Jon: We're both "techies" on the ROADS project in eLib and the EU DESIRE project (which for us is basically "EuroROADS"). I do two days per week on ROADS and a day per week on DESIRE. ROADS and DESIRE are both concerned with the provision of access to network resources via Subject Based Information Gateways (SBIGs); these are services such as SOSIG, OMNI and ADAM. As a result both Martin and I do the odd bit of work on metadata and resource discovery standards.
I also work in the University Library's systems unit for one day/week (looking after CD-ROMs, doing database work, fiddling with out Talis OPAC and other miscellaneous library systems stuff), in our PESS&RM department for half a day per week (doing biomechanics modelling) and up to half a day per week for the eLib ACORN project as a technical consultant.
Jon: I did both my undergraduate and postgraduate degrees at Loughborough and for the most part just slipped into doing the local jobs because I'd volunteered help or done smaller contracts in the past. For example I was known to the Library chaps because I'd written the software for the BLR&DD funded ELVYN project (an electronic version of a journal delivered in HTML) and helped out on a prototype of our WebOPAC. I got into ROADS and DESIRE because Lorcan Dempsey got in touch with Martin and suggested that we might be interested in joining in on a proposal for an eLib project based on what he'd seen of us on various mailing lists. ROADS and DESIRE were lucky breaks for me because I was planning to just do odd jobs for a few days each week and then spend the rest of my time doing the "Internet stuff" I enjoyed unpaid; these project pay me to do it! :-)
Jon: As a few people know, I'm not too enamoured with Java; its definately currently in the over-hyped-solution-looking-for-a-problem category for me. It always strikes me as C++ with the useful bits removed, which is not unnatural considering its heritage. Having said that, the hypewave has reached such a point that too many companies and individuals have staked too much on Java for it to suddenly collapse and go away and it does appeal to me more than ActiveX (which looks like a security nightmare waiting to happen). Maybe when Java gets a better standard GUI toolkit that AWT, faster JIT compilers and I can think of a good reason for using it rather than a C program or a Perl script for myself I'll give it another go.
Jon: Caching is cool (see, I do like some things... :-) ). In fact its more than cool; its vital to the future scaling of the Web IMHO. By pumping all access for remote documents through a campus cache, a site can "save" an appreciable portion of its bandwidth _and_ improve the percieved performance of the network for its end users. Getting every site on JANET to have a campus cache in the same way that every site has a mailer would be a very good thing to do. More so if these campus caches then took part in a cache mesh whereby they asked regional siblings and national parent caches for objects that they didn't have. I think that the institutional caches are a vital part of a national caching strategy because enouraging invidual users to point their web browsers at a single set of national caches is bound to lead to hotspots on the network far faster than if local or regional caches could satisfy the request first.
One thing that really irks me is the current trend that some vendors have of wanting to generate "dynamic content" that is purposely designed to "cache bust" (ie not be cachable). Their argument that they want to tailor the information that they return to individual users just doesn't hold water for me; what's the point of a set of dynamic data tailored to my exact needs (or what the vendor tells me my exact needs are at least) if the network becomes so slow and congested that I can't actually get any of it? Many people will already have experienced how slow the Net can get on a weekday afternoon; I consider vendors that are pushing cache busting products out to be adding to this congestion and thus be the worst kind of bad network citizen.
Jon: Heh, John has put this question in because he knows it'll guarantee to have me frothing at the mouth. :-) I look after both the networked and standalone CD-ROMs in the Pilkington Library here in Loughborough University and they are a never ending tale of woe. Don't get me wrong; a few CD-ROMs are model products that you just slide into the CD-ROM tower, build a fileset for and away you go with no problems (the Guardian CD-ROM fits into this category - a top product that I'll highly recommend to any library). If only all of them could be like that. Most require some jiggery-pokery to be done to get the software working over the network but once done, updates are easy.
However there are some CD-ROMs that not only are a pain to install but also are a nightmare to maintain. These are the ones which have frequent updates, with each update requiring a new installation of the software (usually to add some feature that wasn't needed or asked for) and, better yet, encryption keys that mean that the CD-ROM suddenly stops working when you least expect it to (usually with either an obtuse error code guaranteed to scare the living daylights out of most non-IT literature users or no error at all so that as a systems support person I get to play "guess the problem"). You sometimes get crappy technical support from vendors to add to this, although I must admit that most of the vendors I deal with have now got reasonably good - H.W.Wilson still stand out as one of the best with multiple trans-Atlantic phone calls on their dime to help fix a minor problem). Throw in one A5 page for the installation instructions (that naturally assume nothing goes wrong) and idiosyncratic user interfaces and I think we can safely say that CD-ROMs are a pain.
