atlas-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mandy Chessell <mandy_chess...@uk.ibm.com>
Subject Re: Rename trait to classification
Date Wed, 28 Sep 2016 14:12:12 GMT
Hello Suma,
I like that idea.   How would glossary terms fit into this scheme?  Are 
they also a subclass of annotation?  We link terms to assets to indicate 
the semantic meaning of a data item.  However, we may also link a 
classification to a term to say any data item linked to this term is also 
linked to this classification.  This pattern is used by organizations that 
want consistent treatment of data across all data sets/repositories.  It 
is particularly useful when supporting a specific regulation that requires 
a particular type of data to be governed in a particular way.

This picture shows an example of a classification of "Sensitive" linked to 
a glossary term "Compensation Plan".  There are two specialisations of 
compensation plan in the glossary called "Annual Salary" and "Hourly Pay 
Rate".  These glossary terms are then linked to the schema of the data to 
identify that 45324 is a salary that should be considered sensitive.



All the best
Mandy
___________________________________________
Mandy Chessell CBE FREng CEng FBCS
IBM Distinguished Engineer
IBM Analytics Group CTO Office

Master Inventor
Member of the IBM Academy of Technology
Visiting Professor, Department of Computer Science, University of 
Sheffield

Email: mandy_chessell@uk.ibm.com
LinkedIn: http://www.linkedin.com/pub/mandy-chessell/22/897/a49

Assistant: Janet Brooks - jsbrooks12@uk.ibm.com



From:   Suma Shivaprasad <sumasai.shivaprasad@gmail.com>
To:     dev@atlas.incubator.apache.org
Date:   27/09/2016 16:19
Subject:        Re: Rename trait to classification



Thanks Mandy for the detailed explanation. To add to what David has 
already
stated,

We could expose two higher level ATLAS types/concepts which are based on
current Traits.

1. Classification - that are locked down and that have a classification
scheme to group together a set of classification concepts
2. Annotation - that are free form and are not locked down

So Traits could be renamed as Annotation and Classification is a
specialization of an Annotation with an additional classification scheme
associated with it.

Thoughts?

Suma





On Mon, Sep 26, 2016 at 5:48 AM, David Radley <david_radley@uk.ibm.com>
wrote:

