Saucemaster

I somehow felt that rating this film (or the short film “Boy” that preceded it) was inappropriate. Although I personally found both films very affecting and at times enjoyable, I think that judging the films based on their entertainment value is missing the point entirely. I’m sure that neither Welby Ings, Gregg Araki nor Scott Heim need any validation from me, any more than Gregg needs idiots like Philip Ruddock trying to ban his film in Australia.

Was looking forward to seeing this new work from Wim Wenders, but a P1 bug landed in my lap at an inopportune time (4:30pm on a Friday arvo)…

Question: Hey Nonexpert, my girlfriend drags me to IKEA almost every weekend and it’s driving me crazy. What should I tell her?

Answer: There is no known treatment for IKEA addiction. The best you can do is learn to survive.

IKEA WALKTHROUGH v2.3.1

IKEA is a fully immersive, 3D environmental adventure that allows you to role-play the character of someone who gives a shit about home furnishings. In traversing IKEA, you will experience a meticulously detailed alternate reality filled with garish colors, clear-lacquered birch veneer, and a host of NON-PLAYER CHARACTERS (NPCs) with the glazed looks of the recently anesthetized.

Link to Ikea Walkthrough

Modern U.S. propaganda is as good as the old stuff!

Phoenix has once again produced an album of worth with “Alphabetical”. It’s sometimes strange to listen too, with synthesised honky tonk sounds and all, but it’s never dull.

So I’ve got a problem. We have lots of code, hundreds of thousands of lines of code. Code written in COBOL, JCL, REXX, C, Delphi, Modula-2, Java, VB, C++. Code running on Z/OS, Linux, Windows. Some of it smells, but other code is fresh, and ready to execute.

Problem: How can I find the bits of code I can and should use when I write yet more code?

So we write design docs, program specs, test scripts, etc. in Microsoft Word. How can I search this big ball of proprietry file formats? And even if I could search them, the docs I find would be out-of-date, or just plain wrong.

So we have UML diagrams in Microsoft Visio. This is even more useless than the Word documents. Can I search it? No way.

So we have some source code in Microsoft Sourcesafe, some in CA Endevor, and some on personal hard drives just waiting for the disk to fail. But at least the source code is definitive, the horse’s mouth so to speak.

The current process for searching code: ask the person who sits next to you. This sucks. No wonder every developer keeps writing the same code to do the same things over and over again.

So, I want a search engine for all our source code.

A search engine that indexes source code stored in version control systems, or on hard disks, or on the web, or whatever.

A search engine that parses every language I care about, can filter by language, indexes on class names, method names, variable names, comments.

A search engine that ranks based on how many other pieces of code calls this piece of code. The more a piece of code is used, the more likely it’s of use, the more likely I can use/reuse it in my code.

I want Google for code.

Update: koders.com seems to do just what I want, and there is a version for use inside an organisation.

I was speaking to one of our security guys last week about a project he is spinning up to sort out a manager’s complaint that he had to log on four times to check his email, access his files, check group discussions and whatever else he needs to do. Our organisation is basically all microsoft all of the time, and we’ve still got sign-on and multiple identity problems!

Problem: I need to manage multiple identities for all the different web sites I visit and applications I use.

His project is going to investigate products that provide a single sign-on feature so our managers will only need to type their passwords once instead of four times. This product will be installed on our back-end infrastructure: web servers, proxy servers, directory servers. This project is likely to cost six figures!

Anyway, I was thinking about the problem for a while, and I had a revelation. People are always going to have to provide different credentials to different web sites or applications or devices. Signing in to Amazon will always be different to signing in to check my work email.

Anything that targets the back-end to solve this will fail. The manager above will still have to log on to a partner’s web site by hand - or his hotmail or gmail account, or the mainframe, or SAP, or whatever.

So the best place to solve this problem is on the client. The client should manage the multiple identities people have, and perform the sign-ons on the user’s behalf. Mac OS X already has something like this - the keychain. It’s a pragmatic and smart solution to a annoying problem.

This is the only sensible way to help someone manage multiple identities. Trying to force our back-end servers to understand the same single identity just might work within an organisation, although even that is unlikely, but can never work across organisational boundaries.

So if we can’t get our managers to use a Mac (because of the Microsoft hegemony), maybe we could at least get our managers to use Firefox, which has a cool password manager for easier multiple identity management. Single sign-on for free, who would have though it!

There were 3 papers in my session at AusWeb (Royal Pines Resort, on the Gold Coast), with 4 presenters, the session chair and an audience of about 20 people. Mine was the second paper, I was up there for about 18 minutes or so going through my slides and demonstrating the template editor and FO output plugins. After all 3 papers, there was a general Q&A forum where all of the presenters fielded questions.

One issue raised was why we had used Java so heavily rather than a pure XSLT system (XHTML and OfficeML). Someone else asked “what does the Oracle database do?” There was also some speculation about how well the Excel output would handle 3rd party XML, formatted using a custom FO template. I explained that we did not yet have a XSL 1.0 compliant implementation, and to the extent that we had not implemented all of the FO elements there would be visual discrepancies between the ideal (PDF) output and its spreadsheet equivalent. A Web developer from RMIT also raised a question about maintaining all of the FO templates in the face of changes in the XML Schema Definition. This is definitely something we should look at w.r.t. XDO/FRM (some process for taking an existing FO template and updating it to match a revised XSD)

There is definitely a resistance at AusWeb to “death by Powerpoint” or being “talked at”. Allan Ellis, the conference chair, made the point that a format of brief presentations followed by general discussion makes better use of the physical co-location of the delegates. I feel that you are much better able to demonstrate the depth of your knowledge of a topic by fielding questions, rather than by delivering a rote presentation. It also means less pressure to “entertain the audience” and more pressure on the attendees themselves to read your paper and come prepared with intelligent, insightful questions.

In total, there were about 100 delegates at the conference although some years there has been up to 350. Besides technologists and academics, there were also a number of T&L people (educators), librarians and the armed forces. The food and accomodation at Royal Pines wasn’t too bad, and evening activities such as the trivia night and BBQ were great opportunities to meet people who were using XML for a wide variety of purposes. I will give a brief overview of some other papers and poster sessions that I attended:

Keynote speakers were not that great – the IBM guy, the lawyer but particularly the entrepeneur were clearly there to push an agenda (and/or to sell stuff). The librarian from Berkeley gave a fantastic talk, however, that covered a heap of areas and issues in his field.

13 papers were presented in the Technical & Standards track, although due to overlap I was not able to attend all of them.

XUP seems to have a lot of potential as a next-generation paradigm for developing highly interactive enterprise GUI. There is a big shift away from installing applications on the client computer, and the current trend of using ECMAScript to implement interactivity in proprietary software has some major downsides. The alternative proposed by XUP is to define the user interface using XML on the client, defining event listeners that trigger a SOAP request (either synchronous or async), which is handled on the server. Lower maintenance costs, coupled with tighter control over IP (since all business logic is implemented on the server side) could make for a very tidy solution.

DotNotelets and DotTegular are interesting from the perspective of Agile methodologies. This Web application presents a user interface for creating a hierarchy of resizable, moveable notes. Clicking on a note drills down to the next layer of detail. Although originally developed for concept maps, it seems like an excellent tool for documenting use cases (the “planning game”).

Xinq (XML-inquire) made excellent use of XQuery and the XML:DB API for rapid prototyping of searchable and updatable browser interfaces.