As to the future of CD-ROMs, I think we can also safely say that they'll be with us for some time to come. They're still a very efficient way of shipping large quantities of data around relatively cheaply. Because they're mounted locally, you also only have local congestion to deal with. However we're already seeing a move away from CD-ROMs to online services on the Internet; at Loughborough we've replaced a couple of CD-ROMs with access to OCLC FirstSearch for example. This reduces management time to practically nil; all the mounting of the databases is done at the remote site. How well this will perform over congested network links and how well it is accepted by the end users remains to be seen. One thing is for sure though; at the moment as soon as one CD-ROM product is removed because an online version exists, another quickly steps forward to take its place. There are a lot of CD-ROMs out there containing data that a lot of end users want and that isn't going to change overnight. Future high capacity optical discs are also going to ensure that we'll be mounting little silver discs well into the next century.
Jon: Yuck. The few times I have to use it really make me wonder why people actually buy the damn thing. Its user interface is somewhat less than intuitive and yet still make some simple operations really tricky. I seem to have the nack of walking up to Windows95 machines and completely screwing them up in no time at all - this just shouldn't be possible for a normal user in this day and age. Give me a copy of Linux and the fvwm window manager any day of the week. And before someone says, "Ooh, but how do you read Word documents?" the answer is that I read them using Word, loaded off a Novell server into Wine running under Linux on an AMD586-133 based PC with 8MB and then displayed on my Sun under X11R6. Works a treat. Though Word isn't my choice of application for text generation (I'm a die hard LaTeX user; far more powerful and lets me concentrate on document structure rather than distracting me with glitz and wizards).
Jon: Until very recently I used X Moasic 2.7b5 almost exclusively, with a bit of lynx thrown in occasionally. However I've now faced up to the fact that the NCSA X Mosaic series is dead (the 2.8 alpha series just stopped) and so I've actually got a copy of Netscape 3.01 for X11 open on the desktop at the moment. Whilst Netscape gives me working tables (some thing that the 2.7 series of X Mosaic never did manage) I still find I miss some of the nice X Mosaic features such as the right button menu that let you do an Alta Vista search based on anything highlighted in any X11 application or go straight to a URL highlighted in any app, and the ability to closely configure which sites you don't want to ever see images from (which means that you get rid of the banner ads completely and get a lot more information on the screen).
Jon: I think it is; unlike some people I don't expect that every eLib project will be a complete and total success or failure; every project is likely to find out useful things, even if those things are "the way we tried to do it didn't work". Finding out that something doesn't work and showing some reasons why it doesn't work is still a good result in my book. Despite what has sometimes been said, eLib is definately a research based programme.
Also if eLib is supposed to be encouraging the use of electronic resources in academia then its already working; the ANR SBIGs have increasing numbers of hits each month, more users are introduced to IT, new resources are brought online, etc. As I said above, our library is moving from CD-ROMs to online resources as a service provision (ie a non-eLib funded activity) and we've got a WebOPAC that allows us to do neat things with the catalogue data (like link directly to the online version of Ariadne for users not sitting at the OPAC-only terminals in the library).
Jon: I don't know about ROADS being exciting; its been fun doing it and I think we've produced some useful output but I don't want to overhype what we're doing too much. I think we will see a national set of SBIGs all talking to one another at some point, and indeed some of the ANR services have already played with getting a single ROADS front end to allow searches to multiple backend databases. When we ship ROADS v2 later this year with centroid support I expect to see alot more of this happening, as it allows the SBIGs to provide much better support for multidisciplinary searching for quality resources. However I must admit to still not being clear on the politics going on behind the scenes on such a federated information gateway and what funding model the SBIGs that form it will have (I'm just a techie after all; its not really my concern). It will be interesting over the next year to see what happens in this arena.
Jon: He stops me from playing the same tape in the office for 12 hours solidly (we have the stereo on continuous play and I sometimes don't notice that the Ozric Tentacles album is on its 6th pass of the day) :-) Seriously, no, Martin doesn't really have any annoying habits, which is probably why we still get on after nearly 10 years (we were Freshers together). I probably talk too much for him though (especially when I still talking to myself whilst I'm thinking).
