WebMink

Entwined Issues [WebTech]

I have respect for both Tim Bray and Joe Wilcox. So how should I react to the comments by one about the other? Responding to Joe's article on C|Net, Tim is quoted on Scripting News saying:
This story is just silly and technically illiterate. The XML tags are already in the hands of thousands of beta developers. XML is an open and quite self-describing file format. Who needs the damn schema?
The problem here is one I've been expounding in the vigorous debate over on Sam's blog. It seems to me there are two issues being blended here, causing confusion. Tim is responding to one of them; Joe is expounding the other.

First is the XML expression of the long-time desire of the SGML community to have tools that create and respect markup within documents. Office 11 appears to be coming out with support for this capability and much of the discussion seems to be about the ability to use Office 11 to create and view arbitrary XML. It's to be welcomed, no doubt. The SGML/XML community has been promised it before and are quite rightly delighted, as long as it happens this time.

Second is the ability of the normal user to open and save their work in formats that can round-trip to other users with full integrity and function throughout the journey. As far as I can tell, Office 11 makes no concessions in this direction as it neither supports the formats others use nor documents the formats it uses itself - see Sean's comments. This is the area the OASIS office formats TC is trying to address.

Dare Obasanjo of Microsoft says:
Office 11 supports LegalXML, OPML, RSS, DocBook, XBRL, etc
but the problem is, as JDD points out, there's no other word processor used by ordinary folk (the ones who pay for all the Office licenses in existence and hence pay for Microsoft to stay in business) that either has its format supported or is likely to be able to read and write the default format in which Office will store its files (which as I understand it is still a proprietary binary format). Or am I wrong? Will the new Word save files by default as XML? Will the format used be available with no strings attached so others can support it? Will it be contributed to the OASIS TC? Watch this space...

Meanwhile I'm with Sam Ruby:
I still think there are two sides to this story. On one side, the ability of MS tools to adapt to formats that users can describe will be an incredible step forward. On the other hand, this doesn't explain an unwillingness to working with others to describe the semantics that PowerPoint 11 uses to capture a pie chart.
It seems that the great new features being introduced for power users and web services integration (my first area above) are being used as an excuse not to join the OASIS TC and not to provide interoperable productivity apps for the average user (my second area above).

posted at 6:07 PM (UK) | Comment? (0 so far) | links to this post | |

Monday, December 16

Locked in by open services? [WebTech]

We need to be constantly cautious as we engage with Web services. They offer a fast-path to lock-in unless we stay aware. For example (and without any criticism of Amazon intended), if I create a program that relies on the new Amazon.com web service interface, I may be using an open technology in the form of SOAP, but the data flowing through the service is in a format used only by Amazon and reflects their business model and world-view - indeed, it is their proprietary property. If, in the future, I wanted to use multiple or different suppliers, I would have to rewrite all my code to use the data format they preferred (presumably using their world-view).

The openness of SOAP does not prevent me from being locked-in to the vendor's business model, object model and data formats. Indeed, web services can provide a hyperdermic needle through which the foreign DNA can be injected into my systems. As an industry, we need more initiatives like UBL if we are to be free of vendor lock-in.
[Originally written on the front page on September 25, 2002]

posted at 4:45 PM (UK) | Comment? (0 so far) | links to this post | |

Defining Interoperability [WebTech]

Full marks to Joe Wilcox in seeing through the techno-FUD to the real issues in his insightful piece on Office 11 today on C|Net. The issue with Office 11 is not whether it uses XML or not - who wouldn't - but whether the way the XML is used creates room for true interoperability (and it seems Sean agrees...). Microsoft in this case defines interoperability as "the ability to interoperate with other members of the vendor's product family (possibly including partners)" whereas what I think we're all looking for is initially "the ability for different tools serving the same purpose to losslessly exchange files". The latter is true interoperability today, the former is the benefit we are being sold under the guise of 'web services'. Tomorrow, true interoperability will also include the provision of replaceable net-connected services, but as I have said elsewhere that's not on the agenda for web services yet.

posted at 4:42 PM (UK) | Comment? (0 so far) | links to this post | |

(cc) [SocioPolitico]

A landmark day today for the world of intellectual property protection and the Creative Commons with the launch of their new DIY license selector and of "Founder's Copyright", a scheme to reduce copyright periods voluntarily to 14 years which I would love to see widely used in the IT industry where publications are typically discardable after 14 months. Once again we see Tim O'Reilly giving a lead (although admittedly his risk is probably small considering most of his books fall into the '14 months' category). The huge stride many of us are really waiting for, though is their "intellectual works conservancy", hopefully a protective home for key software assets as well as for learned tomes.

posted at 11:40 AM (UK) | Comment? (0 so far) | links to this post | |

Sunday, December 15

Ugliness of the Unfamiliar [WebTech]

I was sitting helping Alastair with his homework just now and encountered a common problem. He was typing an essay for his English homework and was cursing the word processor because, unlike the one he was familiar with, it didn't have the font he wanted. The culprit in his mind was OpenOffice.org 1.0.1, the office suite I have installed in the kitchen (shared) computer. In fact, the problem was that the computer is new (the last one had a motherboard failure) and the font he likes, 'Wide Latin', is not one of those installed by Windows 98 by default. But in his mind, the issue was that the package he was familiar with, Microsoft Word, was not installed on the computer and he was being forced to use something apparently inferior because it did not have the font he was used to.

It's my belief that this problem, the ugliness of the unfamiliar, is the biggest threat to the adoption of non-monopoly software. Every time the user finds something unfamiliar, they treat it as bug in the new rather than either learning the alternative approach or discovering the true defect in the underlying system. It's depressing to see this week all the stories about the critical defects in Microsoft's virtual machine being headlined and reported (two different activities by two different writers by the way) as bugs in 'Java'. The story linked even tells users to do the worst possible thing in response:
Microsoft said the most serious of the flaws could enable a website to compromise a user's system and perform actions such as changing data, loading and running programs, and reformatting the hard disk. According to Microsoft's website the vulnerabilities range from one that is rated "critical" to others that are "important", "moderate" and "low." It said users should scrap older versions of Java VM and install the latest version.
Users would be much better served to install the Java Plug-in from Sun, which avoids the problem by complying with the Java specifications.

The really grievous issue here of course is that the bugs Microsoft is fixing relate to the use of their platform integration, built in violation of their contract with Sun, proving to give unauthorised access outside the sandbox and into the host computer - something I remember commenting on at the time as unhealthy and dangerous (I worked for IBM at the time). I, we all in the Java community, were right to be concerned then and it's clear that Sun were right to take court action against the issue too.

posted at 7:49 PM (UK) | Comment? (0 so far) | links to this post | |

Quiet week [MetaBlog]

Well, that turned out to be a quiet week for blogging. As most of you know, my activities are distinctly seasonal, with speaking engagements aplenty during spring and autumn and lots of thinking time in summer and winter. This week was the week of my one December speaking engagement. I was at a conference in Brussels at the European Commission convention centre discussing the possible future convergence of web services, grid computing and the semantic web. Apart from that I have been severely distracted by installing a Linux system at home and discovering all the Unix skills I had back when I worked for Burroughs are still viable. So apologies for the lack of breathless prose this week.

posted at 7:38 PM (UK) | Comment? (0 so far) | links to this post | |