Alaveteli in Australia

74 views
Skip to first unread message

Alex Sadleir

unread,
Jan 11, 2012, 7:33:48 AM1/11/12
to alavet...@googlegroups.com
Greetings! I have had the Alaveteli project in my todo list for several months now, hoping that someone else would do it for Australia before me - many times people say on twitter "Wouldn't it be great if we had a website where you could make FOI requests?" and I say "who's going to set it up?"

So I started trying it out tonight - I have an EC2 account that I like to use with Singapore (Asia Pacific) but the AMI isn't on that region so Amazon was out (the US ones can be very laggy for me). I do have a VPS in the US though running Debian. However, the dependencies of this project in config/packages pulled in most of build-essential (gcc/g++ etc - probably so rubygems can be compiled locally) and lots of X11 libraries which I didn't have on a headless server. 
I don't use Apache but rather lighttpd. The instructions here seemed to work https://wiki.archlinux.org/index.php/Lighttpd#FastCGI but they don't seem very fast... but enough for an initial development server. 

Now to work out the hard stuff like how do I go about customising the site for my country/requirements without impeding/complicating future upgrades (I think the themes also include changing the views and thus the help text on the page...), how do I import agencies if the admin panel has links like https://secure.mysociety.org/admin/foi/body/list (eep!) and getting a list of agencies in the first place. Oh and more fun stuff like deciding on a domain name (I don't think individuals can register .org.au) and setting up the mail server.

Hope you'll hear from me again soon! :)

Seb Bacon

unread,
Jan 11, 2012, 7:44:37 AM1/11/12
to alavet...@googlegroups.com
Hi Alex,

Welcome to Alaveteli :)

On 11 January 2012 12:33, Alex Sadleir <max...@gmail.com> wrote:
> Greetings! I have had the Alaveteli project in my todo list for several
> months now, hoping that someone else would do it for Australia before me -
> many times people say on twitter "Wouldn't it be great if we had a website
> where you could make FOI requests?" and I say "who's going to set it up?"
>
> So I started trying it out tonight - I have an EC2 account that I like to
> use with Singapore (Asia Pacific) but the AMI isn't on that region so Amazon
> was out (the US ones can be very laggy for me). I do have a VPS in the US
> though running Debian. However, the dependencies of this project in
> config/packages pulled in most of build-essential (gcc/g++ etc - probably so
> rubygems can be compiled locally) and lots of X11 libraries which I didn't
> have on a headless server.


Great. Which version are you trying? In theory the master branch is
always at a stable release, though I accidentally commited some things
to it a while ago :) Quite soon I'll be doing a 0.5 release which
among other things has a tidied up way of handling your theming, so I
recommend using that. (You can use the release branch now, or wait a
week or so when I'll make the final release).

> I don't use Apache but rather lighttpd. The instructions here seemed to work
> https://wiki.archlinux.org/index.php/Lighttpd#FastCGI but they don't seem
> very fast... but enough for an initial development server.

That is useful to know. Keep us informed if you continue to use it
and we can add it to the docs.

> Now to work out the hard stuff like how do I go about customising the site
> for my country/requirements without impeding/complicating future upgrades (I
> think the themes also include changing the views and thus the help text on
> the page...)

Yes they do...

One recommendation about customisations: it can be tempting to
customise the workflow, the states that a request can be in, etc, to
match interesting edge cases in your own jursidiction. After launch,
the feature design principle with WDTK quickly became "design for how
it should be, not how it is". We realised that we're partly building
an advocacy platform, and could use the software to hide complexity in
the FOI system by pretending that any complexity that wasn't relevant
to the user didn't exist. Examples: don't worry about the differences
between different types of FOI law; don't stick boilerplate to the
start of the FOI request, which makes it hard to read and really
shouldn't be necessary (even if the law suggests it is); etc. You
can't always ignore things that shouldn't exist, but it's a good
starting principle.

> how do I import agencies if the admin panel has links like
> https://secure.mysociety.org/admin/foi/body/list (eep!)

That's part of the whatdotheyknow-theme which was the default theme
installed in older versions of Alaveteli. That was a bad default
theme to package for the reason you've just found!

