Wire: Interview with Icarus Sparry

What do do you?
My job title is ‘Computer Officer’, and I do everything connected with computers. I am essentially third line support, as I was not brought up to tolerate fools.

How did you get into this job ie what did you do before?
I used to work in Electrical Engineering, working on data transmission systems over mobile radio links.

Firewalls in Universities - essential security feature or inconvenience to open use of networked technologies?
Unfortunately they are an essential security feature - the Internet has 2 million plus hosts connected to it, and as I don’t know the administrators of each of these hosts personally, and therefore am unable to vouch for their good motives, we have to protect ourselves.

Firewalls also help implement policy. At Bath it has been decided that using MUDs is not an appropriate use of our equipment. There are notices to that effect, but without a firewall they are just so much wallpaper. With a firewall the access can be blocked.

Firewalls also help people configure machines correctly or in a more sociable manner. All Bath WWW accesses go through a local cache, which helps reduce international bandwidth requirements. However none of the popular browsers come with a setup file designed to directly fit our needs. Without a firewall someone could just install say Netscape, and start using it. They would get worse performance, and use more bandwidth than if they use out cache. With the firewall in place, it just doesn’t work, so they have to make the effort to find out about the cache.

Home pages - should students be allowed to use time and network resources (is there any significant impact) on creating and mounting their home pages? And should staff, for that matter?
Well, for better or worse, the World Wide Waste of time is here to stay, at least for the next few years. It is a very useful source of information, and even more so with search engines like AltaVista. The problem is that too many people put up information and then do not spend time keeping it up to date. The support implications for students creating home pages is fairly small, so there is no reason why they should not be able to do so. As search engines like AltaVista exist, staff can justify putting time into creating pages as it can lead to contacts and ultimately research contracts.

Networking charging - various debates are ongoing onto whether network access charging schemes for universities should be changed - your views?
I think it would bring back a sense of reality to the use of the network. While there is nothing wrong with genuinely spare bandwidth being used for recreational use, it currently impossible to do anything sensible using the transatlantic links in the afternoon. Because of this, and the important National services based at Bath, we have invested in a link to Pipex. Unfortunately this is expensive (12k per year for a 64k line), but it does deliver the needed bandwidth. Obviously it is better to buy bandwidth in bulk, but until we have technologies like RSVP in place to make sure that we get the bandwidth we are contracted to get, then there is little to be done. Of course, if the Universities paid 10 pounds per month per student then the resulting money would solve almost all of our networking problems!

What Web browsers do you use, and why?
Lynx, mosaic, telnet and Netscape. I try not to download the silly pictures that people tend to clutter their pages with, and Lynx works well for that. The choice of telnet may surprise a few people, but it is a nice, clean way of getting textual information, and also shows me exactly what is going on. I use mosaic if I need to look at the pictures, and Netscape if I have to look at some Netscape extended site, but I tend to discard those on the basis that if they are not prepared to make the information available to as wide an audience as possible, then they probably don’t put much effort into the data either.

How significant an impact do you think that mobile code, such as Java applets, will have on the development of the Web as an information resource?
I think it will have very little impact. Obviously it means that you can have nice pretty moving pictures on your screen, but I have them turned off anyhow, and more usefully you can have forms verified locally, rather than having to transmit the faulty data. The network computer idea, which should enable state of the art small machines to again become less than 500 pounds (as the BBC micro and the Spectrum were) is obviously a good. Having machines which you just plug in, and switch on will make administration simpler. Byte-coded systems are a good idea, but more work needs to be put into the security design. Currently ActiveX says allow anything and relies on the producers of the code signing it to ensure that it is correct. Java allows you to do almost nothing by default, but users can enable it. This opens up a whole range of social engineering attacks.

Javascript - what’s the point? Is there one?
Javascript only confuses the issue w.r.t Java. Too many people do not know there is any difference.

Pornography. Much has been made, often hysterically, in the mass media, about pornography on the Internet, especially on the Web and Usenet. What steps should or could academic institutions take to prevent students from choking up the networks by accessing such material
Nothing can be done to absolutely stop it. Reasonable steps need to be taken to reduce the risks that people inadvertently stumble across it, so banning people from displaying it on screen in public places is reasonable and proper. Informed adults should be allowed to view what they wish, but owners of equipment have a right to say what their equipment should be used for. So if a student wants to buy a CD full of pictures of Cindy Crawford wearing little or no clothing that is fine, but don’t use our computers to view it, or our network to collect it - a picture is worth ten thousand words.

One wish for what you would like to see happen in the world of networking and the Web in the next year…
HTTP over TCP over IP to be dropped as a protocol, in favour of something that uses lighter weight connections and a hierarchy of caches.

Interviewee Details