Push Button Paradise

Nice little bit of JavaScript from Daniel Glazman. I wonder how many people are running with JavaScript disabled these days... (I use server-side PHP to accomplish the same thing on some of my sites) Link: http://disruptive-innovations.com/zoo/20040830/HTMLoverlays.html -m

Orbeon, one of my Ten Favorite XForms Engines, has gone open source under the LGPL: http://www.orbeon.com/company/pr-oss-announcement As of right now, it's all up on Sourceforge.

Significantly, they write: "there is a greater opportunity in building a services and value-added business around an enterprise-grade open source business integration suite than selling proprietary software and services".

So XForms on the server is in great shape. XForms on the client is underway in Mozilla. I think there's one piece still missing for a truly seamless transition story, but I won't have any news on that until later... -m

My server had an aging installation of Red Hat 8 on it. It was starting to show almost IE-like signs of flakiness, due mostly to nearly every piece of hardware getting replaced recently, while mostly original configuration files remained. Time for a refresh, and I picked SuSE 9.1. I like it.

Your download options are a 27 Meg boot ISO which installs the professional version from FTP, or a 700 Meg "personal edition" ISO. I went for the 700, but immediately added back in the hooks to easily get the professional stuff, like a Samba server. (In YaST, add this file site: ftp.suse.com/pub/suse/i386/9.1 ).

Despite using Gnome for years, I'm also trying out KDE, mainly just to see how far it's come. Pretty far, I'd say. It's gorgeous, and at least as usable as XP. On my new $300, 17 inch LCD, it looks great. -m

Maybe it's just me, but Safari as of version 1.2.3 on OS X 10.3.5 isn't including buttons in the DOM Level 0 access (either document.formname.buttonname or document.formname.elements). If anyone can verify this or knows of a workaround, let me know. -m

Here's a convenient datatype library I'm kicking around. It's based on the unfortunately-widespread XSD part 2 system.

A datatype is defined like this: dt:string-20, or dt:string-10-20. The first is a string with a maxLength restriction of 20; the second is a string with a minLength restriction of 10 and a maxLength of 20. The base part has to be a datatype defined in XSD part 2 or XForms 1.0.

This is particularly useful for forms, where it's common to put such restrictions on existing datatpes, and it's inconvenient to have lots of little one-step restrictions piled around somewhere.

It's especially useful in conjunction with the XForms extensions to Examplotron: <fname eg:content="dt:string-30">First Name</fname> maps to a limited-length input control. -m

The esteemed T. V. Raman also wrote a book on XForms, for Addison-Wesley. Now it's available online too. No obvious word on what license it's under (nor could I even find a copyright notice, but maybe I'm dense). Link: http://safariexamples.informit.com/0321154991/book.html -m

I've spent some time recently looking at the remarkable symmetry between Examplotron http://www.examplotron.org and XForms.

The basic idea of Examplotron is that an instance document is a schema, with little or no annotations needed. XForms takes a similar approach: the instance data is in example of what gets submitted. Any variations are due to user interaction. In version 0.7, Examplotron even has an eg:assert attribute, allowing Schematron-style arbitrary assertions. From here, it's a short step to add in indications for XForms properties, like required, relevant, readonly, calculate, and constraint. More on this later... -m

John Gilmore keeps trying to fly within the US without having to show ID to get on the plane. His site is at http://freetotravel.org/ I did it yesterday, more than once depending on how you count...

From home I did a web check-in for my America West flight, and printed out a page that vaguely resembled a boarding pass. At the airport I went straight to the security queue. There I did have to show an ID to get in the queue, that that is a mindless (and ultimately pointless) exercise of matching a name to a semi-official looking piece of paper. No computers or are involved;.no intrusive database profiling or blacklists.

To actually get on the plane, all I had to do was tear the sheet of paper along the dotted line and give one half to the lady at the podium. No ID. After sitting in the unmoved plane for 1.5 hours, the flight was canceled due to hydraulic problems.

Another lengthy wait in a line, and I'm re-routed through Las Vegas. "I'm speaking in San Fran at 1:30. Can you get me there?" The flight leaves in like 2 minutes, so I'm raced onto the plane, without showing ID and without a boarding pass.

The connection in Vegas is basically instantaneous, as they are holding the plane a few extra minutes just for re-routed passengers. All I have is a different computer-printed page handed to me in Phoenix. Here, I am asked to show ID to get on the plane, and I oblige. She takes my paper and I board the plane with no documentation at all, just a verbal "you're in seat 19-D".

Returning the same day, I check-in at a Kiosk, go through security (showing ID to get in line), and board the plane without taking out my ID again. Basically, getting into the terminal without ID is harder than getting on a plane without ID. That seems...backwards.

So basically, I agree with Gilmore, Schneier, et. al. The system is a huge hassle, needlessly restrictive, and largely ineffective. That's not to say we'd be better off with no checks, but it sure would be nice to see some analysis and justification for what's going on there. -m

P.S. I made it to my talk, just a few minutes late, during which time the remarkable Bill Trippe covered for me.

