building a task list

[Paul Miller]

I'm just having a terrible time building a list of tasks really
similar to the one hiveminder presents using the API. I'm trying to
build something really really lightweight, since it runs on a phone
(webos).

Sorry if this seems more like a rant than a question. It takes a
lotta space to see what I mean.

I've been using http://hiveminder.com/=/action/DownloadTasks.yml to
download tasks and I've been reasonably happy with the results. I can
pass in my pro-user search list tokens in the query field (slash
separated) and I get results like this:

- accepted: 1
attachment_count: 0
complete: 0
completed_at: ~
created: 2009-09-06 07:43:39
description: "blah blah blah"
due: ~
group: ~
id: xxxxxx
last_modified: 2010-04-17 11:40:28
next_action_by: 'Paul Miller <jet...@gmail.com>'
owner: 'Paul Miller <jet...@gmail.com>'
priority: 3
record_locator: xxxx
repeat_days_before_due: 1
repeat_every: 1
repeat_next_create: ~
repeat_period: once
repeat_stacking: 0
requestor: 'Paul Miller <jet...@gmail.com>'
starts: ~
summary: blah blah blah
tags: hiveminder webos
time_estimate: ~
time_left: ~
time_worked: 1d8h
type: task
will_complete: 1

But I'd also like to be able to show the before and after stuff that
hiveminder does. Notice that it doesn't seem to appear when you
DownloadTasks. So then I tried http://hiveminder.com/=/action/TaskSearch.yml
where I can pass my pro search list tokens to the "tokens" param
(undocumented) and I get this:

- accepted: 1
attachment_count: 0
complete: 0
completed_at: ~
created: 2009-09-06 07:43:39
depended_on_by_count: 1
depended_on_by_ids: 1005037
depended_on_by_summaries: blah blah blah
depends_on_count: 5
depends_on_ids: 1038539 1038540 1038542 1038543 1038546
depends_on_summaries: 'blah blah blah blah blah blah'
description: "blah blah blah"
due: ~
group_id: ~
id: 849559
last_modified: 2010-04-17 11:40:28
last_repeat: 849559
milestone: ~
next_action_by: 42673
owner_id: 42673
priority: 3
project: ~
record_locator: TXNQ
repeat_days_before_due: 1
repeat_every: 1
repeat_next_create: ~
repeat_of: 849559
repeat_period: once
repeat_stacking: 0
requestor_id: 42673
starts: ~
summary: blah blah blah blah blah
tags: '"hiveminder" "webos"'
time_estimate: ~
time_left: ~
time_worked: 1d8h
type: task
will_complete: 1

The first one is more immediately useful, as you can see. Since webos
is pretty much just prototype.js with fancy widgets, I'd really rather
stick to download tasks since then I just punching things into
prototype Template objects.

The only way I can see to fill in those requestor_id: 42673 and
owner_id: 428282 is to go back and do a web request for every one of
those... that's quite a lot of traffic for a webos application,
particularly where you might be in a situation with 30B/s speeds, so
that's right out the window.

I can do an O(n^2) operation to combine the two task list queries into
one list that has both template usable stuff like requestor: "Paul
Miller" instead of requestor_id: 8383838. That seems like just about
the only realistic option unless someone is aware of a single request
that lists tasks by tokens and provides enough information to actually
list the tasks.

--
You received this message because you are subscribed to the Google Groups "Hiveminder API Hackers" group.
To post to this group, send email to hivemin...@googlegroups.com.
To unsubscribe from this group, send email to hiveminder-ap...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/hiveminder-api?hl=en.