[wdmmg-dev] New version of the explorer using bubbletree

[Rufus Pollock]

Moving on-list ...

On 22 June 2011 10:33, Gregor Aisch <gka at vis4.net> wrote:
> Hi Rufus,
>
> I'm absolutely fine with your refactoring suggestions. Will refactor the
> examples and throw out the unnecessary stuff once you're finished with the
> /lib/aggregator.js, ok?

I wanted *you* to factor stuff out into lib/aggregate.js since you
know where the code is in bubbletree that calls out to the aggregator
and then converts to tree format ...

> If you still want to have some of the configuration vars inside the hash url
> (eg the drilldown array), we could define a clear url specification that
> helps the explorer and the visualizations to correctly interpret the urls.
> For instance, we could define some kind of separator char like '~' that
> devides the hash url in two parts
> ...explorer/#/2009/cofog1/cofog2/~/grand-total/health/

seems good. I wonder if we could reuse some of the format from the
aggregate api.

> This way, both the explorer and the viz would still have the full
> flexibility on how to build and interpret their urls.

Agreed.

Rufus

> Am 22.06.2011 um 11:15 schrieb Rufus Pollock:
>
> Did you see my email below? Any response, especially on items directed
> to you? I'd be grateful in your assistance in doing the suggested
> refactor (if you agree).
>
> Rufus
>
> ---------- Forwarded message ----------
> From: Rufus Pollock <rufus.pollock at okfn.org>
> Date: 20 June 2011 10:50
> Subject: New version of the explorer using bubbletree
> To: Gregor Aisch <gka at vis4.net>
> Cc: wdmmg-dev at lists.okfn.org
>
>
> Hi All,
>
> I've extracted the dataset explorer into a pure js app:
>
> <http://demos.openspending.org/openspendingjs/app/explorer-v2/>
>
> You'll need to select something from the drop-down (e.g. cofog1) and
> hit redraw to see anything (and wait for a long time).
>
> Next steps:
>
> 1. Reintegrate this back into openspending.org/dataset/explorer
> 2. Allow for provision of default drilldown options
> 3. Have year slider
>
> Some thoughts on the bubbletree code (@Gregor primarily):
>
> * Bubble tree should just take a standard tree struture
> * All code to get that tree (whether off-disk, from openspending API)
> should be separate (and outside of bubbletree)
>
> Hence:
>
> * Have a single option to bubble tree: data (or similar)
> * Get rid of all other data options (if we really want we can keep
> some helper stuff in Loader but let's make simple data provision
> primary).
> * Move getTree functionality into ./lib/aggregate.js (or similar)
> * Refactor examples
>
> Regards,
>
> Rufus
>
>
>
> --
> Co-Founder, Open Knowledge Foundation
> Promoting Open Knowledge in a Digital Age
> http://www.okfn.org/ - http://blog.okfn.org/
>
>



-- 
Co-Founder, Open Knowledge Foundation
Promoting Open Knowledge in a Digital Age
http://www.okfn.org/ - http://blog.okfn.org/