I'm away for the day, speaking at Seybold in San Fran. -m

SCO Group alleges "non-literal" copying of their valuable Intellectual Property into Linux; presents non-literal evidence to the court

In a stunning non-literal announcement, the Lindon-based SCO group today explained the legal strategy behind their attacks on the Open Source operating system known as Linux. "Intellectual property cases are complicated," said a non-literal spokesperson for SCO, "you can't just point at anything in any of our court filings and say 'yeah, that's evidence'." In a comprehensive series of Motions for Partial Summary Judgement, IBM lawyers severly criticized the apparent lack of any basis whatsoever in SCO's case. A non-literal analyst and frequent supporter of the SCO Group said, "I am certain that, in the end, SCO is going to win this one--though perhaps not literally." -m

This URL: http://www.formsplayer.com/community/samples/appearance.html seems to trip up the validator. Looks like a legit namespace goof at the top, but later on there are spurious Relax NG validator errors. -m

In the fairly short term, I'm going to attempt to install a subversion server on my web host, the wonderful Dreamhost. Have any readers attempted this? If so, please share your experiences and I'll put them up here. -m

I installed Nvu 0.41 today on the wife's computer. (Amusingly, downloaded from the Linspire site) On the initial screen, a sidebar prompts you to enter web site details, like the http and ftp addresses to access it. From there, I handed it over, and she had no problems managing from there. Great job guys! Any chance of this showing up on fink? -m

With all the recent attention, the number of votes for XForms in the Mozilla tracking system ( http://bugzilla.mozilla.org/show_bug.cgi?id=97806 ) has jumped up to the 600 mark. Only restore supoprt for MNG has more votes, with 718, and the closest competitor after that is PGP plugin with 421 votes. -m

With help from IBM and Novell, the Mozilla foundation has launched an effort to implement XForms in the browser. Link: http://www.mozilla.org/projects/xforms/ This is fantastic news, and it fits in well with some other things I am currently planning but not yet announcing. -m

From BoingBoing: an online version of Lewis Carroll's scrapbook. I have a special connection with Carroll (via Hofstadter) here: http://www.xml.com/pub/a/2002/09/25/linkoffering.html

I also keep a scrapbook of sorts, called It's the data, stupid, which powers this blog among other things. Unlike paper versions, though, the electronic version is inherently sharable, backup-able, searchable, and repurposable. Link: http://international.loc.gov/intldl/carrollhtml/ -m

If you'll be in the area, you might be able to see me at the Seybold San Francisco 2004 conference at the Moscone Center (link: http://www.seybold365.com/sf2004/conference/?sessionid=4640&s=session_detail_fm ) speaking on E-Forms: Key Requirements and Design Decisions, moderated by Bill Trippe. Thursday, August 19, 1:30 pm. -m

Colloquy, as I've written before, is a great IRC program for OS X. One minor nit: it didn't work with the Daily Chump bot (http://usefulinc.com/chump), specifically, it didn't show NOTIFY messages sent to a channel. Didn't as in past tense. I went on the IRC channel, chatted with the developers, and within 10 minutes, the fix was checked in. Man, I love open source. Link: http://www.colloquy.info -m

Does anyone else think the new paper used by Dr. Dobb's Journal is icky? -m

Last night one of my non-technical friends contacted me with an unusual problem: any attempt to access a search engine page would immediately crash his browser. A complete AdAware scan didn't help. Installing Firefox did. The tide is shifting. -m

Slides from my tutorial: Real-world XForms available at http://dubinko.info/events/XTech2004/ The same slides were used for the shorter session, just skipping about two-thirds of them. :) Like this blog, everything there is personal opinion, not representing my employer. -m

Tuesday. The phone rings at 7:30 am. Hmm, I don't remember asking for a wake-up call. It's the conference chair: 'Can you give a half-day tutorial on XML Schema Languages, starting in about an hour?'

After a few moments of groggy contemplation, I accepted, but asked for a co-presenter.

Fortunately, the original speaker, the remarkable Eric van der Vlist, left PDF notes of 90 pages. I could cruise through the material at 2 minutes-per-page, keeping a good pace. I covered XSLT-as-schema, RelaxNG, and XML Schema. Kip Hampton made a heroic effort, and presented on Schematron, which he uses on a regular basis.

Mike Fitzgerald, author of the now-available XML Hacks also helped out for a few minutes after the first break, sharing his experiences and generally encouraging the audience to give RNG a try.

Overall, a good experience. I tried to include little extra bits of experience, like using examplotron, or editing (with XML Spy) the Schema for XForms 1.0, and later converting it to RelaxNG. For the most part, the students didn't feel cheated by the cruel twist that left the regularly-scheduled speaker stuck at O'Hare. -m