What are our design principles?
Mike Pearson
of design principles/requirements, to ensure we are all thinking the
same things?
For instance, I actually view the overarching problem as a set of
issues about AUTHORITATIVE data, OWNERSHIP and LONGTERM ACCESSIBILITY.
What are my requirements as a USER?
* I own my data
* I want to put it on the Net, so I can share it, access it anywhere
* I want to store my (functional) data in a single (functional) place
* I do NOT want to go back and ever move/transform my data
* I will allow (functional) applications access to my data
* I will be annoyed if any new application breaks my existing
applications
A simple set of requirements, but they break most of the existing
Internet business models of the last 5 years. As an example, think
about your photos in Flickr.
The technical implications are potentially things like a five-tier
software architecture:
http://www.opengroup.org/architecture/togaf7-doc/arch/p4/views/vus_sw.htm
So what is the bigger problem? And what are the fundamental
principles / requirements needed to resolve it?
Thanks, Mike
Chris Saad
I suggest you make a new page for it here on the google group and
start an outline - and encourage others to contribute!
Cheers,
Chris
maguffyn
> I suggest you make a new page for it here on the google group and
> start an outline - and encourage others to contribute!
Ian
Chris Saad
Henry Story
On 7 Jan 2008, at 23:57, Mike Pearson wrote:
> What are my requirements as a USER?
>
> * I own my data
I have a problem with the idea of ownership of data. Simply put: you
cannot own data really.
First of all it makes no sense. Data only has value if you can trust
it. Unless you have built the notion of trust into the notion of data,
but I think that may be the concept of information...
What you can be is responsible for what you say.
If I publish a foaf file that says that my name is Henry, that I work
for Sun, and that I have a number of friends, then this is a statement
I am making. People who don't know me at all will not be able to
distinguish me from a spammer. For them that information is less
valuable. For people who do know me on the other hand, whome I have
given my foaf file, this is good information because it comes from the
source of the entity (me) who has a reason to keeping this ifnormation
up to date: people will know how to contact me etc,...
In this file I can also make statements about other people. I could
say that jane loves joe. Should I not be able to say that? Why? There
is an economic reason which may reduce my desire to do so, in that by
saying something like this I am increasing my work. Now I have to
track jane and joe's love life, and it is known that this changes a lot.
So I don't believe so much in "ownership" of data. Rather putting
graphs of relations on the web is a speech act. The value of that act
depends on who makes it.
>
> * I want to put it on the Net, so I can share it, access it anywhere
yes. I also want other people to be able to link to it.
>
> * I want to store my (functional) data in a single (functional) place
No. All places on the web are identified by URIs. There is no reason
why I need to put all my info at one URI.
> * I do NOT want to go back and ever move/transform my data
Moving data is a problem because part of the value of information has
to do with the idea of permalinks. Moving information will break other
links to it.
> * I will allow (functional) applications access to my data
yes
Mike Pearson
understandings of the terms, so I will try to explain my understanding
better.
On Jan 9, 8:59 am, Henry Story <henry.st...@gmail.com> wrote:
> On 7 Jan 2008, at 23:57, Mike Pearson wrote:
>
> > What are my requirements as a USER?
>
> > * I own my data
>
> I have a problem with the idea of ownership of data. Simply put: you
> cannot own data really.
disposing of a thing. People expect to own their data.
Electronic data has the characteristic that you can make copies, and
have two or more versions of the data. Then you get interesting
issues like you gave someone else a copy of your data and they then
sell it, change it or deletes it.
As an example of the ownership issues -- I gave many of my details to
Facebook. Some of those details will eventually be out-of-date. If I
no longer use Facebook, do I have a right to tell them to stop using/
publishing my "out-of-date" data? In the future, if I gave Facebook
access to the authoritative source for that data, do I have the right
to revoke access in the future (that's an ownership right).
