Re: [model] section 3.2.4 Metadata

On Wed, 2006-27-12 at 15:33 -0500, Joel M. Halpern wrote:
> I don't think what you propose will work.
> To be more specific, I can not udnerstand how I would act on that if 
> it were agreed to.
> 

The text needs to be more relevant. 
[This does not imply the description of what metadata is irrelevant;
there is just too much of it and it just gets into the way of making the
relevant point].

More than anything the text goes into a lot of verbosity. Example:
-It talks about metadata that is visible within a "chip" which it refers
to as "internal". What is a  "chip" in Forces context? Is it an FE, a
LFB, a collection of LFBs? After weeding through all that text, you find
out that after all Forces doesnt deal with "internal" metadata. 
- Also, i am not sure i understood the relevance of "implicit" metadata.
The example used didnt seem to make a lot of sense to me.
- The text also repeats similar information in more than one location. 
Example 3.2.4.2 para 3 is repeated again in 3.2.4.4 second to last
paragraph, etc.
- I could go on ..
If i was writing a tutorial for some hardware person where some of the
terms used are common, I would have seen the necessity to explain things
that way. However, if i am dealing with Forces, then it is simpler to
assume that we are talking about metadata related to LFBs.
Define that it means a 32 bit tag with a value and explain details
around that construct.

> I am not going to claim that section 3.2.4 is particularly 
> readable.  It is hard to read.

Ok, at least we agree on that. 

> But unless we define what we mean by meta-data, and how it can be 
> used, it is meaningless to define how the model works with 
> meta-data.  It won't help anyone.
>
> The basic problem is that different people mean different things when 
> they say "meta-data."  

Agreed, 
I think you need to define what metadata is in a paragraph or two -
theres plenty of literature that exists that defines it using more text,
you just need to reference that. I think you need to have more text on
how we use it (than on defining it). Those two items should not exceed a
page.

> For the ForCES work we mean a particular kind 
> / set of meta-data, used in a particular fashion.
> If we leave out the explanation, people will think we mean all sorts 
> of other things.
> 

I would start with wikipedia (http://en.wikipedia.org/wiki/Metadata) 
It does good job and it has many references. 
On the convergence of what it means in Forces, you could narrow it down
to a page and get to the point faster.
I see relevance in describing that metadata within forces as:

a) definition
b) usage and lifecycle
b-i) constraints within forces
b-ii) defined operations 

Do you want me to take a crack at reducing the text? 

> No, the schema does not capture most of "what is meta-data".  The 
> schema does not capture all the aspects of the LFB definitions.  It 
> can't.  We did not, for example, include a programming language in 
> which to formally specify the actions of the LFB class.

Looking at the metadata operations in particular: Assuming they are
going to be used in some LFB(s), isnt it important that two different
people defining two LFBs that may be used together within a graph
converge on some representation of "READ" and "WRITE"
Are you saying, perhaps, that metadata operations are local to an LFB
and that they can define them as they see fit?

The other angle:
Since we have defined metadata to comprise a 32 bit tag with a value;
in order for those two mythical LFB writers described above to interop,
would it not make sense to have both the tags and the values globaly
defined and hence IANA regulated?

> There are two reasons we describe meta-data:
> 1) For both the human being and the CE, it is important to know how 
> the meta-data flows, so that it can be understood how operations relate.
> 2) Some LFBs have tables that either generate meta-data or are 
> indexed by meta-data.  It is important to have a way to clearly 
> specify which piece of meta-data is being generate / consumed.  (In 
> some cases, one even has tables to specify which piece of meta-data 
> is being manipulated.)
> So we need some significant discussion of meta-data.  (This is in 
> contrast to frame-type which is almost purely for human consumption.)
> 

Well, I think we are starting to. 

cheers,
jamal