Wednesday, June 17, 2009

Role of the web in bushfire warnings

The 2009 Victorian Brushfires Royal Commission is addressing the issue of the role of the web in providing warning to the public. Professor John Handmer, author of "Handbook of Disaster and Emergency Policies and Institutions", gave evidence on 16 June 2009. The statement is not yet online (the commission secretariat told me they have some "technological issues" with statements at present), but the Transcript of Proceedings is. Below are some excerpts dealing with the web and Internet. I agree with the general approach suggested by Professor Handmer, but would like to see simple efficient web mark-up used for warnings, rather than plain text.
You note in paragraph 16 that the audience for a warning may be hugely variable and towards the end of that paragraph you note that, "People go to different sources. Some community members may be habitual uses of the internet, others might be more likely to turn to the radio, others might use personal networks. There are different preferred modes of receiving information." How does that then impact on the way that one should take care to disseminate warnings?---Ideally - I mean the community at risk is infinitely diverse. Each individual, we could argue, has a unique preferred way of receiving a warning, but at some level we have to stop, I suppose. But ideally the modes that are the preferred ways for that community at risk to receive their information should be the modes that are used, given whatever is practical, and that means, almost always it means that there would be several modes.

So it would be preferable in your view to use the internet as well as ABC Radio and perhaps even give consideration to other modes like phone calls or Twitter sites?---Yes, that's right. They are all reasonably technological means. One could argue that in many communities to ensure that the more vulnerable people - it depends on the community - are reached, we would probably need to get into the local networks, the personal networks or the community networks to try to activate, if you like, the neighbourhood to make sure that people who may not receive warnings via those modes receive them either by direct personal contact or some other way, and that they make sure that they are in a position to take what sort of protective action is needed. But this is tapping into what we call the informal warning system. Is there another benefit to disseminating by more than one means, namely in case of failure of one means or imperfect delivery of one means during a crisis?---That's right. We would argue that reliance on any single mode of dissemination is pretty risky, partly because it is not going to get to everybody no matter what it is and, secondly, any single mode is subject to failure or congestion or interruption.

The next aspect you turn to in your statement is timeliness and you note in paragraph 17, "A warning should be delivered in a timely manner so as to allow people to confirm what they have to do and take action in time." Is that a feature you have noticed in your research, that people usually seek confirmation from further sources before they act?---There are two things that come out of the research, main things. One is what you have just said, that people will almost always seek confirmation. Officials will, too. But people at risk will seek confirmation usually by mobilising their personal networks or if they hear something, read something on the web, listen to the radio or TV or ring somebody or vice versa. This is pretty normal and we have found often people - they also might want to ascertain the location of other household members. There are a number of things go on typically before people take action. The other thing we have noticed is that very frequently people receive the warning or at least understand that the warning is important to them too late to do anything useful. ...

Websites. Can we go to question 5, which starts at page 0018, and you note in paragraph 67 that web-based material has really become the primary source of information in our society. In paragraph 69 you make some points about who uses the internet. You say that even though it seems ubiquitous, in 2006 about a quarter of Victorians didn't have internet access. So, although that is a declining proportion, that needs to be kept in mind. That comes from the census data, is that right?---That's right. So it remains the case that the web is not a fix all. One would need to keep in mind promoting messages through ABC Radio and other means?---That's right. The point there is that a proportion of households, and they are likely to be people who are more vulnerable, elderly people and so on, do not have web access. It is also an interesting thing that people who promote the web as a vehicle for warnings have an implicit assumption that people are out there actively seeking their warnings on the web. We don't have evidence for that.

That's an important point you make at point 3: "Websites offer a passive form of warning. That is, they don't alert you to come and read them, although you will find the message if you go and look for it"?---That's true. There are a variety of ways of overcoming that and making websites active through all kinds of tools that can send the messages to you now, Widgets, Twitter and so on. But, nevertheless, the basic principle is that a website is a passive form of warning.

It could be used in conjunction, though, couldn't it, with those other tools you mentioned. If there was a SEWS signal played on the radio or an automated phone call or a text message, part of which suggested looking at a website, that might combine the call to action with finding more information on the website?---It could, or it could simply be that the material on the website is sent to your mobile phone or whatever by one of these devices and there are several possibilities with that.

You note over the page on 0019 some issues about currency and reliability and the issues which may arise when a website is under heavy demand. We touched on this when you spoke of your own experience on 7 February. Is there a way to address the situation when websites are under heavy demand and therefore slow down or even become inaccessible?---They tend to slow right down, that's right. There are a number of ways of addressing it.

Probably the simplest way is for people to take the information off the site automatically and feed it onto other sites or other systems. In the fires on February 7th the material from the CFA site was re-posted, if you like, via Twitter. There was an unofficial site, CFA updates, which was a Twitter site, and that is still active, actually. That was one of a number of sites that on the day took material unofficially from the site. There is a way of doing it which is quite legitimate and CFA encourage it. So, that's one way. What that does is take the load off the site. Another way is to ask people not to use it or to restrict access, but that doesn't seem very promising to me, given that we actually want people to use it, but that's a standard response. Otherwise, there are a number of technical ways of doing this which I outline in the paper. They are basically about reducing the degree of interactivity with the site, so that when you go into the site you don't actually - what you get is just sitting there. The amount of processing power that site needs to use is limited one way or another. Things like graphics, logos and so on, which we have more and more of them on our sites, are pretty hungry for memory.