> Hi Hermanth,
> I appreciate your feedback and openness. It was an interesting point you
> made about which roles were authoring traits and terms. I guess this is
> not something Atlas would police.
>
> The current traits could be :
> 1) locked down so only the governance team could update them; in that 
case
> they would be classifications that governance rules could act on.
>  or
> 2) Not locked down so a wider audience (business personas) could create
> them.
>
> I am suggesting:
> - renaming traits to classifications for use by the governance team.
> - using terms as glossary terms for use more widely by business users.
>
> Does this work - or am I missing something ?
>       all the best, David.
>
>
>
>
>
>
> From:   Hemanth Yamijala <yhemanth@gmail.com>
> To:     dev@atlas.incubator.apache.org
> Date:   26/09/2016 13:09
> Subject:        Re: Rename trait to classification
>
>
>
> Hi David,
>
> Reg. the point I made about sharing traits - I don't want to give an
> impression that this as an agreed upon point. Apologize if I conveyed
> that sense.
>
> It is a fact that Atlas today has two concepts that are slightly
> related: Traits (aka Tags) and Business Terms. The latter was new in
> 0.7. IMO, it is important that the Atlas community tries to converge
> on an unambiguous definition of these concepts as the product would be
> driven around these.
>
> With respect to this thread, I am trying to fit in whether
> "classification" is a new concept. Or it overlaps with one of the two
> existing ones (which we are trying to rename).
>
> I am certainly not a domain expert on this in any sense :-) - so
> hoping that others who are would provide guidance (@aahn - ping?).
>
> Thanks
> hemanth
>
> On Mon, Sep 26, 2016 at 2:59 PM, David Radley <david_radley@uk.ibm.com>
> wrote:
> > Hi Hermanth and Mandy ,
> > Thanks for your feedback.
> >
> > It does seem like these are de-facto industry terms in the governance
> > industry; the reason I say this is that looking around the web I see
> quite
> > a few uses of the words governance classification in different domains
> > (including in the Atlas documentation!).
> >
> > I was not aware of the idea that traits and terms would be authored by
> > different roles - thanks for your explanation. What is coming up for 
me
> is
> > :
> >
> > I think business users should be able to add new business terms (maybe
> > going through a workflow and a governance curator then sorting out
> > inconsistencies), as they are the most expert as the language they 
use.
> > Classifications could be authored by different teams, for example 
levels
> > of confidentiality (in Mandy's example) would be dictated by the
> > governance team. Governance rules would run on these classifications.
> >
> > You say "So, it is hard to use traits in a shared sense or expect to
> have
> > conventional usage" . I notice the Atlas tutorial did not give me this
> > impression, as the example of a trait/tag is PII.
> > Your description of traits implies they are more like free form labels 
.
> > If this is the intent for traits, then it does not make sense to 
rename
> > them to classification. Maybe traits should be called labels; so their
> > name is more in line with their expected usage. Though we should 
change
> > the tutorial!
> >
> > A business term is a type of classification -a semantic 
classification.
> We
> > could add in the concept of classification which Business term and
> > Business category  (Jira 1186 ) inherit from. This would allow us to 
add
> > in confidential classifications and classifications schemes to 
organize.
> >
> > I look forwards to your thoughts,
> >       all the best, David.
> >
> >
> >
> >
> > From:   Hemanth Yamijala <hyamijala@hortonworks.com>
> > To:     David Radley <dev@atlas.incubator.apache.org>
> > Date:   26/09/2016 05:33
> > Subject:        Re: Rename trait to classification
> >
> >
> >
> > Hi,
> >
> > Are these de-facto industry terms in the governance industry? If yes,
> > would they make more sense to explore as part of the Business Taxonomy
> > feature that's currently in alpha in 0.7, rather than the trait 
system?
> >
> > One differentiation we've been trying to express is that traits (also
> > referred to as tags in some places in Atlas) are free form and left to
> the
> > user using them. So, it is hard to use traits in a shared sense or
> expect
> > to have conventional usage. So, traits would probably be a tool for a
> data
> > scientist to quickly annotate something for their own discovery usage
> > later.
> >
> > Business taxonomy, on the other hand, is something we are thinking as
> used
> > to express standard classification, even if only within an 
organization,
> > but maybe even across industry domains etc. They would likely be 
created
> > by data stewards with knowledge of the domain and their usage would
> follow
> > established practices (authorization controlling who can do what).
> >
> > Not sure if what we're referring to as "classification" here fits the
> > "traits" or "business taxonomy" side more - trying to understand...
> >
> > Thanks
> > hemanth
