Identification of Express Version

[Phil Spiby]

Dave,

I think we may be getting somewhere here!



> -----Original Message-----
> From: owner-wg11 at steptools.com [mailto:owner-wg11 at steptools.com]On
> Behalf Of Dave Loffredo
> Sent: Friday, May 03, 2002 10:15 PM
> To: wg11 at steptools.com
> Subject: Re: Identification of Express Version
>
>
> >>>>> "Phil" == Phil Spiby <Phil.Spiby at Eurostep.com> writes:
>
> Phil> Sorry if you think I am being a little dim here, but I don't see
> Phil> a problem!
>
> There may or may not be a problem -- that's what we are trying to
> explore with this discussion.
>
>
> Phil> Of course the ASN.1 number will change if a TC is published, but
> Phil> so far we have only published TC's when there has been a clear
> Phil> technical need to go through the process. Although some people
> Phil> believe the two TC's issued against edition one only fixed up
> Phil> minor editorial issues, they both had varying effects on
> Phil> conforming parsers (TC2 much more so than TC1 admitted).
>
> Yes, I think we are getting closer to something here.
>
> So this version id is more about parser compliance than about
> language compliance.
>
> In my opinion, way to address parser compliance is not to force AP
> developers to add version numbers to every schema.  Rather, it is to
> add TC1, TC2, etc to the PICS, and then have the customers demand
> these features from the vendors!
>
>
> ------------------------------
> A little thought experiment
>
> What if this ID was in place when TC 1 & 2 came out?
>
> Well, all of the IS schemas in existance would have an ID in place
> requiring version(1).  Would TC1 apply to them?  You would probably
> put verbage in TC1 saying that it would.
>
> From that point on, everyone would just put version(2) in any new
> schemas.   Now it's starting to get confusing to the schema writers
> because the IRs have a mix of version(1) and version(2) in them.
>
> What have the vendors done at this point?  Would they update their
> parsers any faster than before?  Probably not.  Will they refuse to
> process versions > 1?  Not likely, unless they want to annoy their
> customers.  But if enough vendors do refuse to process versions > 1,
> the AP developers will avoid calling for newer versions!
>

If there was a syntactic element at this point, such as the ASN.1 text, then
the vendors would either be:
 1. processing this syntactic element and parsing according to this level of
conformance,
 2. processing this syntactic element, possibly issuing a warning that not
all features were supported, then validating against a previous version of
EXPRESS
 3. ignoring the syntactic element, and just validating against a previous
version of EXPRESS
 4. expecting a specific syntactic element and then falling over if an
unexpected (probably later) one occurred.

In my opinion, options 1,2 and 4 are acceptable to the developers, since
they say effectively:
 1 - The vendor is aware and implemented the most recent changes.
 2 - The vendor is aware, but not implemented the most recent changes.
 4 - The vendor is either un-aware, or has made a business decision not to
implement the latest changes.

A developer could decide to use tools from vendors supporting 2 & 4 if they
do not depend on schemas which use later versions of EXPRESS.

> So in 2002, we would have schemas with version(1), version(2), and
> version(3).  Seems like this makes life more confusing for the schema
> developers, without changing the status quo on parser compliance.
>
>
> 					       - Dave
>
>

- Phil