Atom 1.0 format is Recommended

The new Government recommendation is to use the Atom feed format.

• Atom it is an official specification managed by the IETF
• Atom uses the ISO 8601 date format
• Atom provides a mechanism to explicitly and unambiguously label the type of content.

The new format is fully understood by all common feed readers and aggregators, something that is not the case with the previous specification which was a customisation of RSS 1.0.

In addition to the standard Atom element structure, the prefix ‘nz’ is defined as an Atom extension for New Zealand specific elements.

nz:dtend is defined as an optional custom element for publishing the closing-date for consultations and jobs.

A value list of categories/tags is defined: nznews, nzjob, nzconsult, nzevent.

Resources

• Download the draft 2008 New Zealand Government Feed Standard
      in Open Office .odf format or Microsoft .doc format
• The Atom specification is at http://www.ietf.org/rfc/rfc4287.txt
• Previous RDF (RSS) e-GIF standard (2003)

Examples

We have provided a number of example feeds at http://eg.elabs.govt.nz/feeds/

• Example Atom feed
• Example complex Atom feed
• Other examples

Example Atom feed code

Here is a code sample showing the recommended structure.

The example feed viewed in Internet Explorer 7

Consultation

The consultation period runs to 5pm, 13th August 2008.

Please post comments here or on the Government Web Community workspace.

Download the presentation Feeds for Thought in Open Office .odp format (3.5G) or Microsoft .ppt format (3.6G)

A global IP address is needed for anyone who wants make a web server or any other service available on the internet. The predictions are that the world will run out of IPv4 addresses in about four years (given current consumption). IPv6 introduces a far larger address space, a different syntax or way the addresses are constructed, simplified routing and adds increased security. Our goal is to set up a small isolated IPv6 network, then to connect this to the internet and test what works and what doesn’t.

In part one I attempt setting up the isolated network with various operating systems, wireless and DNS, then testing a few applications.

Phase 1: Create a self-contained simple IPv6 network

Step one: Equipment

• Recently vacated desktop (P4) - installed Debian 4 dual booted with Windows XP
• D-Link DI-524 wireless router
• OS X Leopard Macbook (built-in wi-fi)

Step two: Network this equipment together

I created a network with the Linux box connected to a D-Link DI-524 wireless router. The Mac laptop is connected to the wireless router via wireless
Linux box <……cable…..> D-Link router <- - - - wireless - - - -> Mac laptop

I statically set IPs to Linux box at 192.168.0.2 and Mac laptop at 192.168.0.3, D-Link router is 192.168.0.1. Time to check if pings work:
Linux box from Mac - success
Mac from Linux box - success

Step three: See what works IPv6-wise straight off

Try IPv6 pings:

(localhost address in IPv6) on Mac - success

on Linux box - success

So far so good!

Ran ifconfig on Linux box to check what IPv6 address the ethernet device had - fe80::208:2ff:fefb:30db so I tried pinging that - failed (unknown host)
Ran ipconfig on Mac to check what IPv6 address the wireless device had - fe80::21b:63ff:fe04:7cf1 so I tried pinging that - failed (No route to host)
tried cross-pinging - failed again

Time to look up configuring IPv6 on debian. I came accross http://people.debian.org/~csmall/ipv6/setup.html and found I might need to add this to /etc/network/interfaces:

this didn’t seem to do much

Unfortunately via Google I found out the D-Link (Firmware version V2.03 dated Fri, Dec 09 2005) doesn’t really do IPv6 so we’ll continue on anyway and see how far we get.

I then tried adding IPv6 to /etc/modules on Linux box.

Aha! help from http://lists.debian.org/debian-ipv6/2005/06/msg00051.html suggests the real way to use ping6 (on Linux box ping itself):

- success

now to try pinging the Mac from the Linux box:

- well well, success - and with ping times that suggest it is actually working. Realised I don’t actually need the “iface eth0 inet6 static” section in /etc/network/interfaces on the Linux box because it works automatically without it

on the Mac ping itself:

- success

ping the Linux box from the Mac

- success

I am assuming this is not a true IPv6 network as the router is still only talking IPv4 so all packets that are passing the router are IPv4 and in the case of IPv6 pinging, we have IPv6 encapsulated in IPv4?

Step four: Test applications

SSH

Installed sshd on the Linux box and tried connecting from the Mac
Normal IPv4:

- success (it asks me for a password)
Now IPv6:

- failed (No route to host)
hmmm trying the ssh man page… maybe we need to force IPv6 addresses with -6:

- failed (No route to host)
just checked

and that still works

At this point I decided to replace the dlink router with an apple airport express/basic 4 port switch to see if that was any better for IPv6. Unfortunately not the configuration mentions nothing of IPv6.