Jon: That's a tricky one; I think some parts of Universities need firewalls (say the admin chaps for example) purely because there are so many nasty people out there. However I think that there should be at least some untrusted machines outside the firewall that allow people to experiment with new protocols and ideas. And not just researchers as well; lots of undergrads have good ideas that deserve to be tried out.
Jon: No, don't charge for JANET in general. The US and Australians have gone down this route and all I've seen is academics regretting it bitterly as it results in penny pinching and the stifling of discourse and interaction. I personnally would have been badly effect if blanket charging already existed; for example it would have been unlikely that we'd have got into multicast IP and the MBONE as early as we did and I would have missed out on many valuable conferences and seminars that I have been able to virtually attend over the network, plus it would have affected my PhD work (which included some long haul multicast IP based experiments). Indeed unless I could have "played" with multicast IP freely to see how it worked, I might never have taken the route I did in my PhD and some of our community knowledge would have been lost. As this is what Universities are all about, this would obviously have been bad news.
However, it might be worth penalising sites that don't do their bit to be good network citizens for service traffice; FTP, Gopher and HTTP traffic not originating from a campus cache incurring a charge might encourage some of the more lax sites to get into gear and state such a cache.
Jon: Yes, both staff and students should be able to put information on the Web. Its often valuable to be able to find out something about people before you meet them and there are loads of very useful resources on the Web that are really just bits of people's home pages.
Jon: Well, if you precached all the Playboy and Hustler sites' material in your campus cache then you'd quickly reduce the traffic on your JANET link! :-) Seriously, I think that the problem is blown way, way out of proportion; whilst it is true that a lot of people do look at "porn" sites, I'd say that much of this is because there is so much in the news about porn on the Internet and people are natually interested. There's been porn available over JANET for years; certainly when I was a fresher it was available if you knew where to look. The difference today is that its easier to use the network to find any information, including porn, and because of media hype in the last few years more people are now looking.
If you were feeling a bit fascist and wanted to censor stuff that _you_ consider pornographic you could always enforce the use of a campus cache and then get that to quietly discard retrievals from known porno sites. This does mean that someone would need to be constantly looking out for such sites (and this might well have to be a community project with several sites sharing the load). Using filtering systems at the web browser end is not such a good idea; many people in academia can easily circumvent those tools and they can also remove single terms that change the meaning for documents used for research (a classic set of examples has recently appeared on the web4lib mailing list where all a librarian in the US was testing one such product and noticed that it indiscriminately removed the word "queer" from a set of USMARC records she had downloaded). A last point to bear in mind about the filter products is that they usually censor material based on the vendor's view of what is objectionable and not the view of the library/computer centre/senior management/moral enforcement committee of the University.
Jon: Its got to be to get everyone into using caches (including the vendors of dynamic content, through customer pressure if nothing else). They're a cheap and yet effective way of helping make the bandwidth we've got go further. Be good network citizens folks!
Jon: I encountered the Web sometime in 1991/2 (I forget when exactly). Martin had been playing with it and I'd taken a look as I ran an FTP and gopher archive. At that time we were using the CERN line mode browser, X Mosaic was just appearing in its early, slightly flakey alpha releases and images were few and far between. My first impression was that it was a nice idea but the SGML style of HTML seemed to be a real bind; I'd have personally gone for an extension of LaTeX if I'd have been Tim Berners-Lee as I was very anti-SGML at the time (I'm less so now, though SGML does still seem to be struggling to do what LaTeX has been doing for years). I created some pages for the archive (that mostly hyperlinked to the gopher version of the data) and that was that; I got hooked, CGI came along and the rest is history.
Jon: Throw the Sainsburys loyalty card as far away as possible for starters; its useless on a desert island and I refuse to have one anyway (I don't feel very predisposed to giving a large corporation market profile information in return for the odd free box of doughnuts every few months). Then build a shelter and work out what food I can gather (I assume that this desert island has food sources available; otherwise the next thing I do is use the laptop to send an SOS email to lis-link :-) ). Then I settle down with a bowl of fruit and berries to read my email and get to work on the web. Hmmm, this sounds very attractive; no meetings to go to, nice environment, lots of peace and quiet. When do I get to go? :-)
Thanks, Jon and Martin.