Re: New wiki system: logging in

[Michael Heath]

That OpenID service sounds perfectly reliable; kinda like a somewhat 
simplified Kerberos-over-HTTP. Not a bad system, and doesn't sound 
dangerous. But, it does require that the user has an account registered 
with an OpenID provider, and OpenID isn't widely used at all. I've only 
stumbled on it twice: the first time was a blog that used OpenID to 
authenticate for commenting, and the second time was here.

Its not a bad solution, but it isn't a solution to your "yet another 
username and password to remember" problem. No one uses this protocol; 
users at our Wiki would be creating an OpenID account just for editing 
the wiki. You'd still have yet another username and password to remember.

Site specific authentication is fine. Theres no problem with that. Stick 
with that and then work on (in the future) seeing about integrating it 
with Savannah's authentication.

Michael Heath

Thomas Schwinge wrote:
> For one, ikiwiki supports the usual (username, password) combo for
> logging in.  Yet another username and password to remember.  Or
> alternatively have your browser remember it for you if you trust that
> one.
>
> An idea of mine is to also recognize if someone is logged in to
> Savannah's web services and as well use that as an indicator whether he
> is logged in to the wiki.  But that needs to be implemented first, and is
> a long-term goal.
>
> Now, ikiwiki also supports OpenID for logging in, see
> <http://ikiwiki.info/plugins/openid/>, <http://openid.net/>,
> <https://www.myopenid.com/>.  Does someone have any experiences with that
> service?  I just read a bit about it.  It sounds very powerful.  And it
> sounds a bit dangerous as well.  So, do we want to use it?  Do we want to
> forbid using it?
>
>
> Regards,
>  Thomas
>