Web Services are boring because..
Saturday, March 23, 2002 by Dave Winer.
A few years ago I would be more casual about writing DaveNets, I'd just open my outliner and write for a couple of hours and edit with a light touch and send out a piece.
Lately, I've been doing that on the Web, on Scripting News. Today I strung together a bunch of semi-related essays on the website, and thought what the heck, why not run it as a DaveNet.
So here goes, a loop back to casualness and irreverence. This is the come as you are channel. We're all just folks here, set down and rest a spell, chew the fat and let's have fun.
Everyone's cutting Dvorak a new anal orifice for his piece about web services, but there's a perverted kind of courage to what he says that's worth defending.
Suppose Dvorak doesn't know that there are lots of independent developers doing cool stuff with XML-RPC and SOAP? It's quite possible especially if he reads the reports written by his colleagues in the other big pubs.
Dvorak says the BigCo's don't know what they're doing. Of course! Any idiot can see that. People who say otherwise are lying. It's obvious on its face. If they knew what they were doing they would be selling products, not starting new consortiums to study interop.
So give Dvorak a few points for having the guts to say what's obvious that so few are willing to say.
First, no users are asking for web services. That's as it should be. Web services are about how you implement things, not what you implement. Users care about choice, not being locked in, reliability, security -- so perhaps they can care about web services, after compelling apps ship.
There isn't really much fun in web services unless you're a programmer who likes to play in different environments, or likes to work with other programmers who work in different environments, or thinks the Internet is cool, even if he or she can't totally explain why. Web services are plumbing, and therefore to most non-programmers, off-topic.
The second reason Web Services are so boring is that the Big's send out suits to do the talking. So where there might be a spark of insight somewhere in the bowels of the Bigs, those people don't get a chance to speak, so the confusion is all you hear in the press reports because the people they talk to are confused.
What the press is really saying is that Bill Gates told me this is important, which is pretty close to Bill Gates told me to write this, which is where the big pubs lose me and most other readers with minds.
If I can have sympathy for John Dvorak, I should also be able to call up some sympathy for Dan Gillmor. A friend who talked with Dan recently told me how it's going at the SJ Merc. People walk the other way when they see Dan coming. I can only imagine how Dan really feels about the Merc tossing his archives in the bit bucket and crowding his copy with ads, and burying him three-deep in irrelevant structure. But he's said almost nothing about it on his weblog.
According to Knight-Ridder, Dan writes for Real Cities which is part of Silicon Valley Dot Com, neither of which are brands in any sense of the word. This is backwards. Knight-Ridder and its inventions are the overhead we tolerate to find out what Dan thinks. The structure exists to support the talent. Dan is the talent.
So his conspicuous silence on the technological ineptness of his support team is perhaps somewhat understandable. After a time you get tired of fighting. Dan makes his living writing. So if he quit Real Cities or whatever it is he works for, where would he go? Who would pay him to write for Dan Gillmor Dot Com?
Dan Gillmor's column is currently here, but that pointer, based on past experience, should rot before this day next year, no doubt about it.
I've been reading Steve Pilgrim's weblog, consistently, he's the designated HTML newbie of the Radio community. For Steve, I wish that HTML today were as simple and approachable as it was when I got on board, in 1994. Even then, for a guy like me, a commercial developer with a computer science degree, the leap wasn't easy. Today, it's so much harder, esp with all the confusing advice from techies over the basic stuff.
Now, the thing that made it all come into focus for me was a simple mail-to-web app run by Ohio State University. I remember it so well. It was January 1, 1995. I knew how to use Eudora to send Internet mail (although that was a leap from AppleLink, which I had been using up till 1994). The rules were simple. Send a message to their mail server. It sends back a URL of a page containing the text in your email. So I sent it a simple message. Hello Dave. As I expected the page that it sent back a pointer to contained just that text. OK, so I went to HotWired's home page, did a view source, copied the text, pasted it into an email message and sent it to Ohio. Back came a URL. I went to the page. Wow. There was the HotWired home page with a few broken images. Bing!
For some reason that was the only impossibly difficult concept for me to grasp. A website was just a bunch of text somehow stored so that some piece of software (a Web server) could find the text and then shoot it out to a Web browser. The server wasn't doing anything very complex. The complex stuff was in the text that it was sending (and even that wasn't very complex).
So maybe the answer for the incurably inquisitive, like Steve, is for us oldbies to put up an app like the Ohio State app that shows them how the Web works at its most basic level. Interestingly, Steve and everyone else that uses Radio has exactly such an app, sitting on their desktop. It's called Radio.
Much of the confusion for newbies and oldbies alike comes from the idea that the standards of the Web are moving. But it's a dream. Something as installed as HTML won't change because that isn't how software works. If you want to move you must do it fluidly, especially when the installed base is as large as the Web.
To a newbie like Steve, telling him that the font tag could disappear is like telling someone learning to drive that the volume control on the radio might disappear. It's weird enough to be confusing, but there are so many Web pages that use the font tag and other basic HTML stuff that's been targeted, that it's unthinkable that they could go away.
Just the other day one of the new standards advocates pointed to the major tools by Adobe, Macromedia, etc and said that none of them follow the new guidelines. A few weeks ago someone bothered to look at the home pages of the W3C members and found that almost none of them validate.
So, is the font tag being phased out? It can't be, if it were it would break most websites.
PS: W3C stands for World Wide Web Consortium.