--- Log opened Wed Sep 21 00:00:34 2011
00:30 < mcallan> i can't get any codework done because of all the messages in metagov.  it's worth it though if a guiding star emerges... my kingdom for a guiding star
07:37 < conseo> ok, got up, r u still there?
09:08 < mcallan> conseo: still here, at least for a bit
09:26 < mcallan> i'm reading for a while c, ring my phone twice if u need me
09:59 < conseo> damn, you are still difficult to reach. i often don't see a reason to mail u about small things, but i guess i have to change that
10:00 < mcallan> conseo: i am here again!
10:00 < conseo> hey
10:00 < mcallan> hey :-)
10:00 < conseo> just tried to assemble your number
10:01 < conseo> ok
10:01 < conseo> first, this is my dev setup, i have played a bit with a progress bar: http://whiletaker.homeip.net:8888/voff/xf/?gwt.codesvr=127.0.0.1:9997#s=G!p!sandbox*21&c=DV&a=qUS$
10:02 < conseo> i wanted to talk with you about target definition
10:02 < conseo> *and condition definition)
10:03 < mcallan> ok (not sure i'm looking at your latest code, can't load the dev URL)
10:03 < conseo> ok, no prob
10:04 < conseo> nothing really fancy
10:05 < conseo> this http://whiletaker.homeip.net:8888/voff/xf/#s=G!p!sandbox*21&c=DV&a=qUS$
10:05 < conseo> does not work?
10:06 < conseo> (can only check with a textbrowser w/o ecmascript from outside)
10:07 < mcallan> i was just wondering if that's what you want me to look at - you gave dev URL, and dev URL might be different code - so it is the real thing ...
10:07 < mcallan> yes, it looks ok ...
10:07 < conseo> can you see the cheap progress bar?
10:07 < mcallan> it only goes about 1/5 across the screen
10:07 < mcallan> the green part?
10:08 < conseo> yes 200px, i will ask about the design on the list
10:08 < conseo> i just have a simple question. this is a div element (HTMLPanel) around a span element (InlineLabel)
10:08 < conseo> i set the width of the span to get the progress stuff
10:08 < mcallan> i see
10:08 < conseo> how would i properly lay text in that
10:09 < conseo> if i set the text of the InlineLabel it gets cut, so i need some other element i guess
10:09 < conseo> atm. they are positioned relative, but i guess i have to lay sth. absolute (like you do in your SacSelectionV) over it
10:10 < conseo> (not too important, but i'd like to know how to do this properly)
10:10 < mcallan> text should not be absolute, or it may get clipped vertically
10:11 < conseo> ok
10:11 < mcallan> you want text to have natural height, so lay the progress bar absolutely in the text
10:12 < conseo> so text in div and absolute positioning of span?
10:12 < mcallan> exactly
10:12 < conseo> ok, thx
10:12 < conseo> you are working on the count engine atm.?
10:13 < conseo> have you got some time to talk about target definitions? (tomorrow is ok as well, but we should talk about that stuff a bit imo)
10:13 < mcallan> re engine: no, unfort (all this theory thinking) but i did make a little progress on it
10:13 < conseo> ok
10:14 < mcallan> did you have a suggestion about targetting, or did you want me to look at designing it
10:14 < conseo> Ed has said on the list that he doesn't know a software which runs in proper local scale
10:15 < mcallan> he mentioned ours, and that it's not 100% ready for actual use
10:15 < conseo> (sry for bringing up a lot of issues here, but i would like to discuss some stuff with you and the list feels like overkill (so much too write :-/))
10:15 < conseo> right, but we have proper districts
10:15 < conseo> he didn't seem to know this
10:15 < mcallan> (no prob) ah, i didn't catch that...
10:16 < mcallan> he's right that we're not ready to do full residential registration (we got no streetwiki with proper layouts, etc)
10:16 < conseo> sure, but we have a compatible geographical divison
10:17 < mcallan> yes, we have solid designs (i think) and rough prototype code - and the divisions are well impl'd already
10:17 < conseo> ok, so i'll point him to it?
10:17 < mcallan> sure
10:17 < conseo> ok
10:17 < conseo> have you thought about how to define the targets for accounts?
10:18 < conseo> i have thought about it and i think inline is not bad
10:18 < mcallan> u didn't answer my q about that above - ah ok
10:18 < mcallan> what's inline mean?
