detailed questions about Needle have any relevance 
> to this conversation about data quality, and I'm not sure they're of 
> much relevance to this mailing list at all, so I'm not going to follow 
> up on them any further in this context unless somebody else expresses 
> interest. As Hugh noted and a couple others endorsed, it should be 
> perfectly possible to discuss one system (like the LOD cloud) on its 
> own terms, without confusing contrasts to other systems with different 
> constraints, goals and natures... 
Glenn,

Of course you are 100% entitled your opinions just as I am mine. Here's 
my summary of this thread (via my context lenses of course):

1. You made reference to my views about data quality being subjective in 
the opening paragraph of the very post I urged you (repeatedly) to 
contribute
2. You repeatedly commented on data quality as a critical component that 
undermines my demos
3. You constructed (via interesting triangulation) a correlation between 
my demos and my views on data quality.

You have repeatedly tried to claim that I don't give a darn about data 
quality simply because:

1. You don't share my view on its inherent subjectivity
2. You believe my demos accentuate data quality issues, in a negative way.

All I've tried to convey to you (part of why I expend this much time and 
effort with you in these lengthy threads) is that my demos are 
fundamentally about:

1. Distinguishing Data, Information, and Knowledge
2. Accepting reality that cognitive beholders are wired to observe 
aspects of all observation subjects differently
3. Accentuating the inherent subjectivity that underlies cognition
4. Non existence of absolutes without inherently subjective context.

Bearing in my the 'world view' I've expressed above, all my demos are 
fundamentally about loosely coupling data, information, and knowledge in 
human and/or machine discernible form, courtesy of InterWeb scale Linked 
Data. My URLs are contemporary variants of DBMS views, with the in-built 
advantage of letting the beholder experience his/her/its inherent 
subjectivity by always providing conduits (LINKs) to the data sources 
behind the default projection. Worst case I and the beholder "agree to 
disagree" and when that happens a new View can be constructed from the 
accessible data sources and then shared with others. All of this is 
driven by URIs (be it in Name or Address functionality context).

When I made reference to ODBC, JDBC, OLE-DB, ADO.NET etc.. of yore, I 
was trying to convey to you that back in the day the following wasn't 
possible, in a platform agnostic manner:

1. Queries decoupled from backend DBMS
2. DBMS Views decoupled from DBMS engine
3. Reports decoupled from Data Sources (DBMS engines via ODBC, JDBC etc. 
Data Source Names) and Report Writer Products
4. etc..

Today, via a single LINK I can bury these myopic data flow impediments 
of yore.

When you say to me: Kingsley, your demo is broken because there is a 
questionable owl:sameAs relation, bearing in mind what I've outlined 
above, how are you expecting me to respond beyond: Okay, go make your 
own View and share it via a URL?

I am hoping, that you can at least accept an understanding of what I am 
trying to do, and at least factor that into your comments about my 
demonstrations. In addition, I don't spend my time telling "you" who 
"you" are, even if I disagree with "you". Thus, do try to reciprocate by 
not telling "me" about "myself", we can disagree without that dimension 
in the conversation.

-- 

Regards,

Kingsley Idehen 
President&  CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen

Received on Wednesday, 13 April 2011 16:12:59 UTC