> > * I want to store my (functional) data in a single (functional) place
>
> No. All places on the web are identified by URIs. There is no reason
> why I need to put all my info at one URI.
single (function), in a single place. For example, I put my (photo)
data in a single (photo) data store. I put my (mail) data in a single
(mail) data store. Thinking about this further, perhaps this is part
of the issue. If we break apart (physical storage) from (application
service), we have less tension.
Perhaps then,
* I will pay someone (DATA STORAGE SERVICE PROVIDER) to physically
host my data on the net.
* I will pay for some tools to manage/manipulate my data (APPLICATION
SERVICE PROVIDER).
* I may participate in some communities that aggregate my data with
others (SOCIAL NETWORK APPLICATION SERVICE PROVIDER)
* I will have the right to withdraw access to my data from SOCIAL
NETWORKS.
> > * I do NOT want to go back and ever move/transform my data
>
> Moving data is a problem because part of the value of information has
> to do with the idea of permalinks. Moving information will break other
> links to it.
thinking?):
My (PHYSICAL) data is on the Net. It may be distributed/replicated
across multiple physical file stores. I trust my service provider to
keep it secure and give me access when/where I need it. I would
expect to sometimes change my service provider, depending upon their
performance. There would need to be a process for this. This is
Physical Data Portability.
I do not wish to break my application links to my data. This is
LOGICAL Data Portability. So what is the nature of the permalink
system that links LOGICAL application link to PHYSICAL Data?
Technical solutions can be centralised or distributed. Each solution
has its pros and cons.
Thanks for your feedback,
Mike
Henry Story
> Hi, thanks for your feedback. We probably have some different
> understandings of the terms, so I will try to explain my understanding
> better.
I am as you are trying to explore these ideas. Thanks for playing along.
> On Jan 9, 8:59 am, Henry Story <henry.st...@gmail.com> wrote:
>> On 7 Jan 2008, at 23:57, Mike Pearson wrote:
>>
>>> What are my requirements as a USER?
>>
>>> * I own my data
>>
>> I have a problem with the idea of ownership of data. Simply put: you
>> cannot own data really.
>
> Ownership means the exclusive right of possessing, enjoying, and
> disposing of a thing. People expect to own their data.
But data is not a 'good' that can be owned by someone. It is not like
a car, where there can be only one person who has it, where if one
person has it, the other does not.
Since there is no "good" that can be owned by one person or the other,
there is no way to prove who has the data, and who is meant to own it.
Take your birth date. What if I guess you birthday, or if I deduce it
from other information, have I stolen something from you? What if you
think that Jane loves Joe? Is that information that belongs to Jane or
is information that belongs to Joe? Or is it information that belongs
to you?
Really you are walking straight into first amendment issues here.
> Electronic data has the characteristic that you can make copies, and
> have two or more versions of the data. Then you get interesting
> issues like you gave someone else a copy of your data and they then
> sell it, change it or deletes it.
It is interesting to look at this example in more depth.
So let us say that I give some bits to a service provider to publish
at a specific location. Publishing those bits at that location makes
it clear that I am standing behind the statement those bits make. As I
said people will trust this information depending on the trust they
have in me.
Now imagine this service provider takes those same bits and gives them
to someone else. What are they doing? Are they claiming responsibility
for what the bits are saying? They could be taking on a lot of
responsibility doing this. It could be very dangerous to them. At the
most extreme those bits could say: "kill the person who gives you
these bits".
And if they say that I say what the content of those bits says, then
there are two possibilities:
- they link people to the location where I placed those bits
originally, and the reader can see that information too, so then what
advantage does he get from the going through this intermediary?
- either they don't link to the information or they do but the
reader cannot access that information. The reader then he has to trust
the intermediary to be telling the truth - the same intermediary that
clearly was breaking a contract law. Also the information won't be up
to date. And up to date information is a lot more valuable, because it
can be the difference between truth and falsehood.