10:18 < mcallan> back in moment
10:23 < mcallan> back - i just re-read what u wrote earlier about "inline" but i still don't understand
10:23 < conseo> a target property in the account definition: like account = {{{dollars|0}}}
10:23 < conseo> target = 1000
10:23 < mcallan> right, so a property
10:23 < conseo> yes
10:23 < mcallan> hmmm...
10:24 < mcallan> i can't think offhand of a problem with that
10:24 < mcallan> shd work
10:25 < conseo> me neihter until yet. since account definitions are related to positions it should work
10:25 < conseo> just define a new account explicitly if you want to set a different target for your position
10:25 < mcallan> exactly
10:26 < conseo> have you thought about "conditions"?
10:26 < mcallan> not sure, what are they?
10:26 < conseo> like return + interest rate, free sample of result, open source production, ...
10:26 < conseo> or payment, compensation, ...
10:27 < conseo> it can be done in the position, but it is actually tight to the resource
10:27 < mcallan> i see...
10:27 < conseo> tied :-D
10:28 < mcallan> :-D took me a minute, i was looking the other way...
10:28 < mcallan> yes, it would be part of the pledge, i guess
10:28 < conseo> no prob
10:28 < conseo> ok
10:28 < mcallan> tied again!
10:28 < mcallan> lol
10:28 < conseo> :-D
10:29 < mcallan> part of the pledge, but i don't fully understand the need
10:30 < mcallan> why not let it be informal, for starters?
10:31 < conseo> we can do that
10:31 < mcallan> after all, if the position you are voting for does not suit you, then you can always pull your vote
10:32 < conseo> but it came to my mind that it makes some sense to analyse the resource distribution instead of just define it in the position somewhere
10:32 < conseo> sure, the resource management is only for convenience as our basis allows already to use drafting for it
10:34 < mcallan> not sure i understand res distrib, or management
10:34 < mcallan> (i will sign off soon, because it's end of my day)
10:35 < conseo> i could write something in the draft like: Thomasvonderelbe contributes 50$, Michael 100$ etc.
10:35 < conseo> :-(
10:35 < conseo> i have just got up 3 hours ago
10:36 < mcallan> still here for a bit, if you need me :-)  but pledges will say that
10:36 < conseo> yes, i was just saying that your text drafting concept allows all kind of decisions describable in text (which is everything) already
10:37 < conseo> it is just not very convenient
10:37 < conseo> so targets for now and then we will see how that goes
10:37 < mcallan> ah exactly
10:37 < mcallan> text is *informal* and so infinitely flexible
10:38 < mcallan> yes, agree - formalize as little as possible for starters
10:42 < conseo> but also not programmable, and counting manually is pretty stupid for a digital platform
10:42 < conseo> so i think the resource management makes a lot of sense
10:43 < conseo> but the current concept is ok
10:43 < mcallan> resource management = count?
10:43 < conseo> any hint for how to combine TrackStack with the selected position (i think that SacSelection model reacts on the Tokelets somehow, but i cannot see it work)
10:44 < conseo> mcallan: hmm, yes (?)
10:45 < mcallan> ok, definitely need to count resources!
10:45 < mcallan> yes, sac is done through 'a' tokelet
10:45 < conseo> so the model should react on votespace selection, right_
10:46 < conseo> ?
10:46 < mcallan> as mentioned before (when i got up), you probably won't depend on votespace
10:47 < mcallan> you need ac model
10:47 < mcallan> we don't got that yet, but when we do...
10:48 < mcallan> it will be passed into your constructor
10:48 < mcallan> i guess you need that now tho, right?
10:49 < mcallan> ofc u do, i can start on it tomorrow
10:49 < conseo> not necessarily, i can first mail to the list and ask about how people would like progress to be visualized, i can also fake some stuff
10:50 < mcallan> ok, but i will still start on it tomorrow - we need it for sure
10:50 < conseo> what i need is targets ofc, and i don't know how they will be faked in JSON, should i add them to the counts?
