Forum OpenACS Development: Response to General Alerts

Collapse
8: Response to General Alerts (response to 1)
Posted by Michael Feldstein on
<p>OK, I'm working with Malte to see the latest ACES
interface.</p>

<p>In the meantime, it will help to nail down the functionality
goals in more detail. In particular, I need to know not only what's
likely to get built into the first iteration, but also what is likely to
get built into the second and third iterations, so I can design the
interface to accomodate future needs. So here are some
random questions/thoughts on functionality:</p>
<ul>
<li>Using bboard as our archtype, there are at least two levels of
alerts. First, there are alerts for a package instance (e.g., one
bboard.) Second, there are alerts for particular items (e.g., one
discussion thread in bboard or one ticket in ticket-tracker.) As I
recall, the ACES implementation only allows you to toggle the
first kind (i.e., at the package instance level). Would we want this
package to enable both levels of control?</li>

<li>If the answer to the previous question is "yes," then how do
we handle interactions between the instance-level settings and
the item-level settings? For example, I could imagine somebody
wanting a daily digest of a bboard but instant alerts for one
thread. What options do we want to offer? What kinds of default
behaviors make the most sense?</li>

<li>I can imagine somebody wanting to set alerts on a third level,
in which they want to know about content items of a certain type
across all package instances. For example, somebody might
say, "Alert me to all content items submitted by Al Essa," or "Alert
me to all content items that are categorized under XML," or even,
"Alert me to all content items that contain the words 'cell,' 'phone,'
and 'AT&T' in them." Do we want to plan for this sort of
functionality in the foreseeable future? (If so, then I suspect that
the UI design should overlap with the search UI.)</li>

<li>I understand that we want an interface that makes it possible
to plug in new alert channels (e.g., email, Jabber, SMS, etc.) as
they are developed. At what level of granularity do we want/need
to allow users to control their channel output? At the item-level?
At the instance-level? Both? And am I correct in assuming that
we want users to be able to select output to, for example, both
their pager and their email?</li>

<li>Which of the features listed here are likely to make it into the
first iteration? Which into the second or third? And which should I
just cross off the list entirely?</li>
</ul>