The 0.5 version deploys the Alaveteli-theme, which is a simple theme
suitable for customising to your own needs.

> and getting a list
> of agencies in the first place. Oh and more fun stuff like deciding on a
> domain name (I don't think individuals can register .org.au) and setting up
> the mail server.
>
> Hope you'll hear from me again soon! :)

Great - stay in touch and ask lots of questions -- although things are
getting easier it's still not always obvious how to do things, and
questions will prompt me to update the docs (or at the very least,
will be archived in this list).

Also, if you've not already done so, have a read of the blog at
http://www.alaveteli.org/blog/ -- in particular the post about
volunteers, and Frag den Staat.

Seb


--
skype: seb.bacon
mobile: 07790 939224
land: 01531 671074

Adam McGreggor

unread,
Jan 11, 2012, 9:41:32 AM1/11/12
to alavet...@googlegroups.com
On Wed, Jan 11, 2012 at 04:33:48AM -0800, Alex Sadleir wrote:
> So I started trying it out tonight -

Hooray!

> However, the dependencies of this project in
> config/packages pulled in most of build-essential (gcc/g++ etc - probably
> so rubygems can be compiled locally) and lots of X11 libraries which I
> didn't have on a headless server.

Could you explain what you mean there? We run headless. The libraries
should all be in the debian repositories -- either on your localized
mirror, or via debian.mysociety.org

> how do I import agencies if the admin panel has links like
> https://secure.mysociety.org/admin/foi/body/list (eep!)

"patches welcome"...

(or follow Seb's thoughts)

> and getting a list of agencies in the first place.

We did (do) a fair bit of crowd-sourcing, in addition to seeking the
info from authoritative sources, originally

> Oh and more fun stuff like deciding on a
> domain name (I don't think individuals can register .org.au) and setting up
> the mail server.

For another project, I recently had cause to look at .AU
registrations, and my local knowledge there advised that for .org.au,
some accreditation's needed. Apparently (I don't have access to the
.AU zone), most registrations are just .com.au

For the mail server, for straight-forward compatability, you might
want to use Exim.

--
"There are three principal ways to lose money: wine, women, and engineers.
While the first two are more pleasant, the third is by far the more certain."
-- Rothschild, ca. 1800

Alex (Maxious) Sadleir

unread,
Jan 11, 2012, 10:14:22 AM1/11/12
to alavet...@googlegroups.com
On Thu, Jan 12, 2012 at 1:41 AM, Adam McGreggor <adam-...@amyl.org.uk> wrote:
> On Wed, Jan 11, 2012 at 04:33:48AM -0800, Alex Sadleir wrote:
>> So I started trying it out tonight -
>
> Hooray!
>
>> However, the dependencies of this project in
>> config/packages pulled in most of build-essential (gcc/g++ etc - probably
>> so rubygems can be compiled locally) and lots of X11 libraries which I
>> didn't have on a headless server.
>
> Could you explain what you mean there? We run headless. The libraries
> should all be in the debian repositories -- either on your localized
> mirror, or via debian.mysociety.org
As in I had uninstalled all the X11 libraries to save disk space
previously (while still using php-gd) but one of the packages
suggested pulls them in again from the usual debian repositories.
Probably one of the ones about graphics manipulation/pdfs etc. No big
deal, it was just worth noting publicly so people know you need some
disk space on your system partition for those kinds of things.

>
>> how do I import agencies if the admin panel has links like
>> https://secure.mysociety.org/admin/foi/body/list (eep!)
>
> "patches welcome"...

Yep, I changed branches to release/0.5 from master and the config file
to point to my own theme git repository. Was stumped for a while
before I realised you also have to delete the
vendor/plugins/whatdotheyknow-theme folder if you've accidently
installed it before. Still, it's a good source for the kind of FAQs to
think about.

>
>> and getting a list of agencies in the first place.
>
> We did (do) a fair bit of crowd-sourcing, in addition to seeking the
> info from authoritative sources, originally

Mmm, I think I have a list of all the agencies but just need to work
out which ones won't ever want to be contacted. I think I have to
agree with simplifying things for users (and crowd-sourcers) - there's
such a patchwork of inclusions and exclusions that sometimes they'll
just have to use trial and error.

