[myexperiment-hackers] Re: access to myexperiment-data

[David R Newman]

IMPORTANT! This message has been blind-carbon-copied to you.
Do not reply-to-all or forward it without the author's permission.

Hi Wim,

My colleagues have raised some issues with the data that the public 
database would expose.  Looking at you original request the data that is at 
issue would make your recommender system a non-starter.  To get hold of 
viewing/download statistics for a user they need to opt-in to using your 
recommender system.  The myExperiment team will discuss if there is a way 
of integrating your app so that users can opt in.  I would expect it would 
involve using OAuth at some point, see 
http://wiki.myexperiment.org/index.php/Developer:OAuth, you may need a wiki 
account to view this page?  If you have any questions feel free to ask me, 
as I set up OAuth for myExperiment so have a reasonable understanding for 
how it works.

Regards

Dave Newman

--On Monday, November 17, 2008 15:57:32 +0100 Wim De Smet 
<address@hidden> wrote:

> Hi David,
>
> Sorry to bother you yet again. Has there been any process on evaluating
> the code for your script and if so should I ask someone to run it for me?
>
> greets,
> Wim
>
> David R Newman wrote:
> > Hi Wim.
> >
> > I have written a script.  It is currently going through a checking
> > process with the other developers.  It is likely that another week or
> > so will be needed for them to do this checking.
> >
> > The script (GeneratePublicData.php) is available on the myExperiment
> > SVN in the /lib/maintenance folder.  Instructions on how to use it are
> > in the /doc folder. This script needs to be provided with a username
> > and login to a myExperiment database so is probably of little direct
> > use to you at the moment.
> >
> > Regards
> >
> > David Newman
> >
> > --On Monday, November 03, 2008 08:30:32 +0100 Wim De Smet
> > <address@hidden> wrote:
> >
> > > Hi David,
> > >
> > > have you made any progress on this yet? I noticed a script was available
> > > in the myExp repository, not sure if that's the one you've been working
> > > on?
> > >
> > > greets,
> > > Wim
> > >
> > > David R Newman wrote:
> > > > Hi Wim,
> > > >
> > > > You are correct about the who downloaded/viewed records, these could
> > > > only be provided as anonymous statistics.  I.e. we could probably tell
> > > > you that the same user had downloaded workflow one 10 times and
> > > > workflow two 5 times but that user could not be identified.
> > > >
> > > > On obtaining publicly available data I have written a script to
> > > > produce publicly available RDF, which I should be able to adapt to
> > > > produce public version of the MySQL database.
> > > >
> > > > Regards
> > > >
> > > > David Newman
> > > >
> > > > --On Monday, October 27, 2008 08:49:52 +0100 Wim De Smet
> > > > <address@hidden> wrote:
> > > >
> > > > > Hi,
> > > > >
> > > > > I'm just following up on the mail from myexperiment-topsecret about
> > > > > outside access to the myexperiment data, since David De Roure said you
> > > > > were the man to talk to about getting a data dump. Publicly available
> > > > > data is good but I would really like access to 'who downloaded
> > > > > what'-stats which are probably not publicly available. I don't know if
> > > > > anything could be done in this respect.
> > > > >
> > > > > greets,
> > > > > Wim
> > > > >
> > > > > P.S.: it goes without saying that any non-publicly available data
> > > > > that I
> > > > > might get will stay non-public and only be used for research purposes