Recent builds of the Internet Explorer 6 beta have included the IE implementation of Microsoft's smart tag technology, prompting choruses of disapproval from the media. Which is of course understandable - smart tag technology parses its way through a web page, underlines the words it's been pre-programmed to react to, and inserts its own hyperlinks.
These take you to wherever the smart tag developer wants to take you, entirely without the knowledge or permission of the web site proprietor - in essence, IE6 is taking your HTML and re-editing it locally. The system could be used to convert news stories into vehicles for advertising, without the perpetrator actually having to bother paying the owner of the stories, it could be used to blacken the reputation of, say, Dave Winer of Userland (Connie Guglielmo of ZD Interactive Week has a most entertaining exposition of this very thing here), or it could be used as a cheesy gag to get hits for MSN.
The possibilities are limited only by your imagination, and characteristically Microsoft has therefore gone solely for the cheesy gag in its first IE6 implementation. Smart tags may indeed be switched off as the default when the final product ships, but the default is ON in the build The Register is currently looking at, XP version 2475. Keywords hit on have the tag attached, and if you place the mouse pointer over the word, you can pop up a list of smart tag options/hyperlinks.
So every time Microsoft is mentioned, for example, you get six options; company news, company report, price chart and stock quote (all of these from MSN Moneycentral), official site (microsoft.com, natch) and search the web for Microsoft (MSN search, of course). This seems par for the course for all of the tags currently recognised - six basic options, five of which steal hits for MSN from any site that hasn't yet got around to metatagging smart tags out of the picture. That's what Microsoft says you can do to block them if you don't like them, but we kind of like the Electronic Frontier Foundation's suggestion that you could sue Redmond for breach of copyright, as smart tags create unauthorised derivative works.
Microsoft has ambitions for smart tags beyond feeble hit-swiping stunts, and has already gone some way further with the Office XP implementation. Predictably, this turns out to be only as good as the coding; you can read about Paul Thurrott's experiences in having French holiday destinations and Oakland A's centerfielders thrust at him whenever he typed 'nice' or 'long' here.
Over at XP build 2475 there are some tantalising clues to the IE smart tag ambitions lying around. If you look in Program Files/MSN/smarttag you find a file called msntags.dic, plus what is presumably an associated DLL. It's not entirely clear to us what application is supposed to be used to edit msntags.dic, but you can open it with Wordpad, and the content is just about readable.
It's a very long list of companies, organisations and potential actions. For example, we can't help noticing the words "buy team merchandise" in there. The list seems to go through from 3Com to Zions Bancorporation, with many, many improbable companies in between. It's most likely a prospect list rather than what's happening already - Estee Lauder is in there, for example, but try as we might we couldn't get it to kick up as a smartag on a Web site.
What's Microsoft going to do with it? You can speculate on a sliding scale of paranoia. Say every company in the world gets themselves tagged up to MSN Moneycentral, whether they like it or not; that gives MSN a lot of hits, and maybe puts it in the position of being the primary supplier of information about everybody. And everything. Or maybe Microsoft and/or its partners could sell the links and associated tagging routines. Your company name appears on a web site, so you get to control - for a price - what the reader can do in association with that appearance. If Microsoft sells the service, then it gets all the money, but if a partner sells it, Microsoft still likely gets a percentage for routing it through .NET.
There are a few more smart tag-related files in XP, in the folder Program Files\Common Files\Microsoft Shared\smart tag there's an apparent data file, mstag.dlb, but we haven't figured out what is in it, if anything. There's also a 3x1 pixel bitmap, mstag.bmp - spooky or what? Both of these seem to come with IE6, but there's also a bunch of dlls that come with Office XP, if it's installed, then in a lists subfolder there's stocks.dat (gobbledegook again), and stocks.xml. That one's readable and editable, but would-be dabblers should be aware that deleting references to Moneycentral from stocks.xml is most unlikely to also delete the real moneycentral.msn.com at the same time. Microsoft has however published some instructions and fiddling tools here, if you'd care to try. ®