Tony Bowden

unread,
Jan 11, 2012, 10:32:02 AM1/11/12
to alavet...@googlegroups.com
On 11 January 2012 16:14, Alex (Maxious) Sadleir <max...@gmail.com> wrote:
> Mmm, I think I have a list of all the agencies but just need to work
> out which ones won't ever want to be contacted.

Why would they not want to be contacted? Do you mean simply that some
agencies aren't explicitly exempted from your FOI law? It's probably
worth noting that in the UK we deliberately include bodies that FOI
doesn't apply to, if we think it should! They can, of course, reply
saying "We're not going to tell you that — please go away", but having
to do so in public often makes them think twice about that, and quite
often they're happy to give the information requested anyway.

Tony

Seb Bacon

unread,
Jan 12, 2012, 4:32:10 AM1/12/12
to alavet...@googlegroups.com
On 11 January 2012 15:14, Alex (Maxious) Sadleir <max...@gmail.com> wrote:
> On Thu, Jan 12, 2012 at 1:41 AM, Adam McGreggor <adam-...@amyl.org.uk> wrote:
>> On Wed, Jan 11, 2012 at 04:33:48AM -0800, Alex Sadleir wrote:
>>> So I started trying it out tonight -
>>
>> Hooray!
>>
>>> However, the dependencies of this project in
>>> config/packages pulled in most of build-essential (gcc/g++ etc - probably
>>> so rubygems can be compiled locally) and lots of X11 libraries which I
>>> didn't have on a headless server.
>>
>> Could you explain what you mean there? We run headless. The libraries
>> should all be in the debian repositories -- either on your localized
>> mirror, or via debian.mysociety.org
> As in I had uninstalled all the X11 libraries to save disk space
> previously (while still using php-gd) but one of the packages
> suggested pulls them in again from the usual debian repositories.
> Probably one of the ones about graphics manipulation/pdfs etc. No big
> deal, it was just worth noting publicly so people know you need some
> disk space on your system partition for those kinds of things.

If you're concerned about disk space, note that as the system grows,
you will be storing a whole bunch of raw incoming emails on the
filesystem, plus their parsed attachments, which are often very big.
WDTK (which has been running for about 4 years now) uses > 50G for
storing the emails alone.

There is a config setting to allow you to have the raw emails stored
on a different partition from the software.

Seb

Henare Degan

unread,
Jan 13, 2012, 4:27:45 AM1/13/12
to alavet...@googlegroups.com
On Wed, Jan 11, 2012 at 20:33, Alex Sadleir <max...@gmail.com> wrote:
Greetings! I have had the Alaveteli project in my todo list for several months now, hoping that someone else would do it for Australia before me - many times people say on twitter "Wouldn't it be great if we had a website where you could make FOI requests?" and I say "who's going to set it up?"

So I started trying it out tonight...

Hey Alex, that's awesome!

As you can imagine it's been on my to do list for ages too and I'm super-excited that I'm not the only one :)
 
Now to work out the hard stuff like how do I go about customising the site for my country/requirements without impeding/complicating future upgrades

One of the reasons I've never kicked it off is that the usual getting started spiel for Alaveteli is that you need a good volunteer base to run it.

Have a look at this blog post: http://www.alaveteli.org/2011/07/you-need-volunteers-to-make-your-website-work/

I certainly don't want to be a downer on this effort but it's something we need to work on. I think it would be good to kick off this in parallel with development - maybe fire an email to the OpenAustralia community list?
 
Oh and more fun stuff like deciding on a domain name (I don't think individuals can register .org.au) and setting up the mail server.

I'm sure we can use the OpenAustralia Foundation to pay for and register domain names, servers, etc.

My experience tells me we should have whatdotheyknow.org.au - descriptive names for these types of projects are best I think.

What do you think about using the OpenAustralia Foundation? And the domain name?

Hope you'll hear from me again soon! :)

Please let me know if you want any support setting things up.

