Re: [xml-dev] The triples datamodel -- was Re: [xml-dev] Semantic Web permathread, iteration n+1

["Thomas B. Passin" <tpassin@comcast.net>]

Robert Koberg wrote:

> 
> So you had a contract with the client.

Not even in your dreams!  The contractor is developing this JMS service
fro the state.  They say "If you want to use the service, here is an SDK
which you use as a starting point to modfy the jms client.  Have at it,
fellows!".  Now, the SDK includes the schema and some sample files.  The
sample xml files don't (quite) match the schema - which, BTW, had a few
problems itself I had to fix.

> The client did not stick to the contract

The only contract I can see is the implied one that the contractor will
continue to produce whatever it is they are producing right now.

> and you had to do a great deal of work to make your end work. If it 
> was me, I would say to the client that you have to simply conform to
>  the schema that we agreed to, otherwise we have to do a great deal 
> more work and we will have to charge for all this work.

Well, we can tell the contractor, but they have a working system that
other are already using, so changes like this are probably low on their
priority list.  In other cases, when we have been able to get into the
process earlier, we have been - sometimes - able to get the other party
to adjust their stuff.  Usually, though, there is an awfully big tail
and we are a small dog.

We are not even working for the state of CA in this case, we are working
for the US Department of Transporation.  There is no way we are going to
be able to lean on the US DOT to lean on the CA DOT to lean on their
contactor to change anything.

Better to be good at xslt!

> If the requirements changed then adjust the schema accordingly. That
> is much simpler than doing the work you did.

The requirements didn't change.  The contractor is using a draft version
of the final schema, which is very different from the final one anyway,
so the only sensible change they should be making is a change to the
final schema.  That would *really* break all their users - except us, I
could get it working in five minutes.  That's because I transform their
stuff to the current (hopefully final) schema form anyway.  I would just
stick in an identity transform and be done.

> It is so simple to produce a valid document.

Tell that to all the sites that supposedly produce XHTML but that don't
pass the validator.  It's simple, perhaps, for _us here_, but apparently 
not always.  Anyway, one a system gets in place and working with the 
wrong stuff, it is hard to change when you have lots of users using what 
you have.

> Hopefully you charged for it -- wait, I take that back being a
> California taxpayer. It almost sounds like negligence...

Just the way things usually work on government jobs.  And not just the 
government ...

Cheers,

Tom P

-- 
Thomas B. Passin
Explorer's Guide the the Semantic Web (Manning Books)
http://www.manning.com/catalog/view.php?book=passin