Wednesday, November 27, 2002

My Intent...

Both Jorgen Thelin and Mark Baker referenced a recent posting of mine Glad to hear... They both brought up some good points that make it worth a revisit. To help Mark I will tie it back to an aspect of web services and/or REST depending on which side of the debate you fall on.

I feel architecture by intent is best applied to system level problems rather than individual components. An ideal place to apply it is considering the intent of a service in a service orientated architecture. A major issue in designing services is the appropriate level of granularity to expose, one rule of thumb is that it should be explainable in business objectives rather than technology. A good measure of the business objectives are the intent of the service.

Tuesday, November 26, 2002

Semantic Web....

The Semantic Web has been getting discussion again - one of the more interesting is Spontaneous intelligence and the Semantic Web by Phil Wainewright where he calls into question the AI aspect of the semantic web. The semantic web, like many other approaches before it, will not create intelligence but will organization information better to enable us to draw more intelligent conclusions. Though if we do not have the good analytical thinkers to use the information....

Saturday, November 16, 2002

Glad to hear...

One of my favorite topics merged with one of my favorite sites today. Slashdot has an article about how the 2.6 Linux kernel is going to be faster than the 2.4 kernel. Am I too optimistic or might this reverse the current trend. I think both the major Linux and the other major non unix operating system vendor are guilty of requiring hardware upgrades to support software upgrades.

This is a design problem - throwing features into a product without considering the implications is just being lazy. We are at the point where we do not need to upgrade hardware platforms every 12-18 months unless poor software forces the issue.

We are not considering architecture by intent rather we are developing architecture by accidental consequences. Architecture by intent is the process of deciding what you wish to expose to users and how you want to expose it. Deciding to throw in a more features and not considering the impact is not an intent - it is unintended consequences.

Is it not about time we consider architecture by intent ?

Monday, November 11, 2002

Changing the landscape rather than following the established path

There was a recent article in the New York Times and then Slashdot Operating Systems are Irrelevant, the article is part advertisement part call to action. I agree to some of the key points summarized in the Slashdot article.
  1. Operating systems should be irrelevant
  2. We should be able to access our data anytime, anywhere.

I installed the software mentioned in the article and played a little with it - the basic idea I get but I am still trying to see how to use it effectively (RTFM?). What I think is more relevant is that the software runs on Windows only. If you are trying to create a revolution you need to associate with revolutionaries.

Where are the revolutionaries though? One problem with the whole Linux desktop approach is that it is trying to copy Windows and emulate it. OpenOffice is great but apart from a few nice features its main advantage is cost; which I think is an advantage, but to change the world there has to be more than dollars involved.

Instead of arguing over the merits of KDE and GNOME the open source community should be trying to revolutionize the environment and make the operating system irrelevant. Removing the need to understand c:\ or /usr/local for most users would change the computing landscape. KDE versus GNOME will only slow down adoption, remember OpenWindows versus Motif.

If the open source community wants to change the world the operating system should be irrelevant and the user experience should be dramatically better rather than a copy of a copy(Windows) of a copy(Macintosh) of the Xerox Star. In this the author, David Gelernter and I agree, we are working with a metaphor that has its roots in the early part of the last century, the file cabinet.

There is an incredible amount of energy and talent in the open source community, could it not be focused better than creating another file folder metaphor?

Wednesday, November 06, 2002

Print media and Blogs

A while ago I posted an entry suggesting that blogs may be the death of the technical trade journals Where is the information. I subsequently had an email exchange with Jon Udell where he disagreed and stated that he was going to try to change my mind. Well I would like to congratulate Jon and his colleagues at Infoworld for starting to change my mind. They are starting to blend blogs and print in an useful way and I see a synergy among the blog world and Infoworld. Anyone seen other trade magazines doing this or is this unique to Infoworld?

Tuesday, October 29, 2002

Mapping blog relationships

I came across a tool a few months ago that I found interesting Kartoo, last night it occurred to me that it could be used to visually map relationships between blogs.

It is pretty cool - go to Kartoo type in your favorite blogger and a map appears of relationships. One of the more interesting things is that you can see the concepts that link blogs. The search can be refined by adding or deleting concepts - hours of endless fun. Other interesting things - put in the names of two bloggers and see how they are related.

Apart from being fun this is also an interesting way of mapping the spread of ideas and/or concepts, if the search in Kartoo could be limited to just people with blogs then idea mapping could become possible. How about a private search just using RSS feeds - interesting anthropological experiment conducted in real time.

Now if I could just calculate the degrees of separation....

Thursday, October 24, 2002

Standards and Profits

Standards and Profits

I have been on panels, and in interviews with analysts and press over the last few weeks and the subject of standards (and pseudo standards) has come up repeatedly. The software industry is obsessed with standards and the politics of creating them. Perhaps CS departments should offer a specialization in standards!