I've taken the liberty of setting up a project in our issue tracker if you'd like to use that: http://tickets.openaustraliafoundation.org.au/browse/WDTK and created the first ticket: http://tickets.openaustraliafoundation.org.au/browse/WDTK-1

We can log issues specific to our install there and I've made you a developer so you can mess with tickets on the issue tracker. How does that sound?

I'm looking forward to bringing Alaveteli down under!

Cheers,

Henare

Alex (Maxious) Sadleir

unread,
Jan 13, 2012, 5:46:30 AM1/13/12
to alavet...@googlegroups.com
On Fri, Jan 13, 2012 at 8:27 PM, Henare Degan <henare...@gmail.com> wrote:
> On Wed, Jan 11, 2012 at 20:33, Alex Sadleir <max...@gmail.com> wrote:
>> Now to work out the hard stuff like how do I go about customising the site
>> for my country/requirements without impeding/complicating future upgrades
>
>
> One of the reasons I've never kicked it off is that the usual getting
> started spiel for Alaveteli is that you need a good volunteer base to run
> it.
>
> Have a look at this blog post:
> http://www.alaveteli.org/2011/07/you-need-volunteers-to-make-your-website-work/

I'm more concerned that nobody will actually use it!

There is a tiny number of FOI requests actually being published by
departments but supposedly they release 3,500 non-personal document
sets a year. eg. ATO got 501, there's 13 on their website. What
happened to the other 488?


>
> I certainly don't want to be a downer on this effort but it's something we
> need to work on. I think it would be good to kick off this in parallel with
> development - maybe fire an email to the OpenAustralia community list?

Yep, was on the todo. I'm still getting my head around the technical issues...

>
>>
>> Oh and more fun stuff like deciding on a domain name (I don't think
>> individuals can register .org.au) and setting up the mail server.
>
>
> I'm sure we can use the OpenAustralia Foundation to pay for and register
> domain names, servers, etc.

Along with the potential disc usage, the memory and CPU utilisation of
just the 3 test records is too much for a tiny VPS. So I could
definitely use some extra resources.

>
> What do you think about using the OpenAustralia Foundation? And the domain
> name?
>
>

> Please let me know if you want any support setting things up.
>
> I've taken the liberty of setting up a project in our issue tracker if you'd
> like to use that: http://tickets.openaustraliafoundation.org.au/browse/WDTK
> and created the first ticket:
> http://tickets.openaustraliafoundation.org.au/browse/WDTK-1
>
> We can log issues specific to our install there and I've made you a
> developer so you can mess with tickets on the issue tracker. How does that
> sound?

Great, I'll move some of the discussion over that way then.

Seb Bacon

unread,
Jan 17, 2012, 3:56:05 AM1/17/12
to alavet...@googlegroups.com
On 13 January 2012 10:46, Alex (Maxious) Sadleir <max...@gmail.com> wrote:
> On Fri, Jan 13, 2012 at 8:27 PM, Henare Degan <henare...@gmail.com> wrote:
>> On Wed, Jan 11, 2012 at 20:33, Alex Sadleir <max...@gmail.com> wrote:
>>> Now to work out the hard stuff like how do I go about customising the site
>>> for my country/requirements without impeding/complicating future upgrades
>>
>>
>> One of the reasons I've never kicked it off is that the usual getting
>> started spiel for Alaveteli is that you need a good volunteer base to run
>> it.
>>
>> Have a look at this blog post:
>> http://www.alaveteli.org/2011/07/you-need-volunteers-to-make-your-website-work/
>
> I'm more concerned that nobody will actually use it!
>
> There is a tiny number of FOI requests actually being published by
> departments but supposedly they release 3,500 non-personal document
> sets a year. eg. ATO got 501, there's 13 on their website. What
> happened to the other 488?

If you do a good bit of publicity, that will help ensure that people
use it. The OKF Germany folks did some great PR work to launch their
site: http://www.alaveteli.org/2011/11/frag-den-staat-experiences-from-germany/

We also use Google Grants to buy adwords for WDTK (keywords like
"freedom of information act", "foi request", etc), which now account
for only 2% of traffic, but were really useful at the start. These
days it's overwhelmingly our organic search long tail (fed by our
archived requests) that really drives new requests (followed by direct
visits. Adwords (but mainly people searching for "what do they know")
and then Facebook referrals also contribute a reasonable amount).

