Re: [Forces-protocol] Data Encoding

Weiming/Joel,

On Wed, 2004-11-17 at 09:48, Weiming Wang wrote:

> [Weiming]My thought on this is, we do not have to recognize if it is a KEY or
> not. It's just a kind of element in a path.

Ok, then we need a path TLV. And as Joel has been pushing for eons now,
separate the path from data; but allow recursive relationship.

> > Lets defer your idea of decomposing the ID above for now. I think we can
> > still achieve the goals without it.
> [Weiming]By separate the fieldIDs from Indexes, as Joel just proposed? I just
> have a worry on this expression. Pls see my reply to Joel.

I am not quiet sure i have parsed what you and Joel are talking about.
My approach may even be a third one.

> > Of course we could make KEY_DATA a TLV so we can have variable sized
> > DATA for a KEY, eg to repeat above:
> >
> > Path=4
> > T=KEY_INFO, length = 4 + xx
> >    T=1,  Length= 4 + 4 // 1 is ID for j1
> >    V = 100
> >    T=2, L = 4 + 4
> >    V = 200
> >
> > The above wastes 4 bytes per value because of TLV and is clearly more
> > complex to parse, but we could say thats the cost of doing this.
> [Weiming]This still applies to where KEY is with one ID, not with multiple level
> IDs as 1.2.
> 

To solve that we need something that allows variable path (in addition
to variable data) like:

T=PATH, Length = ZZ
  flags(16bit)=0 IDcount(16bit)=1, ID=4
  T=KEY_INFO, Length = xx
    T=PATH, Length = 4+4
      flags(16bit)=0 IDcount(16bit)=1, ID=1 //j1
    T=DATA, Length = 4+4
      V=100
    T=PATH, Length = 4+4
      flags(16bit)=0 IDcount(16bit)=1, ID=2 //j2
    T=DATA, Length = 4+4
      V=100

Essentially, KEY_INFO contains some content/data as well and reuses
PATH and DATA TLVs. So there could be a nested PATH/DATA pair.
Like i said this maybe the third view (with other two being what you and
Joel are saying)

Essentially grammar becomes:

OPER_TLV : = <PATH-DATA>
PATH-DATA := <PATH_TLV> [PATH_EXTENSIONS] [DATA]
PATH_TLV := flags IDcount <IDs>
PATH_EXTENSIONS := BLOCK_NOTATION | KEY_INFO
KEY_INFO := PATH-DATA
BLOCK_NOTATION :=   BLOCK_RANGE_INDEX_TLV | 
                    BLOCK_COUNT_INDEX_TLV |                
                    BLOCK_LIST_INDEX TLV

We dont distinguish between indices and IDs

I havent thought clearly about this but the recursive nature may
introduce loops. 

Thoughts?

cheers,
jamal

_______________________________________________
Forces-protocol mailing list
Forces-protocol@ietf.org
https://www1.ietf.org/mailman/listinfo/forces-protocol