Archive of UserLand's first discussion group, started October 5, 1998.
scriptingNews outline for 3/8/2001
2 DaveNets: A tip for Yahoo, Wishin and Hopin.
Microsoft: SOAP Toolkit Beta 2.0.
Evidence of webness in SoapLand. Today MS started pointing to us from their SOAP and Web Services pages. Thanks!
Now, what is Hailstorm? "One developer who plans to attend the meeting.. said Microsoft is looking to extend the capabilities of IM as a standalone application and turn it into a software infrastructure that can be used to build many types of applications." Gotta get some of that.
Dann Sheridan: "I believe that Hailstorm is the .Net serivces cloud that captures all six million Microsoft developers and doesn't let them go."
Update: I will be at the Hailstorm briefing next Thurs.
Meanwhile, I'm meeting with Jeremie Miller of Jabber tomorrow afternoon. Lots of interesting rumors flying around IM in the next few weeks. I hope to learn more from Jeremie tomorrow.
Intuit is operating an XML-over-HTTP programmable database called QuickBase. It's free. Interesting.
Evan Williams has a screen shot of the new Word. It looks just like what we want. If only it could send SOAP messages, we'd be done. (Email doesn't turn out to be enough. You have to be able to easily edit a document in Word, to have it fit into a content management system.)
Globe and Mail: "A 21-year-old Canadian Web entrepreneur is planning to circumvent the imminent demise of Napster Inc.'s controversial Internet song-trading system by setting up a clone of the service on a so-called 'data haven' platform off the coast of Britain."
DJ Adams: "In this article, I'd like to show you how easy it is to put together a simple system for pager-style notification of incoming mail, using some standard tools, Perl and, of course, Jabber."
The Snowcrash plug-in "runs a mini HTTP server inside WinAmp."
The RadioShack 2.4GHz Wireless Room-to-Room Audio/Video Sender "allows you to use your home stereo and theater components in any room of your home." $99.
Tech Interview is realllllly groveling for the links. Gotta admire their style.
Here's a disgusting and pathetic idea.
This is disgusting and pathetic and funny!
Danny Goodman: "In March 1981, I quit my last job to become a freelance writer. Twenty years later, I'm still at it, so it's time to celebrate."
Caterina explores happiness in re the subconscious.
News.Com: "..Koogle lamented that Internet companies had erred by giving away too much content to consumers in the dash to gain market share and notoriety. Consumers have been loathe to pay for any Internet content, from news bulletins to stock tips to downloadable music and software."
The cool thing about my life is that I get to meet such smart people, and lots of the thanks for that goes to the Web. When I was interviewed for the Newsweek article about weblogs a couple of weeks ago Deborah Branscum asked what practical use weblogs serve. Good question. I told her that there was no way XML-RPC and probably SOAP could have happened without weblogs. Community projects require good communication, and the Internet, if it does anything, is an unprecedented communication enabler. Mail lists are fine, but they can easily get locked up and frozen, unable to move. Weblogs are different, it's possible on a weblog to speak and not give equal weight to every point of view. This begets more weblogs, giving people choices, they gravitate to the places that make sense for them. With this, movement can happen. Ask a question, get an answer, move on, ask another. After a few years, the results are visible.
Perhaps I got to the bottom of the controversy over WSDL. SOAP w/o an IDL is 404 Not Found all over again. You call a procedure and it's not there or its interface changed and your script fails. What does an IDL do to prevent such an outage, or to help you debug an outage that a thoughtfully worded error message wouldn't? Of course in this debate I'm taking the side of linkrot. It's our friend, it makes this stuff implementable, and keeps the gorillas from swamping and crushing The Little Folk.
Speaking of thoughtfully worded error messages, at UserLand we have a template for messages, which we always (Murphy-willing) follow. It goes like this. Can't xxx because yyy. The first part is important because the user must be told what the software was trying to do when it failed. When there's an error, part of the "user model," like it or not, is the idea that each user operation is decomposed into thousands or millions of procedure calls, any one of which could fail. So the xxx part of the message gives you a clue which one of those internal procedures is the one that couldn't do its thing. The yyy part is supposed to give you a clue how to cure the ill and get on with your life. So a good error message includes both bits of information. By having a default structure for the message, it saves the programmer from having to be creative, in English, while he or she is being creative, in code. And it causes the programmer to worry about the poor users, how will they know what went wrong?
BTW, it's totally true, theoretically, that every procedure could fail, but does this mean that programmers must check every operation to see if it worked? No. But I've seen people try. The code goes like this.
x = y + 5Well, that's ridiculous. If the CPU isn't working your test isn't going to work either. And the user probably has a few other clues to this, like the OS doesn't boot up.
if x != (y + 5)
scriptError ("Can't xxx because yyy.")
This page was archived on 6/13/2001; 4:57:00 PM.
© Copyright 1998-2001 UserLand Software, Inc.