The confusion comes from the fact , people assume companies are creating standards for altruistic reasons. This is not to say that many individuals are not working very hard to create a better world through the standard bodies and I applaud their efforts.

Companies are out to make a profit and produce returns for their investors. The standards process is just a way to maximize this return. A good rule of thumb is the more effort a company puts into a standard the higher the return they expect and hence the more skeptical we should be of it. It is totally appropriate for us to question our own and others motives for creating and supporting standards. "The market demands standards" is a true statement, customers like standards but the actual standard they get may not be optimized for customers, it may be optimized to generate more revenue for the vendor.

Standards are great and we need them to make the development process easier, allows products to interoperate and be more secure, but lets be pragmatic and look at individual standards to see who benefits the most.

Friday, October 18, 2002

Identity and idea distribution through Blogs

Identity and idea distribution through Blogs

One of the subject that is consuming a lot of my time at work (apart from travelling, hence the lack of blog activity) is federating identity between organizations. I have found that my best source of raw material has been blog. Here many smart people are freely sharing ideas, (thanks John Udell,Phil Wolff,Eric Norlin). The impact of this sharing is hopeful lead to faster and better decision making for many of us.

It would be interesting if we could create time based context map of ideas in blogspace - see how ideas flow together morph overtime - something for the RSS aggregators to think about.

Thursday, October 10, 2002

The pillars of integration - a shaky foundation

The pillars of integration - a shaky foundation

EAI and other integration technologies assume the existence of information silos that need to be integrated. This is the accepted wisdom today. More and more complexity is built on this assumption, when the problem is really the information silos. Building more complexity on top of complexity does not make the problem go away, rather it makes the problem more complex.

Instead of trying to integrate complex silos of information - should we not be considering why the silos exist in the first place? If the problem of the isolated silos of information was removed integration would be significantly simpler. Make information easily accessible, then applications can become transient creations, the whole dynamic of the enterprise changes.

I believe this is where the next wave of innovation will come from, it is a wave we desperately need!

Wednesday, October 09, 2002

Shape of Blog Space

Shape of Blog Space

This morning I took a little detour in blog space off my usual track and spent an enjoyable 45 minutes wandering around in various minds. Having been reading "Linked: The New Science of Networks by Albert-László Barabási " recently ( highly recommended for those of you interested in networks), I thought there may be a semantic map in blog space that relates to the map of the Internet.

If we want build software to optimally extract, collate and present the thoughts of blog space we really need to understand the semantic map. Otherwise we will again be reduced to brute force approaches to the problem of the information explosion. This would I think be a step backwards as there are a many interesting views being presented in the blogverse. Perhaps this is a good focus for the current RSS/RDF debate which appears to be a debate over style rather than substance. If there was a real problem that needed to be solved the debate will become more concrete.

So I pose the problem, can RSS move towards a semantic mapping tool from a headline aggregator, could we use it to fly through a conceptual map of blogspace. The possibilities seem endless if I have hooks to tag RSS descriptions I can create my own personal conceptual map, getting a little far out there can we start to extract true meaning from our collective thoughts. There raw material is there, we create it everyday it just needs the infrastructure to make sense of it.....

Tuesday, October 08, 2002

Reliable Web Services

Reliable Web Services

While I agree to Jon Udell's posting Reliable Web Services that web services will converse in a loosely coupled asynchronous manner. I do disagree that this requires message orientated middleware to happen. Reliable message delivery is a function of the protocol among the nodes not the transport, i.e., you can build reliable messaging using UDP Tibco/Rendezvous or over HTTPR, I have even seen JMS over HTTP. Therefore to make SOAP more reliable it is more around defining the semantics of the interaction rather than the protocol that SOAP is bound to. This is similar to JMS, which is an interface rather than a protocol and the messaging semantics are defined at an abstract level allowing the implementers to use any appropriate transport.

It is going to be interesting to see if web services develop in to a rich set of XML standards for messaging or does it become relagated to being a hitchhiker on a reliable transport.

Monday, October 07, 2002

Like minds leading to Polarized Societies?

Like minds leading to Polarized Societies?

I received some positive feedback from Stephen Dulaney (thanks Stephen) after I linked to his blog. This started the thought process of how communities form, I linked to his blog because we appeared to share the same interests, read similar books, and I found his entries insightful. One of the issues, if we all start linking to similar minds, is do we create polarized virtual societies. Therefore should we all create “anti-minds” lists and compel ourselves to read them once a day – in Roman times when a victorious general was given a triumph a slave rode in his chariot whispering in his ear “you are still a mortal”, something to bear in mind as we all join/create like minded communities.

Saturday, October 05, 2002

Where is the information?

Where is the information

Years ago we relied on magazines and periodicals for our information, that has quickly shifted in the last few years to news sites and now to blogs. All my current knowledge comes from blogs of one form or another (I include Slashdot as a blog). Why magazines now becomes the question and how to synthesize the information in blogs or is the distributed nature part of the fun, like delving into an used book store looking for unknown treasures.

