Arnaud's Blog

Opinions on open source, standards, and other things

Format vs Tool (continued)

One thing I should have added to my previous entry is that I believe the reason some people think the tool is more important than the format is because they are confusing the means from the end.

We use tools to achieve specific tasks. Because the tools are what we are primarily interacting with, tools take a prominent role and some people end up thinking that the tools are what matters most. But I believe this is wrong.

The tool is merely a means to an end. The end being to capture, create, process, communicate, and share information. The information is the end game, not the tool. Making this distinction is fundamental. The tool is merely what we use to manipulate the information which is really what we care about.

In this context, having a standard format that can represent the information is tremendously more important than any specific feature a particular tool may have. In fact, having a standard format enables the information to be manipulated using different tools, allowing you to change tools based on your needs and what is available. This in turn leads to having more features at your disposal.

This model is undoubtedly more powerful than being stuck with a single tool, not matter how great that tool may be at a given time, and depending on a single vendor to provide you with all the features you may need or want. Having a standard format enables competition which leads to more innovation and greater tools.

I know not everybody agrees with that last point; some people think that standards stifle innovation, but I disagree and plan to discuss this in a future entry.


November 21, 2007 Posted by | standards | , , | 1 Comment

Format vs Tool – where is the value?

At Goscon last month, Jason Matusow of Microsoft, stated that what matters most is not how information is stored but how you access it. According to Jason the real value is in the tool and this is what you should worry about; the format used to store the information is an implementation detail.

I understand why Microsoft would say that in light of the increasing demand for open standards like ODF. When you enjoy a quasi monopoly status you don’t necessarily want to open your formats and enable competition. But it remains that this argument appears to me as terribly retrograde and at odd with the era we’re in.

Contrary to Microsoft’s claim, I think the tool is no longer the center of interest, the information is. When I made that point at Goscon somebody in the audience applauded and I’m confident that this view is shared by many people but experience shows that what I think is common knowledge is often not. I’ve also learned that only through repetition things eventual sink in. So, I want to discuss this a bit further. Hopefully this will have some value even to those of you who are already convinced.

We’ve all used tools that function like black boxes. You use a specific tool to generate information, and you use that same tool to retrieve that information back. The information is literally imprisoned in some form of storage only known to the application you’ve been using.

When you think about it, if you create a book using Microsoft Word, even though the content of the book is yours, you are not free to access it the way you want. You can only access your own book through Microsoft Word.

But this is a model of the past. It was ok when all we did was to create documents that lived on one computer and stayed there, when sharing a document meant to print it and mail or fax it. But this is no longer acceptable in a world where information is primarily destined to be shared via some digital media, email or other.

The web has demonstrated the power of separating the way the data is represented from the application we use to access it. It is thanks to standards like HTML and CSS that we can all browse the web independently of what computer and browser we use. It is thanks to these standards that people can use whatever hardware and software they like to create and deliver web pages.

Similarly, having been using ODF for a while now, I’ve experienced first hand the pleasure of being able to try new tools as they come out, and switch tool depending on what I’m doing and my liking, all the while without having to convert my documents from one format to another. It may sound like I’m preaching but it is very real. Freedom is exhilarating!

There is no doubt in my mind that people who have had a taste of the freedom provided by this new model of separating the data format from the application will no longer accept the old model. They will no longer accept a model that ties their information to the application they happened to use to create it.

Those of us who are old enough to have known the old model will keep wanting more freedom, and the younger crowd will simply expect it. The future generations will demand it, and will reject anything that doesn’t respect what is fundamentally a right. The right to access YOUR information the way YOU want.

November 20, 2007 Posted by | standards | , , | 2 Comments

CDF and interoperability

Andy Updegrove published an enlightening piece on why the recent claims from the founders of the OpenDocument Foundation regarding the W3C Compound Document Format (CDF) have been puzzling many of us. I just want to add a tidbit of information regarding CDF which is in line with my previous post on XML vs Open.

CDF is just another piece of technology that helps raising the level of interoperability achievable between software components exchanging XML data. It provides us with a formal way of describing how various XML vocabularies are being used together. This is definitely useful and that’s why IBM, for one, has been participating in its development. Yet, this is no magic bullet either.

CDF is merely a framework, a container. As such, CDF itself does not ensure interoperability. Interoperability can only be achieved with regard to a specific “CDF profile”. A CDF profile lists a specific set of XML vocabularies and how they are to be mixed. Interoperability is only achieved between applications that support the same CDF profile(s).

This is applications that not only support CDF but also support every one of the XML vocabularies being used in that particular profile as well as the particular way they are being used together (CDF supports various combination models).

I’m sure you’ve had the same experience as I have with video files you can open but your media player won’t play because it doesn’t have the right codec. That’s the exact same problem. The MPEG video format is a container that lets the player discover what video compression is used in a standard way. This is nice but, as experience shows, it doesn’t guarantee that your player will be able to render all videos, merely that it can figure out what’s in the file and whether it can render it or not.

So, again, let’s be careful not to jump to conclusions too fast. Just like XML itself and many other technologies, CDF is useful but it does not in and of itself guarantee interoperability.

November 9, 2007 Posted by | standards | , | 1 Comment