Seb

Francis Irving

unread,
Jan 17, 2012, 5:41:13 AM1/17/12
to alavet...@googlegroups.com

Facebook, and not Twitter?

In which case, perhaps should have a WhatDoTheyKnow Facebook account
instead of a Twitter account... (And it breaks my prejudices as to how
each site is used)

Francis

Seb Bacon

unread,
Jan 17, 2012, 5:56:38 AM1/17/12
to alavet...@googlegroups.com

If we restrict it to visits that resulted in a new request from new
(not returning) visitors, it's:

32% from search
32% direct
29% from referrals (including 7% from Facebook and 5% Twitter)
7% CPC (our Google Grant)

Overall traffic (not restricted by goal or visitor type) has 12%
Twitter and 10% Facebook.

I'm not sure how we can analyse where people came from within
Facebook. The referrer links are all via the Facebook l.php script.
Anyone know?

> In which case, perhaps should have a WhatDoTheyKnow Facebook account
> instead of a Twitter account... (And it breaks my prejudices as to how
> each site is used)

Well, I guess both are important given the data above. Seems like
Facebook drives more actions.

Seb

Francis Irving

unread,
Jan 17, 2012, 9:14:07 AM1/17/12
to alavet...@googlegroups.com
On Tue, Jan 17, 2012 at 10:56:38AM +0000, Seb Bacon wrote:
> Well, I guess both are important given the data above. Seems like
> Facebook drives more actions.

Ah, turns out there is a WhatDoTheyKnow Facebook page already, which
presumably Myf posts to...

https://www.facebook.com/pages/WhatDoTheyKnowcom/194119890631784

Francis

Seb Bacon

unread,
Jan 17, 2012, 10:11:41 AM1/17/12
to alavet...@googlegroups.com

Only 21 likes though -- seems unlikely it would drive that much
traffic? Or would it?

Michelle Bourke

unread,
Mar 10, 2013, 8:59:06 PM3/10/13
to alavet...@googlegroups.com, seb....@gmail.com
Hey Seb, Francis, Alex, Tony and Adam,

I know this is ages after your conversation started, but having looked at I'm from Oz and keen to be involved.  I run a marketing agency: www.artlivemedia.com and happy to volunteer to lead the marketing side of this.  It seems like you guys have the technical background?

I think there are three key things we can use to get this off the ground from a marketing perspective:
1.  Support form GetUp
2.  Viral social media campaign
3.  Requests on Volunteers.org.au + getup and other independent political organisations (including those for youth - e.g. left right think tank) for dedicated ongoing volunteers.   I know a number of individuals who would be extremely interested in being involved.
4.  The purchase of a domain name and completion of onsite SEO activity
5.  Offsite SEO: Placement of the site on a variety of directories and government/political resources to drive traffic

Would be keen to perhaps organise a skype meeting just to get a sense of where you guys are at with it - or if it's dropped off the radar, to get it back on again

Cheers,

Michelle :-)

Louise Crow

unread,
Apr 2, 2013, 7:56:34 AM4/2/13
to alavet...@googlegroups.com
Hi Michelle,

Thanks for posting to the list! Australia does indeed now have an Alaveteli instance [1], launched by the OpenAustralia foundation. You can read a bit about the latest news on the site on the OpenAustralia Foundation blog [2]. If you haven't been in touch off-list already, you can contact the team at con...@openaustraliafoundation.org.au.

Cheers

Louise


[1] https://www.righttoknow.org.au/
[2] http://www.openaustraliafoundation.org.au/2013/02/06/whats-the-latest-with-righttoknow/

Michelle Bourke

unread,
Apr 3, 2013, 1:49:30 AM4/3/13
to alavet...@googlegroups.com
Thanks Louise :-)

Michelle Bourke 
0400 157 988

Sent from my iPhone
--
You received this message because you are subscribed to a topic in the Google Groups "Alaveteli Dev" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/alaveteli-dev/kbckno4Mxy8/unsubscribe?hl=en.
To unsubscribe from this group and all its topics, send an email to alaveteli-de...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply all
Reply to author
Forward
0 new messages