RE: [election-services] The Future of EML

Paul

It was probably inevitable that we have arrived at a version of EML that is not optimal as the requirement continues to evolve and experience gained as we come to use the schemas. Clearly if we knew back in 2001 what we know now we would have ended up with a better specification. However let's not lose sight of the fact that what we have is far superior to anything else around. I don't get the sense that any of it is wrong, just that it could be better. I don't share your concerns about going into ISO with this version. I don't think we should fear other experts who might criticise it, we have most of the global e-voting expertise within the TC and the Council of Europe working group, and that body of opinion gives us a considerable advantage.

We have adopted the process of modifying the schemas periodically in the light of experience and if what you say about 120 and others being superfluous we should add that to the next update cycle, when we can also feed in any further changes resulting from the next UK and Belgium pilots. We can correct any documentation problems in the current Public Review round so perhaps you and David can post comments to that effect so there is a formal record which we can address at the end of the review cycle.

We have discussed your suggestion before about producing a data model and clearly we are not going to get anybody to fund that work. However I am prepared to put some personal time into this so perhaps we could meet up at your convenience to scope out what needs to be done etc. If the TC are content with our plans we can include that work into the next version of EML.

Regards
John
M. +44 (0)7976 157745

----- Original Message ----
From: Paul Spencer <paul.spencer@boynings.co.uk>
To: eml <election-services@lists.oasis-open.org>
Sent: Wednesday, 11 April, 2007 1:30:41 PM
Subject: [election-services] The Future of EML

In my recent work involving EML, it has been clear that the clarity and
consistency of the specification are suffering badly from the effects of its
continued evolution. Some years ago, I suggested building a proper data
model, but this was turned down on cost grounds.

Typical of the issues I am now finding is that the 330, which was originally
intended to supply a list of electors for the purpose of voting, has
evolved to cover most areas where such a list is required. For example, in
the UK, it is being used to provide a list of electors to credit reference
agencies. This means it is duplicating the 120 message, and I am being asked
which should be used. If there is this confusion, it will lead to
inconsistent implementations. Now I am being asked for similar features in
the 230. It seems logical to remove the 120 and expand the 230, 330 and 630
to cover its functions.

This is just one example. I believe that a thorough review of the schemas,
including the development of a data model, would help clarify many points
and help with consistency of implementation. Since the ISO process is not
one to be taken lightly, I would like to see it done before submission. I
currently worry about the feedback we will get if other experts look at the
current version in detail.

Why can't we just get on and do this work? Because, in spite of the presence
on the TC of several large companies working in the multi-billion dollar
voting industry, most of the technical work is being left to two individuals
who receive no payment. Is it not time for some of the big companies either
to allow some staff to work on this or put up some money so that the current
volunteers can do a decent job?

We should be able to get EML up to the standards of UBL and other
specifications that have had significant resource thrown at them. If not, it
reflects badly on the industry and will end up costing people more as
interoperability becomes harder to manage.

Regards

Paul Spencer
Director
Boynings Consulting Ltd
http://boynings.co.uk

election-services message