Due to the lack of IPv6 support in the wireless routers, I’m going to do away with the wireless side for now and just use a small 5 port switch to connect the Linux box and Mac together - I’ll get the Linux box to do DNS IPv6 style in the future.
First configure Linux box and Mac to work normally over IPv4 - modified the Linux box /etc/network/interfaces to say:

And set the ethernet port on the Mac to 192.168.0.2 subnet mask 255.255.255.0
The Mac now as IPv6 address of fe80::217:f2ff:fede:83e8, Linux box is same as before

IPv4 and IPv6 pings to the respective address local and between each other work fine.

Now we have IPv6 connectivity established, lets put Windows into the picture. Connected up a Windows Vista laptop and ran:

- told me it had an IPv6 address of fe80::2953:ea4:d099:a29c

Pinging from the Vista laptop:

- success pinging the Linux box

- success pinging the Mac box

Try pinging the Vista laptop from the Linux box and from the Mac laptop - this failed. I went into Control Pannel/Network and Sharing Center and turned on Network discovery and File sharing. Now pinging the Vista laptop from the other from the Linux box and the Mac works.

Web broswing

Now to try web-browsing to the Linux box (which has apache running) using its IPv6 address.

I tried this firefox on the Linux box itself http://[::1]/ - success.
I then tried http://[fe80::208:2ff:fefb:30db] - failed (it didn’t even attempt to find the page)
Tried this on the Mac laptop on both Firefox and Safari http://[fe80::208:2ff:fefb:30db] - failed (Unable to connect message)
Tried this on the Vista laptop - http://[fe80::208:2ff:fefb:30db] - success!

So I need to figure out why I can’t goto the Linux box’s IPv6 IP on the Linux box or Mac. I downloaded opera for the Mac and still no luck browsing the the IPv6 address of the Linux box. It seems this might be an OS thing which is strange because I can ping these addresses, but not web-browse to them.

I edited /etc/hosts on the Mac and added the line:

Then tried web-browsing to linuxboxv6 - failed
Tried

- failed
changed the line in the hosts file to:

Now web browsing and ping6′ing linuxboxv6 (from the Mac) - success!
So that means on the Mac even though I can’t web-browse to the IPv6 IP address, I can put the IP address into the hosts file as above (must append %interface_name - en0 or whatever) and specify a name then web-browse or ping6 that name.

Lets try applying the same concept to the Linux box.
Added this line to /etc/hosts

- failed

And on the Vista laptop changing c:\windows\system32\drivers\etc\hosts

- failed

Time to try setting up an ipv6 DNS server on the Linux box. From this article http://www.linuxjournal.com/node/6541/print it looks like BIND9/named is the way to go.

Spent a lot of time trying to configure BIND9 to map names to the v6 addresses. Followed various tutorials on the net with not much understanding or luck.

Decided to get BIND9 to map names to the v4 addresses of the machines then try adding in v6 BIND stuff. I roughly followed this guide for the ipv4 DNS server setup http://www.ubuntugeek.com/dns-server-setup-using-bind-in-ubuntu.html

I have the DNS setup like this (A records):

and told all machines that their dns IPv4 server is 192.168.0.1 - success (can ping the IPv4 names as per above from/to all three machines)

Now I add the following (AAAA or quad-A records):

I told all machines that their dns ipv6 server is fe80::208:2ff:fefb:30db - success (can ping the xxxxv6 names as per above from/to all three machines)

Lets try web browsing using the IPv4 and IPv6 names to connect the apache server on the Linux box:
http://debianv4 on the Mac - success
http://debianv6 on the Mac - success
http://debianv4 on the Linux box - success
http://debianv6 on the Linux box - success
http://debianv4 on the Vista laptop - success
http://debianv6 on the Vista laptop - failed - hmmmm! - I’m reading http://technet.microsoft.com/en-us/library/bb727035.aspx to see if that explains anything

Now I want to try the configure BIND on the Linux box to advertise itself as a IPv6 DNS server - so I don’t have to manually put in fe80::208:2ff:fefb:30db as the DNS IPv6 server to every machine on the network.

http://people.debian.org/~csmall/ipv6/setup.html appears to suggest radvd for router advertising.

Installed radvd on the Linux box and spent lots of time reading about how it might work.

Based on http://dufo.tugraz.at/~prokop/ipv6/ipv6-linux-routing.pdf
I edited the /etc/radvd.conf file to contain:

and restarted radvd.

I removed the DNS server fe80::208:2ff:fefb:30db from the Vista laptop’s IPv6 network config, then disabled/re-enabled the network interface, then I checked network connection details and good news:

• I saw a Property called Site-local IPv6 Address fec0::1:2953:ea4:d099:a29c%1 appear
• the IPv6 Default Gateway had been automatically picked up as fe80::208:2ff:fefb:30db

Unfortunately it says the IPv6 DNS Servers are fec0:0:0:ffff::1%1, fec0:0:0:ffff::2%1 and fec0:0:0:ffff::3%1 - was hoping it would say fe80::208:2ff:fefb:30db

So it looks like radvd has done gateway advertisement and prefix advertisement but not advertised the DNS.

