Re: [Suggestion]model naming convention

I Can certainly live with a requirement that 
non-IETF names for ForCES elements use specific defined prefix.
As I said, just tell me what it should 
be  (preferably, with more than one person liking the choice.)
VND-
EXT-
PRV-
?
(Just to be clear, I appreciate Tom's clear note 
on the subject.  But someone has to pick the 
prefix.  Historically, if I pick a prefix everyone will hate it.)

Also, Tom has pointed out that I left out something.
The named data types also need to be 
unique.  They don't have numeric identifiers, but 
they do have names that are used to refer to 
them.  So I need to add that to the IANA considerations.

Obviously, the prefix syntax neeeds to match 
whatever convention, if any, we adopt for naming these things.

Yours,
Joel Halpern

At 06:06 AM 10/18/2006, tom.petch wrote:
>Sorry for being slow to reply.
>
>My main concern is with LFB Class, rather than a 
>lower level of detail, and that
>is largely because that is the one that is 
>called out in IANA Considerations and
>so given particular treatment.
>
>I agree with the space allocated for numeric class identifiers but see a need
>for more precision with the names.  The 32 bit number did leave me wondering
>just how many of these were expected and am reassured by the view that
>standards-based would be O(100); I would expect vendor specific to be greater,
>say O(1000).
>
>So, I would then want an obvious way of telling a standards Class name from
>vendor specific, and like the idea of doing it on the basis of the first
>characters.  X- for vendor-specific is a bit 
>ugly but is widely used.  But since
>the number of standards-based ones is quite small, it might be simpler to
>reserve an initial letter(s) for these, such as F or R or S (or f or r or s)
>
>A secondary consideration is to identify a Class name from any other
>sort of name but am less clear how much of a 
>need there is for this ie should I
>usually
>be able to tell that a name is a Class name from 
>its context; given the somewhat
>loose syntax of XML/XSD (/Xpath/X...), may be not in which case, a further
>constraint on other names would be helpful.
>
>Beyond that, I like to keep name spaces simple and so would not impose further
>constraints.
>.
>Tom Petch
>
>----- Original Message -----
>From: "Jia Fenggen" <[email protected]>
>To: <[email protected]>
>Sent: Monday, October 16, 2006 2:50 PM
>Subject: [Suggestion]model naming convention
>
>
> > In the current model,an LFB is written without an clearly naming
> > conventions,the LFB name and element name are always begin with uppercase
> > an all kinds of  datatypedefs also begin with uppercase,should we define an
> > naming convention to formal the naming of various names.I suggest
> > following:
> > 1 All data types defined using datatypedef by user should have a name begin
> > with uppercase character,inside an complex data types,each element name
> > should begin with lowercase character;
> > 2 The attributes capabilities and events element names of should begin with
> > lowercase character.
> > an general rule is that if the element is to be referenced by others,it
> > should begin with uppercase characters,else lowercase is used.An i suppose
> > the user don't directly use the simple types,he should define it's own
> > types before use them.
> > that maybe not complete,what do you think?
> > Yours,Fenggen
> >
> > _________________________________________________________________
> > ÏíÓÃÊÀ½çÉÏ×î´óµÄµç×ÓÓʼþϵͳ¡ª MSN Hotmail¡£  http://www.hotmail.com