WEB WINS

Dick Pountain/17 March 2000/Idealog 68

So it's official then: the future of mankind lies on the World Wide Web. The 'old' economy that made things (emitting smoke and nasty smells in the process) is now over and a squeaky-clean, organic 'new' economy has effortlessly taken over in a bloodless revolution. Rover cars goes into the dumpster, Whitbread is booted out of the Footsie, even BT is seen as a takeover target, while a bunch of virtual booksellers and travel agents have become worth more than all of them put together. Within a few years I'm sure that the teething troubles - like the fact that none of the Web companies are making money, or have a plausible model for how they might make money, or employ anybody - will be sorted out one way or the other, and it will be so exciting waiting to see whether anything then remains of the world economy.

Microsoft has announced that the object/component wars are over and the winner is SOAP, and after an initial knee-jerk scepticism I think that I believe them. HTTP is one of only a handful of genuinely worldwide standards and like other real-world standards (VHS video, the IBM PC) it was the people's choice rather than imposed by committee. Making DCOM and  CORBA objects talk to each other in plain XML text across an HTTP link might just prove to be the Dynorod that's needed to clear the blocked pipes of distributed object systems.

All this being the case I figure it was about time I did something about my own Web site. This was a Noddy sort of effort that I knocked up around three years ago and haven't really maintained or improved during that time. I created it originally just using Microsoft Word's web output and adding some raw HTML with a simple editor called WebEdit. All the Idealog columns went up as separate pages, each converted to HTML using a script (and that became such a chore to update that inevitably I stopped doing it at around issue 40). This time, thought I'd try out a more modern tool and take it a bit more seriously. It so happens that PC Pro gave away a full working copy of NetObjects Fusion v2 on a recent cover disk, so I decided to give that a whirl. Elsewhere in this issue in his Real World column, Tom Arah talks about Fusion v5 and gives it high marks, and I'd have to say even from experience of the older version that I agree with him - Fusion's user interface is pretty good compared to other tools I've tried, and it makes laying out Web sites almost as easy as a DTP program does for local documents. Adding sound files and other bits and bobs is pretty sweet too. Using Fusion's data publishing feature I was able to put up all the Idealog columns as 'stacked pages' that all share the same format and have an automatically maintained contents list on the first page: now I have no excuse for not updating them. You can check out the result at dickpountain.com.

However Tom also remarks in his column that if you use all of Fusion's more DTP-like features then it generates some horrendously inefficient HTML, and I'd have to say I agree with him there too. When testing my site I noticed that a hypertext glossary of networking terms that I'd imported from my old site suddenly became appallingly slow. The glossary was a single page a modest 12 Kbytes in size (I originally compiled it for a Pro feature on ATM years ago), and I'd hand-coded it using anchors for all the cross-references. When I went to see what Fusion had done to it I discovered it had broken it up into about eight separate pages, so that just about any link you clicked on loaded a new page instead of just jumping within the page. Maybe later versions of Fusion don't do that - but I just shrugged and patched it to point to my original file. Next I noticed that a stunning number of files had been generated for the Idealog columns: each column nn became idealog_nn.html, which simply called  header_idealog_n.html and body_idealog_nn.html. However the header was identical for all the pages, so 150 Kbytes of space was being eaten up by 65 differently named copies of the same file. Again my instinct to hack overcame my trust of the tool, and I just deleted all but one of them and hacked the links using a script. Of course the problem with doing this is that you can't then just republish the site at a single button press, - which is the whole point of this tool - because that will undo your hand patches.

Such concern over code efficiency seems very 'old' economy indeed: after all everyone is fighting to offer free net access and pretty soon they'll be paying us to go online, so maybe I have to 'let go' and just pretend that bandwidth and disk space are free. But I can't do that because my site is hosted in 2 Mbytes of free web space, and I have no intention of paying for space. Basically I believe with Dr Johnson that 'No man but a blockhead ever wrote, except for money' and though I can steel myself to give away old writings, the idea of paying to give them away is a bridge too far.