Re: Simultaneous Username/Password Detection

2001-03-23 Thread Daniel Lopez
Hi, Regarding security, the JSDK spec is IMHO too simple and that has forced many people to reinvent the wheel again and again, or get into proprietary solutions. We chose to go our own way to be fully portable and we end up with such a system: .- As we already have implemented a Model 2 architec

RE: Simultaneous Username/Password Detection

2001-03-22 Thread Frank Eggink
I think the major part is a design issue. How do you recognize someone is logging-in after he accidentally killed his browser? Maybe you could use IP addresses for that??? As far as I know most ISPs use simple timeout mechanisms on their POP3 servers (you can reconnect after sometime when the conn

RE: Simultaneous Username/Password Detection

2001-03-22 Thread Allen Fogleson
what we did was store session info in a DB table. (appropriately enough named Session_T (our naming scheme add a T to tables, etc) We store it with the create time and all that. As you say the hard part is when they forget to log out. or winblows locks up on them. in that case what we always do o

Re: Simultaneous Username/Password Detection

2001-03-22 Thread Vaskin Kissoyan
The new Oreilly book "JavaServer Pages" has a great way of doing authentication with a tag and the existance of a "validUser" bean in the session scope. It properly produces the original URL and is container independant. - Original Message - From: "David Morton" <[EMAIL PROTECTED]> To: "O

Re: Simultaneous Username/Password Detection

2001-03-22 Thread Jim Kraai
David, I'm just getting to know servlets in depth, and have recently read (and reread a few times) the servlet spec. Two sections of the spec (2.3) are relevant (I think Orion is at 2.3?): > 3.1 Scope of a ServletContext > > There is one instance of the ServletContext interface associated >