[Ex-plain] Finalizing a version of Explain-- for use in real life

Robert Sanderson azaroth at liverpool.ac.uk
Mon Apr 15 12:59:28 CEST 2002


> I am wondering if a timetable should be put in place to do a snapshot
> of Explain-- so people can actually start implementing it. An important

That would be very useful for me at least.


> Good progress has been made, but every time I turn around things seem
> to be changing again. Just recently there was discussion of "hey,
> maybe all the element names are wrong and we should change the whole
> DTD around". This sent shivers down my spine.

It has only been about 6 to 8 weeks since we started this whole thing. The 
progress that we've made has been very encouraging I think.  On the other 
hand, as it is only 6-8 weeks I think that we shouldn't feel too 
constrained in making changes.  Consider the other projects under the ZiNG 
umbrella and how long they've been going and how many changes they've had 
over their life times!

> So does someone want to have a go at defining the requirements for
> deciding a version is finished enough to make it into the real world?
> For example, is it politically sensible to have 3 levels of conformance?

I think that we should have an 'Implementation Candidate' ala OAI.  Here, 
implement this but know that we're working on an expanded version which 
shouldn't be too hard to upgrade to.

> * Define 1 set of queries (and attributes) to be supported with a
>   definition of behaviour.

Yes.  

> What else? Is it acceptable to have additional levels of conformance
> introduced later (as well as other extensions using namespaces etc)?
> Is it more important to get it right, or get it accepted (and so keep
> it simple).

Currently there is a need for extra places in which to put things later.
Our difficulty in finding a good place for number of Records, last Update, 
piggy-back support, query type support and so forth speaks that our 
pentachotomy needs a little work.  That we've expanded some things past 
their original definition means that the names are somewhat out of sync.

I'm tempted to say:

We should release a version 1.7 in the next week and freeze it.  This is 
the 'if you want to be an early adopter, implement this version' release.
We should then work on version 2.0 (which would have its own namespace for 
elementsetname and namespace defn etc)

By the end of next week we should have finalised an attribute set (or two) 
for version 1.7 and hopefully a profile of searches to implement with 
expected behaviour.  Apart from the protocolVersion as semantic qualifier, 
is there any problem with my proposed attribute set?

After that, we implementers can go off and do our thing until such time as 
version 2.whatever is stable enough to upgrade things to.  I don't really 
foresee massive differences in version 2 apart from some names being made 
more sensible and some bits and pieces moved around to make space for new 
things.


Rob

-- 
      ,'/:.          Rob Sanderson (azaroth at liverpool.ac.uk)
    ,'-/::::.        http://www.o-r-g.org/~azaroth/
  ,'--/::(@)::.      Special Collections and Archives, extension 3142
,'---/::::::::::.    Twin Cathedrals:  telnet: liverpool.o-r-g.org 7777
____/:::::::::::::.              WWW:  http://liverpool.o-r-g.org:8000/
I L L U M I N A T I





More information about the Ex-plain mailing list