[This is long, and not strictly on-topic as it does not deal entirely with mis- or ab-use of available tools. It is a different perspective, however, that I think usefully expands on the issue Meg discussed.]
As a user, I'll grant that both the static and pioneering viewpoints are valid to a certain extent, but that Meg has hit on the key issue without actually naming it or discussing it in its particulars: intent. The examples she provided were of Dilbert-esque people in a position such that they could encourage/press/force their whims on the user through the designer. I do not want someone with zero (or a misguided) sense of responsibility toward 'the user experience' designing or contracting for UIs, and tonyweeg is precisely the kind of developer who drives me nuts, insofar as his post reveals himself. Aside from expressing (at best) nominal interest in the user's experience, and none for the user's opinion of pioneering, he expects that the user will be thankful for his htm-asturbation. This is an attitude too-often coddled by the Dilbert-esque types who, contrary to Meg's specific issue, don't know enough to know when to put the brakes on a developer.
Granted, my own term for pioneering was unspeakably rude, but it suits the post particularly well as tonyweeg's view of development seems wholly centered on self-gratification. This is, of course, despite an attempt to wax lyrical in comparing the use of the 'latest and greatest' web tricks to something historically demonstrably useful, on average. A more correct comparison would be to mountain climbing, in which the 'advancement' is "Looka me!" shouted from a significantly higher elevation than usual.
As to user-gratification... I find it gratifying when the loading of one page element, if stalled, does not stall everything else, leaving me with a blank page for a minute (Akamai, I'm looking in your direction...). I like it when a developer recognizes that html will do a job as well as script, and leaves the script out, and when a site has a functional and clear design, rather than completely unnecessary dhtml drop-downs that take forever to load, or text images that leave me guessing and/or mouse-hovering once I've established that the ones already loaded are not what I need.
I'm delighted when change arrives if it proves to be substantially better than preceding methods, with the speed of change 'allowable' inversely proportional to the improvement it provides. I'll get fitted for a plug in my head if you give me ten minutes to get dressed and have a quick bite to eat, but the speed at which change can happen on the web absolutely does not match the speed at which the average user changes. If you tell me to download the latest version of Flash, I encourage you to hold your breath. It will be as much fun for me to watch you turn blue as it was for you to play with the latest web-tricks.
Some examples that outline why I find being subject to the whim of the pioneering-developer so repugnant they center on pointless and/or 'fun' pioneering that may indirectly or directly degrade performance, or fail to provide any user-measurable benefit over other methods of accomplishing the same thing:
I've never used IE for longer than about an hour (and only as needed) because I started with NearlyDeadscape when it was still viable, upgraded as needed, and I'm holding steady with 4.6. I fully expect and understand that there will be sites out there that 4.6 can't handle, but even 6.x was no boon for one stupid reason. Those familiar with Netscape will know that there is a floating toolbar to provide access to the browser, email, composer, etc. You also know that since 4.x came out, when closed, this bar docks in the bottom-right of the browser window. In 6.x it docks on the bottom-left. After about six or so years, due to the whim of the designers, I need a new habit. I went back to 4.6 so fast I can't even guess at what other new habits I'm going to need.
I still cling to Win98SE as well, as Win2K was no prize. It refused to remember the positions I set for my usual windows because I set them to appear in the exact same spot (same size, bottom-right). Opening several of these windows from the quick launch area (a perfectly valid and welcome addition) causes them to alternately open in the correct location and... elsewhere... On top of this, the shutdown window has been changed from the more-convenient radio buttons to a drop-down. Again, whim, measurably to the detriment of the user, if only by a few extra motions and a new habit.
The intent of the woman who wanted 'taps' was clearly something no more well-reasoned than "Ooooooh! Pretty!", and it is this attitude in web-designers or their bosses, more than any other factor, that drives change on the web faster than the user changes. We already know that the providers of design tools, browsers, etc. have the bottom-line at heart, and advance a hyperannuated style of Detroit's planned obsolecence wherever possible, but the users and directors of users of these tools are in the unique position of being able to provide a buffer. When developers' intent is focused on using the absolute latest (and, OBVIOUSLY, the best) new features, they implicitly try to drag the user along as fast as the demand for profit snowballs above them. But developers can kill a new tool in a few quarters if they en masse decide it provides too little benefit to justify the degree of change required.
Please do either. I'll be happy to fill the pockets of those on the rung above you, but only if you make it worth my while to change that quickly. I'll be just as glad to stick with what I know and love as well; the need for web-developers certainly won't stagnate in the interim... the ability to coherently impart information is a much-needed skill not lessened in import by a choice to do so with tools less recent than "Released yesterday!!!!!"