The idea is not to use them in these emergency situations. In one sense it is an argument for moving to a different website mode in a major emergency when you know the demand is going to be great. I don't know whether I mention it here, but after the tsunami the British Commonwealth and Foreign Office or Foreign and Commonwealth Office website on travel advisories and so on switched to a text only mode for precisely this reason.

And that reduces the memory use?---That's right. It can handle a lot more inquiries.

I note in paragraph 72 you suggest, if we just deal with websites bit by bit, you suggest first of all that it would be useful for there to be one website rather than the DSE and the CFA websites?---A lot of people are arguing this, that there should be one website, but it is a trade-off, I want to say, as well, because if there is one website, all the problems we are talking about in terms of website overload and so on are exacerbated. The solution of course is that there are two sites but they mirror each other's content.

So two sites with the same content or multiple sites with the same content may help?---Yes. I think a single site in terms of content is the ideal, but if we look at the practicalities and the reliability, we are much better off having a number of sites.

Is there also potential to enable information within a website to be hived off, namely to enable people to look at particular messages pertaining to particular parts of Victoria so that they are using different pages or different information at the one time?---Yes, there are a range of devices and so on that can be embedded in sites to do that, and even to send them to the people concerned. You set out all these matters working through to paragraph 80 in the statement. Paragraph 77 is where you deal with the RSS feed. This is the capacity you spoke of for the material on an internet site to be mirrored, if you like, over on a Twitter site?---Yes, but not quite. The RSS feeds really just take key information. They don't take the whole information of the site. That is one reason why they can actually feed information on to sites like

Twitter or even mobile phones if the system is enabled. They take headliners, basically.

Dealing with sirens, which is question 6 - - -

COMMISSIONER PASCOE: Before we leave the websites, a question about the Bureau of Meteorology site which had, we are told, 70 million hits on the day and is used to having a massive - - -?---It is the most popular in Australia, I think, the most popular government site.

I don't know whether you have looked at the features of that site and what enables that site to cope with the heavy demand vis-a-vis the sites that we have just been talking about and whether there are any lessons we can learn from the bureau website?---I'm sure there are, but I haven't personally investigated them, but a lot of the bureau's material is in very basic text form and I think that's probably one of the key features of enabling that site to handle such loads. But I think that would be a worthwhile. I think it is the fourth most popular site in the country. ...

Turning to new technology, question 7, this is a matter you discuss in paragraphs 91 onwards and you refer to the new technologies which have emerged. You make the point in paragraph 93 it is important not to overlook our longstanding communication technologies, including radio. In paragraph 95 you say that it is important to distinguish between new technologies that deal with the centralised systems, such as CAP, and those that relate to individualised information. I take it from what you say here there is certainly a role for new technologies to play and it is a field that continues to develop?---I think the new technologies, in terms of delivering a message, as we were discussing, to the people at risk, have only very recently started to play a major role, but it has been quite quick and now most people in our society, I would say the majority of people by far use either a mobile phone, text, are very familiar with texting and the internet as their normal means of gaining and sending information or whatever. So we have to use them if we want to reach particular audiences and there are many variations of those modes.

Because you mention in paragraph 98 Facebook sites that are mostly post-fire, but Facebook sites, MySpace sites and in paragraph 99 the Twitter site as new technologies being used by portions of the community that ought not be overlooked?---That's right. Some of these played a role, like Twitter sites, in warnings. There is anecdotal evidence that people got warnings on Facebook because they were looking at some aspect of Facebook and suddenly some message came across. But people weren't using Facebook, as far as I can see, for warning purposes but it fulfilled that role.

At paragraph 100 you refer to phones and mobile phones and you make the point obviously they are very familiar. For landline phones, about halfway through paragraph 100, you note the technology which enables locations connected to landlines to be selected which could be used to delimit areas. That might be useful, for example, in any automated phone warning system?---Yes. That's the idea, yes.

You point out the advantages, but also the disadvantages. There may be lack of mobile phone coverage, there may be issues with phone traffic?---And there is a privacy issue with unlisted numbers and so on. But, yes.

Are you familiar with the recent announcement by the Commonwealth government to now establish a national phone automated warning system?---Yes, I am familiar with that. You refer to the common alerting protocol. It, as you mention there, is really a mode of standardising the content of warnings to ensure that it is the same over different modes of dissemination?---Yes. The common alerting protocol relates to what we were discussing a while ago, the write-it-once concept. As you say, it is a standardised message, it has a standardised format and then the idea is that this message can then be disseminated over any number of digital modes. So it has that advantage of speed and also has advantages in being able to go on multiple modes that perhaps would have to be manually uploaded in the past. ...

From: Transcript of Proceedings , 2009 Victorian Brushfires Royal Commission , TUESDAY 16 JUNE 2009, 24th day of hearing

No comments: