Re: [sdd] Alternate definition of artifact

[Julia McCarthy <julia@us.ibm.com>]

Hi Fukui-san,

Your arguments for the usefulness of the term payload also have helped convince me that we need the term. Are you happy with the definition that James and I were discussing?

Payload:

Binary packaged content intended to be delivered, without change, through a solution
deployment.

As for artifact, I think the general definition you reference in your note, does apply in this context. Artifacts are the "objects" that collect together payload and meta-data required to perform a single lifecycle operation on a single resource. I think they are central to any discussion of the SDD. The SDD will contain meta-data and payload for deploying all the resources involved with a solution. The artifacts collect together the information from within the SDD for a single operation on a single resource.

Are you happy with the artifact definition as proposed?

Artifact:

Zero or more files or meta-data used to perform a lifecycle
operation on a resource.



Julia McCarthy
Tivoli Development
Deployment Engine Design
julia@us.ibm.com
349/8156
877-261-0391



Keisuke Fukui <kfukui@labs.fujitsu.com>

Keisuke Fukui <kfukui@labs.fujitsu.com>

04/26/2006 05:57 AM

To	James Falkner <james.falkner@sun.com>
cc	Julia McCarthy/Raleigh/IBM@IBMUS, sdd@lists.oasis-open.org
Subject	Re: [sdd] Alternate definition of artifact

Hi there,

I agree with the below part of the James's comment.

James Falkner wrote:
>
> In that case, the value of defining payload is in allowing payload to be
> handled more efficiently by runtimes, if they know that it is never to
> be changed or parsed.
>
> -jhf-

Here is my opinion regarding to this:
If a file is not parsed or interpreted in the deployment process,
it will be probably used by software itself in runtime. SDD should
consistently handle it as an opaque object in its context. These
may be called payload for the deployment package.

If a file is parsed or interpreted in the deployment process, it
needs to be specified in SDD specification. The data in the file
may be called metadata since the software won't needs them runtime.

While some metadata can describe property (for example, file names,
size, digest value ) of payload files, the metadata should be
kept in separate files from the payload. (payload is not modified.)


I think the critical need for distinction is whether it is
parsed or interpreted in the deployment process. If it is true,
what we need to define the term are the above two concepts. I'm
fine with "payload files", but not sure about files containing
metadata.


To tell the truth, it's not very clear to me why the word artifact
is used everywhere in our discussion. (this may be language problem
of me:-)

In my reference, the artifact is described as:
tool; object; man made object (often referring to primitive tools)

I think that's true for everything we are discussing. We are not
discussing about natural material.

 -Keisuke

P.S.
I'm sorry if I missed the point. A non-native speaker comment:-)