> As an example of the ownership issues -- I gave many of my details to
> Facebook. Some of those details will eventually be out-of-date. If I
> no longer use Facebook, do I have a right to tell them to stop using/
> publishing my "out-of-date" data? In the future, if I gave Facebook
> access to the authoritative source for that data, do I have the right
> to revoke access in the future (that's an ownership right).
I would rather speak of contract law, rather than data ownership. If
people use Facebook as a place they expect up to date information to
be available, and Facebook publishes your information when you have
told them that you are no longer keeping it up to date, then they are
publishing information as if it were coming from you, when it does not
in fact come from you. So they are lying. This can have huge
consequences not just for you, but for all the people who are trusting
Facebook and the data they are publishing. And what could be worse
that loosing trust for an information provider?
Because information meshes very quickly with other information to
build up in consequence more information, make a false claim, can soon
lead to a lot of problems building up. Especially for people who don't
want to take responsibility for what they are saying.
>
>>> * I want to store my (functional) data in a single (functional)
>>> place
>>
>> No. All places on the web are identified by URIs. There is no reason
>> why I need to put all my info at one URI.
>
> My assumption was that most people manage/collect data related to a
> single (function), in a single place. For example, I put my (photo)
> data in a single (photo) data store. I put my (mail) data in a single
> (mail) data store.
My point was that it does not matter where physically your data is,
because you can access it all logically via URLs. I have photos on
flickr and others on my own web server.
> Thinking about this further, perhaps this is part
> of the issue. If we break apart (physical storage) from (application
> service), we have less tension.
Interestingly the web comes with a mechanism to do this. By buying/
renting a domain name, you can move infrastructure provider without
loosing your URLs.
I have done that recently. I moved from verio to gridzones and all my
links still work correctly.
This won't be possible of course for my pictures on flickr. Here we
would need to be able to ask flickr to provide HTTP redirects to get
the same effect.
> Perhaps then,
>
> * I will pay someone (DATA STORAGE SERVICE PROVIDER) to physically
> host my data on the net.
And to make it clear it is coming from you. (they are not just random
bits)
> * I will pay for some tools to manage/manipulate my data (APPLICATION
> SERVICE PROVIDER).
Yes, and hopefully this will be open so that you don't get locked in
by one tool vendor.
> * I may participate in some communities that aggregate my data with
> others (SOCIAL NETWORK APPLICATION SERVICE PROVIDER)
Yes. And the more this happens the less clear it will be where "your
data" begins and other data ends.
> * I will have the right to withdraw access to my data from SOCIAL
> NETWORKS.
Not completely. Once you have told me your phone number, you have no
right to come to me and ask me to forget it. What you can do is stop
publishing it, so that my information will over time be out of date.
>>> * I do NOT want to go back and ever move/transform my data
>>
>> Moving data is a problem because part of the value of information has
>> to do with the idea of permalinks. Moving information will break
>> other
>> links to it.
>
> My definition of what we are discussing (is this what you are
> thinking?):
>
> My (PHYSICAL) data is on the Net. It may be distributed/replicated
> across multiple physical file stores. I trust my service provider to
> keep it secure and give me access when/where I need it. I would
> expect to sometimes change my service provider, depending upon their
> performance. There would need to be a process for this. This is
> Physical Data Portability.
>
> I do not wish to break my application links to my data. This is
> LOGICAL Data Portability. So what is the nature of the permalink
> system that links LOGICAL application link to PHYSICAL Data?
> Technical solutions can be centralised or distributed. Each solution
> has its pros and cons.
The best way, as mentioned above is to own your own domain name. Then
you can easily move what you call your physical data - the way certain
pieces of metal are marked electronically - to other such machines,
without anyone noticing any difference when using URLs.
But this won't always be the case. I have a lot of photos on flickr
now with flickr URLs. I won't be able to move those pictures easily.
Perhaps it should be a right for us to ask flickr to allow us, if we
need to move pictures away from their service. They would have to
provide a HTTP redirect for every picture moved.
> Thanks for your feedback,
Thanks for helping me develop some of these thoughts,
Henry
> Mike