Summary

We have covered the equipment used, experimenting with wireless, setting up a stand-alone network with IPv4 and IPv6 capabilities and using a Linux box to perform DNS.

Issue encountered so far:

• Network not wireless capable due to lack of support from tested wireless routers
• Can’t web-browse to IPv6 IPs on Mac or Linux box
• Can’t web-browse to IPv6 DNS name on Vista laptop
• DNS clients not automatically picking up IPv6 DNS server address

My next steps are to search out how IPv6 tunnels work and do some further digging around on the above issues. Stay tuned for part two!

New Zealand addresses do not follow the US model of street / city / state. The typical New Zealand format is street / suburb / city or for rural addresses street / RD number / mail location.

This typical New Zealand usage does not map well to hCard format (as well as vCard on which it is based). To reflect this New Zealand format, we recommend the following mapping of NZ address elements to hcard elements:

The following is a example of hCard markup:

<div class=”vcard”>
 <div class=”fn org”>State Services Commission</div>
 <div class=”adr”>
 <div class=”street-address”>100 Molesworth</div>
 <div class=”locality”>Thorndon</div>
 <span class=”region”>Wellington</span> <span class=”postal-code”>6140</span>
 </div>
</div>

The hCard creator tool is at:

http://research.elabs.govt.nz/resources/hcard/hcard-creator-0.1.html

If you have any comments or suggestions or have seen a similar solution to the below please leave a comment (no registration or login needed).

The Basics

To enter a macron into web page HTML, you can use the following HTML enities, I’ve also added unicode and UTF-8 for completeness.

Web Content Editors

But what about if you’re not directly creating HTML content but rather entering it through a web application (content management system, blog, wiki, etc.)?

Most content creating web applications use editors, so the easiest way to add macron support is to add a button or keyboard shortcuts to the editor. To test this idea I wrote plugins for TinyMCE, a widely used WYSIWYG editor and for DokuWiki’s built in editor.

TinyMCE Plugin

I wrote a plugin to TinyMCE which adds a new button and some keyboard shortcuts for adding macrons. See it in action below. The plugin is available on SourceForge (the downloads are at the bottom of the page under ‘Attached Files’), extract it to the TinyMCE plugin directory to install. You can also view the page source of the demo below to see how it is included.

Enter text here
To add a macron click on the right-hand ā button or use the keyboard combinations:

• Ctrl + Alt + <vowel> - (lowercase)
• Ctrl + Alt + Shift + <vowel> - (uppercase)

DokuWiki

To add a macron button into DokuWiki:

1. Download userscript.js, macron.png, and icons.tar.gz (or if you prefer icons.zip)
2. Place userscript.js in your dokuwiki conf directory (if you already have a userscript.js append the downloaded code to the end).
3. Place macron.png into you dokuwiki lib/images/toolbar directory.
4. Uncompress the icons into your dokuwiki lib/images directory.

For more information on the DokuWiki toolbar see http://wiki.splitbrain.org/wiki:tips:toolbarbutton.

HTML Forms

For users to enter macrons into HTML forms to submit information I developed a very simple JavaScript toolbar (it uses the JQuery JavaScript library). See below for a demo, click in a text box and then on a macron from the toolbar.

The current version is lacking some features needed in a production website but it should still provide a good example of what can be achieved with the technology. Current features missing:

1. No Graceful Degradation - it should degrade into a plain text macrons which can be copied and pasted if JavaScript is disabled or unavailable.
2. The macron is only appended to the end of the text in the text field, which is usually OK but should instead insert at the cursor position.

The Australian Open Source Industry Community Report 2008 says “…conservative projection of earnings suggests that the Open Source industry generates $500 million in revenue each year, with over 50% of that being directly related to Open Source”

It goes on to say “…around half of the Open Source industry respondents service export markets (45%)”

Should government be more actively stimulating the local Open Source industry?

See the Australian Open Source Industry Community Report 2008 [1.5MB] in PDF.

Or view converted to HTML format by Google.

When the publisher wishes to publish a new article, the feed is simply updated with the new item. This may seem like a complex task for web developers and content managers of websites, but many Content Management Systems (CMS’s) allow feeds to be automatically generated. To see the new items for a feed, one must subscribe to the feed by adding its web address to a Feed Reader (sometimes called an aggregator). Feed Readers periodically check the subscribed feeds for new items. When you open or log-in to your Feed Reader, it will have a way of displaying the new items from your feeds.

Here is an example of a Feed Reader called Bloglines. The 6 feeds that have been subscribed to are displayed on the left and articles on the right

Why not subscribe to email updates or just go to the website?

