[humanities-dev] Federated ID for TEXTUS
christian.morbidoni at gmail.com
Thu Feb 23 09:26:48 GMT 2012
In my experience logging with OpenId is very easy, given that the user
has a gmail, yahoo or other OpenId providers (which I guess is the 99
percents of the cases...isn't it?).
This is how we implemented it in SemLib, give it a try:
Well at least is seems easy to me...do you think such a workflow would
confuse the user?
I'm a big fun of this solution as it does not require your server to
store passwords and so on, thus making things easier for developers.
However this BrowserId that has been mentioned, which I didn't know
about honestly, looks even better at a first glance. In my
understanding the major advantage is that the user does not have to
automatically login to the provider (say google) when she logs in to
the website (say OpenPhilosophy). What I'm wondering is if practically
implementing in would be as easy as with OpenId, in terns of libraries
support for example...
On Thu, Feb 23, 2012 at 9:35 AM, Jonathan Gray <j.gray at cantab.net> wrote:
> Unless we can making signing up with OpenID *really* unbelievably
> streamlined and easy, my vote would be for rolling our own. I would
> guess that users of sites like OpenPhilosophy.org will probably not
> have much technical experience and OpenID sign in could be a barrier.
> On Wed, Feb 22, 2012 at 11:36 PM, Tom Oinn <tom.oinn at okfn.org> wrote:
>> To use TEXTUS users are going to have to identify themselves (or at
>> least to use most of its interesting features). Do we have a
>> preference for how this happens? This is just how we acquire the user
>> ID and verify it, not where the user's data is stored, a few options
>> would be:
>> OpenID - widely used, anyone with a Google or Yahoo! account has an
>> OpenID even if they don't know it
>> Facebook Connect - authenticate against facebook accounts, never used
>> this but it's an option
>> Roll our own - more hassle, users have to remember logins and
>> passwords specific to this system, they'll inevitably end up using the
>> same ones as they use for the email accounts, we have to verify
>> accounts etc
>> As you might have guessed I'm in favour of OpenID or, if we must,
>> rolling our own. I have a strong objection to Facebook Connect due to
>> that company's cavalier attitude towards privacy, user data etc, but
>> it does work for this kind of thing. Of course, if someone were to
>> turn round and point out a cross-OKFN standard way of doing this
>> that'd be even better.
>> humanities-dev mailing list
>> humanities-dev at lists.okfn.org
> Jonathan Gray
> humanities-dev mailing list
> humanities-dev at lists.okfn.org
More information about the humanities-dev