On 14/01/2011 21:37, Francis Davey wrote:
On 14 January 2011 21:05, Mark Goodge<[email protected]> wrote:
It works reasonably well for the land registry. The information is available
to anyone who needs it, but the costs of creating a full database would be
prohibitive even for large corporations.
Its available, but its _expensively_ available. Sometimes you need to
search a lot of titles to be able to reconstruct information that
really ought to be available under one. In practical terms this means
a lot of downloads, and the cost adds up. Fine if its a big
conveyance, but not everyone is in that position.
That's a fault in the way that the data is structured, though, rather
than the way that access to it is controlled.
But that's just from a normal user's perspective. There's a lot of
information locked up in these databases that can't be got out in any
easy way. For example, any kind of study of the nature of land
ownership in England would want to cover a lot of the land registry,
but would be prohibited by cost.
That's true, but that's a different issue. Statistical and overview data
ought to be freely available.
Also, reverse searches _are_ sometimes sensible. Finding out that X is
the director of a lot of other companies is (I think) something that X
should not be able to hide, even if X can hide their address.
Without an address, though, or some other unique identifier, that can't
be done anyway. Names are not unique. One of the effects of the change
in the law which allows company directors to use the registered address
of the company as their registered address is to make it impossible to
know if, say, the Fred Bloggs who is a director of company X is the same
person as the Fred Bloggs who is a director of company Y.
I'm not suggesting that price-rationing is an ideal solution. But it is one
option, and may possibly be more effective than some others.
Anonymising the data properly? Or removing particularly sensitive information.
Really, this is about deciding what should be made public and what
should not. The idea that only organisations with the money or effort
to circumvent security-by-infeasibility is not (in my view) a good
one.
It's also about the restrictions placed on the reuse of public data, and
how those restrictions are enforced.
Eg, the liberal democrats once bulk mailed (as in posted via the Royal
Mail) all electors in the City of Cambridge (I think it was). They did
this without breaking any rules about getting a computer readable
database of the electoral roll or scanning it or anything, but the
_hard_ way by having lots of volunteers go through it by hand.
I think that's a misuse of the data. It would be illegal for a
commercial organisation to do that, I don't see any reason why political
parties should be exempt.
So, in my view, either data is something that you are happy with being
public or not. There really isn't room for a middle way.
The problem is that we pretty much have to have a middle way for some
data, even if it's a kludge. The electoral register is the classic
example. Restricting access to it per se would be undesirable for all
sorts of reasons. But a completely open database of the electoral roll
would be equally unacceptable. There are a lot of people (eg, victims of
domestic violence) who, for very good reasons, don't want their current
address to be obtainable via a reverse search of the data. In the past,
this wasn't a problem because copying the data in bulk was physically
impossible and a reverse search can be made near-impossible simply by
controlling the way that the data is presented. If you open it up to
electronic distribution then neither of these is true any more.
The debate then becomes about what is and is not suitably public
(sorry if I'm repeating myself - I plead having been to Southampton
today). In the case of the land registry, we don't really need to know
names of owners, just whether two owners are the same or not, or are
the same as some other person somewhere else in government data (both
impossible questions to answer with the data as it is currently
recorded of course). When conveyancing you want to check that the
person you are dealing with has title, but there's no way of doing
that anyway.
For Companies House, its reasonable that the fact that a particular
person is a director of a company should be public, but maybe their
address not.
The difficulty with both of these is that a name does not establish
identity. I have a fairly uncommon surname, but even so if you Google it
you will find other holders of my name who are not me. So you can't tell
whether two land owners are the same or not, or whether a particular
person (as opposed to a particular name) is a director of a company
without some other unique identifier such as an address.
...
I seem to have found myself in the position of defending - or, at least,
of being perceived to defend - restrictive approaches to government
data. That's not really my intention, and I don't want to give the
impression that I would actually choose some of the options that I've
presented in the course of this discussion. But these are real
objections that will be presented by those who really do believe in
them, and even if I disagree with their preferred solutions I do agree
that they have a genuine concern which needs to be addressed.
Mark
_______________________________________________
developers-public mailing list
[email protected]
https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public
Unsubscribe:
https://secure.mysociety.org/admin/lists/mailman/options/developers-public/archive%40mail-archive.com