Consider you had thirty websites that you wanted to check for content updates. It is true that you could bookmark those sites in a web browser and check them every day. That would, however, be time consuming by having to open each site, click around and manually scan for changes in content. Even if those sites offered email updates, firstly you would have to give your email address out to them - increasing the risk of spam should they pass your email address on, and if you changed your email address, you would need to go back to each site and update them with your new address. Secondly the email updates are mixed in with all your other emails making it a bit of an effort to find old articles.

With web feeds, just visit each site once, identify and add the site’s feed to your Feed Reader, open your Feed Reader daily, scan article headlines and read only the articles that interest you.

Feed Readers

There are three categories of tools that allow you to read feeds:

Web feeds in Government

Reading web feeds

If your desktop is locked down such that you can’t install a Feed Reader application and your web browser/email program does not have the ability to read feeds (this is common with a Windows XP/Office 2003/Internet Explorer 6 setup), then the easiest option is to use a web-based reader.

Some web feeds in Government

• In Development Posts
• Sustainability news
• NZ Govt news

Advanced uses of web feeds

Web feeds can have more advanced uses than just reading new articles. Just a few examples are:

• watch for new video posts by your favorite Youtube user
• see comments posted about an article or blog that you are interested in
• watch for content changes in a wiki page
• be notified when someone sent you an event invitation in Facebook
• be notified when there are new items for your search at Craigslist
• watch for must-have items on Ebay using RSS Auction
• learn a new word every day from Dictionary.com
• watch for events added to a public Google Calendar
• follow your favorite comic strips on Tapestry Comics

Summary

Web feeds are a convenient way for websites to push content to you rather than you having to do all the browsing and clicking around. Feed Readers give you one place to go to keep track of all the articles that are of interest to you. Since subscribing to feeds only requires you to enter a web address into the Feed Reader, there is a danger of collecting too many feeds and spending all your day reading them! It pays to frequently assess your collection of feeds and delete ones that are no longer relevant to your interests. I suggest trying one of the readers in the above list - web based readers are the easiest as they don’t require installation of any software.

Web feeds are becoming increasingly popular as a way of delivering content to users. Once you have tried web feeds you will never go back to email subscriptions - that’s for sure!

You can view the discussion at http://webstandards.govt.nz/index.php/Talk:Handling_404s.

Our original error page wasn’t awful - it kept our page look and feel and had the site navigation down the right-hand side. But it gave no help to users on how to find the page they were looking for.

To improve the page I followed Ben’s suggestions and added some instructions to help visitors find what they were looking for. I also removed the 404 error as this is irrelevant to users trying to find a page.

To check out the new and improved 404 page see http://research.elabs.govt.nz/nonexistantpage.

In addition to the instructions on the page, we’ve made it smarter by attempting to guess where the user was trying to get based on the URL. Inspired by this Housing New Zealand Corporation 404 page we set up a similar page where a site search will be performed from the page name in the URL and results displayed under a ‘Suggestions’ heading.

See some examples in action:
http://research.elabs.govt.nz/research (searches for ‘research’)
http://research.elabs.govt.nz/category/web-service (searches for ‘web service’)

The algorithm is very simple:

1. get the page name from the URL e.g. http://research.elabs.govt.nz/somepath/somepage.ext
2. replace any underscores or dashes with spaces
3. decode any encoded characters e.g. %20 will become a space
4. search and display under suggestions (if there no results nothing is displayed)

If you have any suggestions, improvements, or examples of better practice please leave us a comment.

See http://blog.privcom.gc.ca/index.php/2007/11/05/a-friend-of-a-friend-of-a-friend-knows-youre-on-vacation/

Work has concentrated on developing practical governance concepts around devolved stewardship. How to make sure you have the right version of a schema, who decides what that is and who’s responsible, how to agree on best practice in a fast changing environment and so on.

A reference group have developed three key governance documents that are out for review and feedback. The principles document sets out the essential features and rules for managing and sharing technical resources based around common standards. The operations guide sets rules for practical governance including the use of supporting tools such as the pilot object registry being run by State Services Commission. The registry can be viewed here: http://registry.e.govt.nz. The Communities Guide attempts to define the main technical communities of practice that need to be involved in governance arrangements and that need to share technical information.

The consultation process is being run by Standards New Zealand who will receive and collate all the feedback. To view the documents and provide input please use the following link:

http://www.standards.co.nz/web-shop/?action=browseDrafts&draftTypeId=3&mod=drafts

The consultation period ends on Friday 9 May 2008, after which the documents will be updated and issued under the New Zealand e-Government Interoperability Framework.

Even if you’re not interested in making submissions, at least take a look at the documents - they’re short. Eventually these concepts and practices will make a very real difference for people working on government systems. The ability to easily find, understand and use the interfaces between systems in an all-of-government context is the key to efficiently implementation.