This is the mail archive of the frysk@sources.redhat.com mailing list for the frysk project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Assistant refinement


Sami Wagiaalla wrote:
I am suggesting this:

user starts up frysk:

[1]
A page pops up with a list of previous sessions the buttons new and delete to the side.
in the bottom the buttons [next] and [finish].
clicking [finish] at any point starts the monitor/debugger depending on was was saved from before.
this enables the user to select the session and start it in two clicks.

No. "next" is part of an assistant, and the Session Manager is not an assistant. The Session Assistant is for either creating a "new" session, or picking and [re-]"open"ing an existing session. In the case of re-opening an existing [monitoring] session, the operation is expensive and potentially fatal, the user needs to be presented with a check-point confirming what's about to be monitored so that they can confirm it. That means at minimum of double-click, click.


If the desire is just to reduce the mouse-click count then there are better ways of doing this. For instance:

- debug session of a "new process" doesn't need process-picker and/or observer-picker. For it "open" jumps to the source code window, starts the process and runs it to main().

- let the user explicitly edit the session so that it skips the p-p window. When combined with specialized logic, it would let the user quickly monitor specific processes based on an arbitrary criteria. For instance, a session to "Monitor the top 2 processes".



user clicks next/new


[2]
they see the page currently referred to as the process picker.
In the bottom the buttons [back] [next] [finish]

user clicks next

[3]
they see this: http://sourceware.org/bugzilla/attachment.cgi?id=1629&action=view
(maybe here if they choose "debugging" (should be Debuging Session :)) then next is disabled)



Assuming the differentiation between debugging and monitoring remains, then I think this question should be asked up front. That way the the first thing the user is presented is:


-> a summary of what is happening

-> a question as to what they want to be doing initially

That way the user gets a warm-fuzzy that frysk "knows" what they want and is being customized for that need. With out that up-front question one is wondering: "where is this going?", "what is this doing?".

To give an example, consider connecting a machine to a network. As a user, you have an ethernet cable in one hand and the lap-top in the other. The first question you expect is: "which network type?" "eithernet", "wireless", "modem", ... ;; and not "is this using dhcp?" say - "DHCP for what"?

Andrew


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]