10:50 < conseo> ok
10:51 < conseo> also atm. SacJS(_*).holdCount returns the count for the whole poll, which doesn't make a lot of sense imo
10:51 < conseo> we can only judge positions, but not the poll
10:51 < mcallan> re JSON: you will have to get info through model though, so it's all part of that
10:52 < conseo> (it is ok to show the sum in the center)
10:52 < mcallan> (i mean part of des/coding of ac model)
10:52 < conseo> get a coffee, brb
10:53 < mcallan> not sure what u mean "we can only judge positions, but not the poll" - but that's correct for hold count, i think, per API docs
10:54 < conseo> back
10:54 < mcallan> for hold count, see red #'s in figs 1 and 2: http://zelea.com/project/votorola/d/theory.xht#medium
10:56 < mcallan> and sac's are always for entire poll ofc, never anything less
10:58 < conseo> right, but targets are bound to positions (and knowing that the poll has collected 1 million $ does not tell me how it is assigned to consistent positions
10:58 < conseo> it is just an external figure to know how engaged people are
10:58 < conseo> (but this is not interesting for the TrackStack, right?)
10:59 < mcallan> right, track stack needs ac i guess, not sac
10:59 < conseo> ok
11:00 < conseo> is this discussion ok, or should i figure that out myself
11:00 < mcallan> maybe needs sac for some stuff, or sometimes (like no ac selected), but mostly needs ac
11:00 < conseo> yes
11:01 < mcallan> not at all, discussion is interesting
11:01 < conseo> and it needs the ac which might also be "up-roots" and not defined in the position itself
11:01 < mcallan> yes
11:02 < conseo> is the json supposed to serialize the tree, or is it only serializing a level of voters
11:02 < conseo> ?
11:02 < conseo> (or do we have to parse that stuff on server or client side?)
11:03 < conseo> atm. sac is just transfered as a list
11:03 < conseo> (fake list)
11:04 < mcallan> it's done by levels, and sac is entire poll - but mostly you don't need to worry about that yet, i guees...
11:05 < mcallan> i guess all your info will come from ac model
11:05 < mcallan> very similar to sac model, tho
11:05 < conseo> but it might be necessary to fetch the target from a "higher" level
11:05 < conseo> ok
11:06 < conseo> no problem, i am just trying to figure that stuff out
11:06 < mcallan> sure
11:06 < conseo> was pretty impressed by your smart widget composition in SacSelectionV, including keyboard shortcuts and smart css
11:06 < conseo> btw
11:07 < mcallan> (thx, huge amount of work) what to do when there is no ac selected is interesting
11:07 < mcallan> use sac?
11:07 < mcallan> then no target
11:07 < mcallan> hmmm
11:07 < conseo> yes, which is pretty bad for a progress bar
11:07 < conseo> maybe we can show a different widget with some other information
11:08 < conseo> (but this can wait)
11:08 < mcallan> you're ahead me... yes u need to go uptree i guess
11:08 < mcallan> but up which branch?
11:08 < conseo> pledges come from down the tree and targets are set upwards (or same position)
11:08 < conseo> imo this should be done in the counting engine
11:09 < conseo> it is insane to recalculate and fetch all that data in all clients
11:09 < mcallan> not suggesting that
11:09 < conseo> (but you have mentioned html5 storage already, so maybe we can even store the trees locally)
11:09 < conseo> ok
11:10 < mcallan> none of that is problem, and tree *is* already cached...
11:10 < mcallan> see here: http://zelea.com/y/vw/xf/#c=DV&s=G!p!sandbox
11:11 < mcallan> or i guess here: http://zelea.com/y/vw/xf/#s=G!p!sandbox&c=DV&a=qUS$
11:11 < mcallan> which target do you show, Az's?
11:11 < mcallan> see, no ac is selected yet
11:12 < conseo> (hehe xf crashes my irc client, because xf crashes the embedded webkit preview)
11:12 < mcallan> you only got sac (with no target) unless you default to an ac
11:13 < mcallan> *here* you get ac: http://zelea.com/y/vw/xf/#s=G!p!sandbox*1&c=DV&a=qUS$
11:13 < conseo> Votes is selected (which is a sac, right?)
