On 5/31/07, Morgen Sagen <[EMAIL PROTECTED]> wrote:

Yes, so I was asking why wouldn't the server itself be the thing to
triage untriaged items?

because the server should be a dumb repository, not something that
takes action to arbitrarily change data without a client being
involved.

we could certainly make a simple, default-value-setting rule like "any
item that is stored without a triage status has it automatically set
to NOW".

choosing whether or not to change an item's triage status based on how
much time has elapsed since the last time an item was stored, or
anything of any greater complexity of the setting of a default value
at creation time, is a client's responsibility, not the server's.

So I asked what about the case where a
large number of items are all in the DONE bucket?  Doesn't that have
a similar traffic problem?

yes, and we have options as to how to display those items in the web
ui. i don't recall specifically what we decided last week, but i think
we at least strongly considered only showing the most recent n DONE
items. something like that.

but, while the amount of data that is transferred between the server
and the client is part of the issue, it's not the entire issue. the
amount of post-processing that has to be done on the client after it
gets the data, and the number of updates it correspondingly has to
send back to the server, are also concerns.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Open Source Applications Foundation "Design" mailing list
http://lists.osafoundation.org/mailman/listinfo/design

Reply via email to