- From: Steven Adler <adler1@us.ibm.com>
- Date: Tue, 7 Jul 2015 15:24:14 -0400
- To: Laufer <laufer@globo.com>
- Cc: Data on the Web Best Practices Working Group <public-dwbp-wg@w3.org>
- Message-ID: <OF5696B2D1.C55397D3-ON85257E7B.0069C4CC-85257E7B.006A96EF@us.ibm.com>
Hi Laufer,
Thanks for this lucid explanation. I really understand you now. Let me
try to summarize what I think I am reading:
1. "A Dataset" is a theoretical construct, an abstract thing, like a
corporation, an entity that may or may not be comprised of organic life
forms. But certainly it is a collection of other things - documents,
locations, concepts, computers, networks, chairs, desks, maybe people.
2. "The Dataset" is a real thing, an entity comprised of data (columns and
rows and/or bits and bytes).
3. "The Dataset" may be an instance of "A Dataset" - the real thing is an
example of the theoretical construct.
4. "The Dataset" may be distributed to other people, computers, and even
to abstract things like corporations.
Do I have that right?
Best Regards,
Steve
Motto: "Do First, Think, Do it Again"
|------------>
| From: |
|------------>
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|Laufer <laufer@globo.com> |
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| To: |
|------------>
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|Steven Adler/Somers/IBM@IBMUS |
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Cc: |
|------------>
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|Data on the Web Best Practices Working Group <public-dwbp-wg@w3.org> |
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Date: |
|------------>
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|07/07/2015 02:58 PM |
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|------------>
| Subject: |
|------------>
>--------------------------------------------------------------------------------------------------------------------------------------------------|
|Re: reviewing the BP doc |
>--------------------------------------------------------------------------------------------------------------------------------------------------|
Hi, Steve,
As I commented before, the words abstract and instance have different uses.
And I think is better to avoid its use without a clear definition. For
example, when we are talking about Object Oriented things, an instance has
a specific meaning.
I think that the fact that the access to data from a Dataset is made
through its distributions, this does not mean that a specific Dataset is an
abstract thing. An "Organization", for example, IBM, is not an abstract
thing. But I cannot reach IBM (I don't know if I am expressing my thoughts
in a good way). I think "Organization" is an abstract thing and IBM is an
instance of "Organization". In that way, "Dataset" (the concept, the class)
is an abstract thing. But a specific Dataset, an instance of the abstract
concept "Dataset", is not.
But people use the word instance for other things. In the same way,
abstract. The data that people can access, the data they can touch is what
DCAT calls a distribution. Is this an instance of the Dataset? In terms of
the informal use of the word instance, I think the answer is yes. But (IMO)
is not an instance when we think about OO.
If we have an API where someone could ask for a collection from a Dataset
using a query, each of the answers will be a subset of the Dataset. Could
we say that these subsets are instances of the Dataset?
The audience of our BP document may have technical different levels (I
guess), so I think that if we could avoid this confusion, it will be
better.
Best Regards,
Laufer
2015-07-07 14:42 GMT-03:00 Steven Adler <adler1@us.ibm.com>:
Laufer,
Good discussion. Are you saying that there are instances of Datasets
that get distributed when someone downloads them?
Dataset
Best Regards,
Steve
Motto: "Do First, Think, Do it Again"
Inactive hide details for Laufer ---06/30/2015 08:37:25 AM---Makx, I am
really confused now.Laufer ---06/30/2015 08:37:25 AM---Makx, I am really
confused now.
From: Laufer <laufer@globo.com>
To: Makx Dekkers <mail@makxdekkers.com>
Cc: Data on the Web Best Practices Working Group <
public-dwbp-wg@w3.org>
Date: 06/30/2015 08:37 AM
Subje Re: reviewing the BP doc
ct:
Makx,
I am really confused now.
In what moment of this discussion I proposed to change this model?
My participation in this post was motivated exactly because I have
identified a proposal to change that model and I expressed my opinion to
contribute.
And, again, the word abstract was not introduced by me in this post. And
this word was the main reason for my opinion. I commented to take care
using the words abstract and instance.
I also agree that the group don't have to change the DCAT model (we don't
have time to that), but I don't see any problem If someone in the group
identifiy things missing in that model, and we discuss and even decide to
insert a note in our document about this.
But as all the things in the group, is a matter of proposals and votings.
Laufer
Em terça-feira, 30 de junho de 2015, Makx Dekkers <mail@makxdekkers.com>
escreveu:
Laufer,
Let’s not use words like ‘abstract things’ because we might have
different ideas what that phrase means.
What is clear is that dcat:Dataset and dcat:Distribution are
classes in the DCAT model. DCAT defines what they are and how they
are related. DCAT is also clear about how those classes relate to
the physical data files or to the endpoints that give access to the
actual data.
I think we should restrict the discussion to that model. If not, we
might end up developing a different model, and I am not sure that
this group really wants to go there.
Makx.
From: Laufer [mailto:laufer@globo.com]
Sent: 29 June 2015 22:30
To: Makx Dekkers
Cc: Data on the Web Best Practices Working Group
Subject: Re: reviewing the BP doc
Ok, Makx.
I know this DCAT diagram. I am comfortable with this. And in this
model, both Datasets and Distributions are not abstract things.
What it is not comfortable to me is to consider that a Dataset is
an abstract thing.
Laufer
2015-06-29 16:04 GMT-03:00 Makx Dekkers <mail@makxdekkers.com>:
Laufer,
Ø I think we have to be carefull about using the words
abstract and instance.
Agree. Let’s not use those words.
Ø From the discussions, it seems to me that the Dataset is
an abstract thing with instances that are the distributions.
Ø This is what I have understood from the posts from
Bernadette and from you. And (until now) I do not agree with
this.
This is *not* what I have argued. Please look at the diagram
and examples in section 4 of DCAT
http://www.w3.org/TR/vocab-dcat/#vocabulary-overview. That
section gives an overview of the modelling approach of DCAT.
I would agree that there are many other ways you could model
this space, but DCAT is just what it is.
Makx.
--
. . . .. . .
. . . ..
. .. .
--
. . . .. . .
. . . ..
. .. .
--
. . . .. . .
. . . ..
. .. .
Attachments
- image/gif attachment: graycol.gif
- image/gif attachment: ecblank.gif
Received on Tuesday, 7 July 2015 19:25:06 UTC