11:13 < mcallan> yes
11:13 < conseo> ok and US$ is ac
11:13 < conseo> no
11:13 < mcallan> no
11:13 < conseo> it can be ac or sac
11:13 < mcallan> sac too
11:13 < conseo> ok
11:13 < mcallan> only select sac atm
11:14 < conseo> yes the selection is dependant on the position selected in the tree
11:14 < mcallan> ac, yes
11:14 < conseo> so to show an ac we need to know the position
11:15 < mcallan> well, that's an impl detail the track stack does not know - it just knows ac
11:15 < mcallan> see, there might be no votespace
11:15 < conseo> hmm, atm. i only getSac() -> SacJS
11:15 < conseo> sure
11:15 < conseo> but there are positions
11:15 < conseo> (which can be selected)
11:16 < mcallan> but you don't know about positions, only about ac
11:16 < mcallan> only what you *need* to know - that's all i am saying
11:16 < conseo> ok, but then the model has to tell me the ac(s)
11:17 < mcallan> (more flexi like that)
11:17 < conseo> sure
11:17 < mcallan> exactly, you will have ac selection model
11:17 < mcallan> not just sac selection model
11:17 < conseo> ok, sounds reasonable
11:17 < mcallan> (see this runs in diff bridge too)
11:17 < mcallan> ok
11:18 < conseo> how do i approach this model (or do you do it?)
11:18 < mcallan> i can do it if u like, tomorrow
11:18 < mcallan> it will be similar to sac, ofc
11:19 < mcallan> interesting q is still, as i say, what about when no ac is selected?
11:20 < conseo> i cannot show a progress information if i don't have a target
11:20 < mcallan> true
11:20 < conseo> it would be different, so we can think about a different visualization/widget imo
11:20 < mcallan> or default to some ac
11:21 < mcallan> altering the vis too much might be too disorienting
11:22 < conseo> yes
11:22 < mcallan> anyway, something to think about while i work on ac selection model
11:22 < conseo> if we know the registry information for a poll (like contains 5 million voters) then we can show a vote progress bar
11:22 < conseo> if it is related to a district
11:23 < conseo> (but if this is hardcoded in a poll, then it might limit different possible positions (e.g. more local ones))
11:25 < mcallan> vote progress, like turnout as here: http://zelea.com:8080/v/w/Votespace?p=G!p!sandbox
11:26 < mcallan> or here: http://zelea.com:8080/v/w/Rank?p=G!p!sandbox
11:26 < conseo> yes, we can also show slices compared to total count
11:27 < conseo> [................. 37% ..............][......20%......][.....19%......][....13%....]....
11:28 < mcallan> those are end candidates?
11:28 < conseo> yes
11:28 < mcallan> that's ok, if a target is made clear
11:28 < conseo> ehm, target in sac?
11:29 < mcallan> 'cause tracks are all about progress - not state
11:29 < conseo> hmm, i have seen visualisations like this before about state, but then it is not fitting
11:29 < mcallan> target not in sac, in ac - for Votes like all else
11:30 < conseo> ok, but we don't have a target except there is a total district number
11:30 < conseo> then it would be different, with an empty space at the end
11:30 < mcallan> vis is ok, just that purpose of track is not to duplicate what votespace already does
11:30 < conseo> sure, but votespace only shows total numbers, not percentages
11:31 < conseo> (in xf)
11:31 < mcallan> i agree, but...
11:32 < mcallan> purpose of tracks is to show progress to goal
11:32 < mcallan> so just add target, and all is OK
11:33 < conseo> how?
11:33 < mcallan> i mean, track designer can do as he pleases, but the attractor (user getter) is to show progress
11:33 < conseo> how to add a target to a poll?
11:34 < mcallan> targets are in ac
11:34 < conseo> yes, but you wanted a good default visualisation for sac/the poll overall, right?
11:34 < mcallan> i meant default ac
11:35 < conseo> i just don't know how to get a proper target. we could also try to calculate a target from all targets defined (like average)
11:35 < conseo> ok
11:36 < conseo> ok, when will you be back tomorrow?
11:37 < mcallan> choices 1) show nothing, or reduced info 2) show default ac 3) ... not sure
11:38 < mcallan> that's all i can think of, but needs more thinking
11:39 < mcallan> i'll turn on the box as soon as i get up, say in 9 hours roughly
11:39 < mcallan> ping u then
11:39 < conseo> ok, i guess i will be asleep then, but try it
11:45 < mcallan> ok, cu
11:47 < conseo> boah, a ton of discussion on metagov, yet i don't see anything new in it
11:48 < conseo> people are just arguing why we still fail and stay very abstract in the problem instead of taking a simple problem and show why we don't solve it yet
11:48 < conseo> mcallan: hmm, you have avoided that discussion
11:49 < conseo> well, not at all, but at least a big part
11:53 < conseo> cu
--- Log closed Thu Sep 22 00:00:50 2011