Regardless blogs are changing the face of information dissemination, the web increased the speed of distribution of editorial information but blogs are moving raw ideas at the speed of light - how will this change our world?

Successful panel discussions

Successful panel discussions

I have both been on and listened to several panel discussions recently - as well as the usual amount of solo presentations. The most informative and useful information has come from well organized panel discussions as they compare and contrast ideas. Some panels work well some do not, to help understand I have come up with the three rules of successful panels. (There is actually four but I hope that is obvious to everyone - come well prepared! ).

John's guide to panel discussions:

  • Controversial or relevant topic that has distinct sides, broad enough to be interesting but not to broad to loose focus

  • Panelists with divergent views: panels do not work well when everyone agrees

  • Well prepared and provocative moderation, if the moderator has the first two in place they can create dynamics that both inform and entertain the audience

Follow the rules and panels well not only be informative but also interesting. Anybody got other ideas?

Saturday, September 28, 2002

Information drives revolutions..

Information drives revolutions..

We are currently in a lull for the software industry with few new ideas and few successful companies emerging. This has been a topic of discussion recently, some of the blame has been on the economy, some on VC's reluctant to fund new companies. I do not believe either are the root cause. A down economy is a good time to start a company with a new business concept, entrepreneurs and VC's know this. It is all about timing, build the company in the downturn when resources are cheap and available, you are then positioned well to take advantage of the upswing.

The truth is that revolutions require new information flows, to be stimulate the revolution.Today we are moving around the same information and trying to manipulate it in new ideas. (This is part of the recent hangover, so much money was poured into the Internet that we pretty much exhausted most good and bad approaches). It is very hard to extract new value by moving around existing information and trying to create new value. Therefore when looking at new ideas the key metric should be does this release new information or combine existing information in an innovate manner.


What are the criteria that would indicate a technology is going to be transformational, rather than just another twist on an existing approach.When looking at a new idea, it is useful to run it against a few indicators that will help judge whether it will be revolutionary or just mediocre.

  1. Does the technology allow existing information to be collected and assembled in new ways at a lower price point?

  2. Are there new sources of information being created, or does the technology allow new information to be created easily?

  3. Are large amounts of unstructured information being moved into structured data and made accessible through new interfaces?

  4. Is there a new way of distributing the information that will empower a new group of information consumers?

  5. Does new information get delivered to audiences that were unable to receive it before at a price that is appropriate?


Well I think the barriers are caused by propriety information formats and the construction of isolated information silos that cannot interact with other sources of information. The problem still exists of asking the wrong question which is how do I integrate the information silos rather than how do I eliminate the information silos to enable easy access to information. Once information is easily accessible the ability to create/modify applications becomes dramatically simpler and the above indicators become active.

So what is the next thing....

I believe there are two potential revolutions either underway or the foundational pressures are being created to start the revolution:

  • Connecting organizations across firewall boundaries, this satisfies the requirement for new relationships between existing data and in some cases I have seen this is moving unstructured data to structured, unleashing the ability to automate the decision making process

  • Moving to loosely coupled structured data in the same way the web moved us to loosely coupled unstructured data

Friday, September 27, 2002

Ideas are free but execution is key

Ideas are free but execution is key

Over time I think we all begin to understand that while ideas flow like wine, execution is key. How many of us have espoused revolutionary ideas, but have not followed up and several years later seen them born to a different mother and grow into well formed companies. So take your ideas and nurture them and fulfill the vision - do not leave it to others,,,,,,

Thursday, September 26, 2002

Is Workflow natural - if so why is it hard?

Is Workflow natural - if so why is it hard?

After posting last night I sat and thought of some interesting ways to organize the flow of information in my life. It was a workflow problem but installing and learning a workflow product does not really excite me - assuming there is a good open source one. Considering other problems they too were workflow problems, is workflow something we all need? Is the ideal desktop metaphor, save, create context, organize,flow, publish,browse? How about we make the information self describing and separate the content from the description...So I have OpenOffice, Mozilla, XML Schema, etc, I just need an organizing tool (Lucene?)and a flow tool and all my information will be managed, or am I talking about the semantic web, and it already exists?

Wednesday, September 25, 2002

Are VC's and Entrepreneurs the next crash?

Are VC's and Entrepreneurs the next crash?

Well we are almost over the .bomb era, we have all taken our lumps and are ready to get on to the next new new thing, but is it there? Does the cleanup have a few more steps to take before we can say we are ready to start again. The number of VC companies with large amounts of cash ready to spend and the large number of professional entrepreneurs seem to be out of balance with the true amount of innovation we can create. Is the last phase of the .bomb a massive shrinking of the number of VC's and entrepreneurs?