Re: [soa-rm] Metadata

[Ken Laskey <klaskey@mitre.org>]

After an extended period on my day job, it is difficult to know where  
to enter this conversation.  Here are thoughts at random.

1. The focus (which I think I prefer to scope) of the RM is software  
but many of the concepts are equally applicable to non-software  
systems.  Our scoping is not that we will only talk about software  
but that our end product does not have to go out of its way to be  
complete for non-software domains.  The electric utility example  
seems to make a useful point but we don't have to tweak every aspect  
of the RM to faithfully account for such an example's details.  On  
the other hand, we should not avoid useful examples because they  
don't have a strict software interpretation.  (There was a book  
example in Appendix C of the 07 draft that demonstrated the idea that  
there is no one true set of metadata.  I might like to see that  
resurrected.)

2. Who or what processes metadata depends on whether you mean  
tomorrow or in 20 years.  Too little of today's metadata is truly  
machine-processible but much of the purported power of SOA is  
described in scenarios where more automation is assumed than may  
currently be realistic.  (I just got done writing 25 pages on the  
discovery tangent which you will be thankful I will not get into  
here.) We need a way of not restricting metadata to that which is  
solely machine-processible while noting that such is the expected and  
anticipated direction.

3. The RM is conceptual but those concepts must be related to the  
real world in which our clients and sponsors live, otherwise our fine  
work will lack relevance and be ignored.  There are expectations of  
things we will talk about when talking about an SOA, and we must  
either relate to those things or explicitly say why those are  
subsumed in the more general concepts and don't need to be explicitly  
discussed.

4. There was one thing that started at the last editors' ftf and was  
floating on the editors' list and I thought would make it into the  
document but didn't was the following example of RM vs. RA. vs.  
architecture (A).  With apologies to Frank who wasn't sure he liked  
it, here is the example:

> Go back to our house analogy.  The RA captures concepts related to  
> what makes up a house, e.g. room, window, door.  It might include  
> the concepts of food preparation area and personal hygiene area and  
> the relationship that there should be physical separation between  
> the two.  Note that this provides a very North American/western  
> Europe reference and not necessarily one that covers a tent.  So a  
> given RM already provides a perspective.
>
> Given RA concepts, various RAs show how these concepts can be  
> arranged in a useful pattern.  So RA examples would be (sorry for  
> the American terms) a colonial, a split-level, a rambler, etc.  You  
> can play with the pattern but one can say that any given pattern  
> serves a particular set of purposes (e.g. a rambler is on one level  
> for those who want/need to avoid stairs).
>
> An architecture is then a specific plan to build a house or set of  
> houses.  There can still be some variations but you don't do things  
> like moving fireplaces or structural walls, else you have a new  
> architecture.

5. I haven't paid significant attention to our write-up of  
conformance but I think we need to have a summary section that says  
this, this, and this make up SOA and your system looks like that or  
it's something different.  I expect that laundry list will still have  
enough loopholes to allow for considerable creativity.  So I am not  
sure how RFC 2119 will apply; I know I didn't write with 2119 any  
more in mind than would unfortunately now be professional reflex.

Enough for now.

Ken


On Sep 28, 2005, at 6:55 PM, Duane Nickull wrote:

>
>
>>> ... the decisions
>>> based on the metadata are made by humans.
>>>
>
> This is not always true.
>
> DN: Agree.  Not always however decisions made by software are the  
> result
> of humans writing and deploying the software => indirectly, humans  
> have
> made the decisions.
>
> "It appears to imply that descriptions are
> meant for human consumption."
>
> DN: Which would not be the intent.  I am not good at the wordsmithing
> but I think we agree.
>
>
> Frank (with 9 fingers)
>
> Okay - I'll bite.  Why only 9 fingers (inferring either + or - one  
> digit
> based on how you semantically classify thumbs)?
>
> D
>
>