> > ________________________________________
> > From: Mandy Chessell <mandy_chessell@uk.ibm.com>
> > Sent: Sunday, September 25, 2016 9:56 PM
> > To: David Radley
> > Cc: dev@atlas.incubator.apache.org
> > Subject: Re: Rename trait to classification
> >
> > Hello David,
> > I also like the idea of using the term classification.
> > Typically classifications in governance are ordered sets of values
> grouped
> > into a classification scheme.  Is the notion of the classification
> scheme
> > also part of the change you are thinking of?
> >
> > For example, the classification scheme and "unclassified" value which 
is
> > the default classification for any data element that has no
> classification
> > from this scheme associated with it.  The other values are defined in
> > increasing levels of sensitivity.  There are also sub-classifications.
> So
> > for example, confidential has sub-classifications of Business
> > Confidential, Partner Confidential and Personal Confidential.  If a 
rule
> > is defined for "confidential", it applies to all three of the
> > sub-classifications.
> >
> > §Confidentiality Classification Scheme
> > §Confidentiality is used to classify the impact of disclosing
> information
> > to unauthorized individuals
> > •Unclassified
> > •Internal Use
> > •Confidential
> > •Business Confidential.
> > •Partner Confidential.
> > •Personal Information.
> > •Sensitive
> > •Sensitive Personal
> > •Sensitive Financial
> > •Sensitive Operational
> > •Restricted
> > •Restricted Financial
> > •Restricted Operational
> > •Trade Secret
> >
> >
> > The classification schemes create a graduated view of how sensitive 
data
> > is.  We would also expect to see classification schemes for other
> aspects
> > of governance such as retention, confidence (quality) and criticality.
> >
> >
> > All the best
> > Mandy
> > ___________________________________________
> > Mandy Chessell CBE FREng CEng FBCS
> > IBM Distinguished Engineer
> > IBM Analytics Group CTO Office
> >
> > Master Inventor
> > Member of the IBM Academy of Technology
> > Visiting Professor, Department of Computer Science, University of
> > Sheffield
> >
> > Email: mandy_chessell@uk.ibm.com
> > LinkedIn: http://www.linkedin.com/pub/mandy-chessell/22/897/a49
> >
> > Assistant: Janet Brooks - jsbrooks12@uk.ibm.com
> >
> >
> >
> > From:   David Radley/UK/IBM@IBMGB
> > To:     dev@atlas.incubator.apache.org
> > Date:   23/09/2016 17:05
> > Subject:        Re: Rename trait to classification
> >
> >
> >
> > Hi Madhan,
> > That would be great :-)  thanks, David.
> >
> >
> >
> > From:   Madhan Neethiraj <madhan@apache.org>
> > To:     "dev@atlas.incubator.apache.org"
> <dev@atlas.incubator.apache.org>
> > Date:   23/09/2016 16:48
> > Subject:        Re: Rename trait to classification
> > Sent by:        Madhan Neethiraj <mneethiraj@hortonworks.com>
> >
> >
> >
> > David,
> >
> > I agree on replacing ‘trait’ with ‘Classification’. I guess the name
> > ‘triat’ might have been influenced by Scala (and not from Ranger, 
which
> > doesn’t have ‘triat’ in its vocab..).
> >
> > Instead of renaming in the existing APIs, how about we go with the new
> > name in the API introduced in ATLAS-1171?
> >
> > Thanks,
> > Madhan
> >
> >
> >
> > On 9/23/16, 1:35 AM, "David Radley" <david_radley@uk.ibm.com> wrote:
> >
> >     Hi,
> >     I have raised Jira ATLAS-1187. This is to rename trait to
> > Classification.
> >     I know that this would effect the API, so am keen to understand 
how
> we
> >
> >
> >     agree to version the API maybe including other changes. I feel 
trait
> > is
> >     not very descriptive and I assume comes from Ranger terminology. I
> > think
> >     using classification instead brings us into using terminology 
better
> >     representing the Atlas capability and its role in governance use
> > cases. I
> >     am keen to get your feedback. I do not feel that I should just
> submit
> > a
> >     fix like this - I think we need more agreement to account for the
> > impact
> >     on current users. At the same time, we are still in incubation we
> > should
> >     be able to make changes like this to polish the API.
> >
> >     I am looking forward to your thoughts,       David Radley
> >     Unless stated otherwise above:
> >     IBM United Kingdom Limited - Registered in England and Wales with
> > number
> >     741598.
> >     Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire
> PO6
> >
> > 3AU
> >
> >
> >
> >
> >
> >
> > Unless stated otherwise above:
> > IBM United Kingdom Limited - Registered in England and Wales with 
number
> > 741598.
> > Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
> 3AU
> >
> >
> >
> >
> >
> >
> >
> > Unless stated otherwise above:
> > IBM United Kingdom Limited - Registered in England and Wales with 
number
> > 741598.
> > Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
> 3AU
> >
>
>
>
>
> Unless stated otherwise above:
> IBM United Kingdom Limited - Registered in England and Wales with number
> 741598.
> Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 
3AU
>
>



Mime
  • Unnamed multipart/mixed (inline, None, 0 bytes)
View raw message