First Lego League
For the last couple of years I've coached a
First LEGO League team. Our team now has a
web page thanks to the new webmaster, my son!
kgb1001001 at 3:35:14 PM EST
Permalink
|
Blog about this entry
| 
Add to del.icio.us | 
digg this
This entry has comments:
Add your own
And now for something completely different...
Just in case you thought that my only interest was in work, I should note that I've started a new blog entirely on the subject of food. Cooking is one of my hobbies, and it seemed that rather than dilute the content of my website and work blog, it'd be easier to start another one. So, take a look at Urban Country Boy.
kgb1001001 at 9:32:05 PM EDT
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
New Comment Lines Article
Keys Botzum, Geoff Hambrick and I have just published a new
Comment lines article on DeveloperWorks. Brief and to the point.
kgb1001001 at 7:31:19 PM EST
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
WS-Polling Spec hits the street!
The
WS-Polling specification submission has recently been released to the W3C. Even though my name is listed first on the spec, that's purely an artifact of alphabetical order. Doug Davis was the driving force in this specification and he deserves all the credit. I contributed some of the ideas (notably in the "Status Requested" and "Status" header section) but Doug did most of the legwork. There's a good
explanation of the release at XML coverpages. Congratulations Doug!
kgb1001001 at 11:09:58 AM EST
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
Some Astrophotos -- pretty but not great
Last night I got out my trusty LX-90 for the first time in (it turns out) almost 10 months. The occasion was the near approach of Mars which is a once-every-15-months thing, and gave me a good excuse to haul out the telescope. It turned out to be a pretty rotten day for seeing in my area (atmosphere not cooperating -- and neither was my equipment) so the resulting photos I took were aesthetically pleasing but hardly clear. This one of Mars was taken with my webcam and processed with Registax. While I was looking through some old photos I came across one I took of the moon with my Meade LPI back on 4/24/04. Shows the capabilities of my telescope much better...
kgb1001001 at 8:47:08 PM EST
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
Replicating PLoP
I just had a fascinating experience this past week. For years I've been a part of the pattern language movement, and since 2002 I've been a member of the Hillside Group. We recently planned an internal conference within IBM and I took the chance to suggest that we run it like the PLoP conferences -- as a set of Writer's Workshops. I had a hunch that it might work, but it took a bit of convincing to persuade the other organizers that this was the right way to go about things.
Well, the conference worked. It not only worked, but I was able to watch the same kind of community that has been created by the PLoP conferences spontaneously emerge from people who had never participated in Writer's Workshops before. By the end of the conference we heard comments like "best conference I've ever attended...and I've attended a lot!" So what happened? Here's my quick analysis -- mostly this is a confirmation of the observations in Dick Gabriel's The Writer's Workshop and The Work of Making Things...
(1) People loved the interaction that the Writer's Workshop format provided; they said that the discussions were useful and challenging, and that they ended up learning things that they had no exposure to previously and would not have been interested in had it not been for the necessity to participate in the workshop. One comment from a senior consultant was "I would not have chosen to read this paper on my own, but I'm glad I did since I've learned a lot from it..."
(2) People responded very positively to the "politeness" enforced by the Writer's Workshop format. The "courtly" nature of the structure of the discussion and the requirement that dislikes be phrased as suggestions for improvement made people (especially more junior people, or those who are naturally shy) participate more than they would have in a more "normal" academic conference.
By the end of the conference the attendees seemed excited and energized, and are now wanting to make sure we do the conference again next year (using the same format) and are ready and eager to take on other PLoP-like activities like shepherding papers for next year's conference.
kgb1001001 at 9:45:10 AM EDT
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
Reference Architectures and Pattern Languages
I've been thinking a lot lately about the idea of a Reference Architecture. I've been asked to review several for different companies, and I've encountered several within IBM as well. As the above article from DeveloperWorks indicates, they have several things in common; they tend to try to map out the landscape of where development occurs and show which trails to take within that landscape.
Recently, two of my colleagues, Jean-Pierre Paillet and Yaroslav Dunchych, came up with a way of expressing a reference architecture that resonated a lot with me; they describe the RA that they're working on as a set of patterns.
This has gotten me thinking about the relationship between a reference architecture and a Pattern Language in the way that Christopher Alexander refers to them. It almost seems like most RA's that I've seen are trying to be poor-man's pattern languages. They mean to act as maps of the development and deployment territory, but they end up with something that is like trying to describe a landscape in text. Sometimes that approach is useful and you can piece things together in your mind, but often it's just confusing and unhelpful. The thing that most RA's often lack is the links between the decisions.
For instance, many RA's I've seen have different sections (different documents or different chapters) on persistence and object distribution. However, these two concepts are linked; if I know that I will use Session EJB's for distribution that might dispose me to use Entity EJB's more than if I were not planning on using them at all. The list of examples like this goes on and on; decisions in the representation of how you take your data to the screen (for instance -- using JSP or XSLT) may have ramifications in your domain object design.
However, this is how pattern languages are supposed to work. Each decision should lead you to the next decision. If you take a fork in the road from deciding that one pattern applies to your context, then applying that pattern creates the context for another pattern. So maybe, what we need to do is stop writing RA's and start writing pattern languages...
kgb1001001 at 5:19:12 PM EDT
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
Web Services Granularity
Recently I received a question in an email that seemed like it would be worth exploring in a little more detail in a blog entry:
"What is a reasonable number of operations per web service?"
There are at least two major factors that weigh on on how many operations you should put together in a WSDL file:
(1) Are the operations bound together in any way; be it from a business standpoint (they are for the same organization) or from a technical standpoint (they share some underlying implementation parts). If so, then you should try to group them together into a single WSDL web service file.
(2) Do the operations change at the same rate? My article on Web Services versioning discusses how it can be difficult to manage sets of operations that do not all change together. Ideally you would want all of your operations to move from one version to the next at the same time; if they do not do so, then you may create problems for the clients of your web service in that they will have to update their client software more often than otherwise.
In general, some simple rules from object design also apply in WSDL design; for one, humans can only comprehend 7+-2 things at a time. If your service has many more operations than that (which would put 20 out past the edge) then you should consider refactoring your service to make it smaller.
Some people have gone so far as to put each operation in its own web service. I think that is going too far in the opposite direction; you lose any coherence in your operations, and it also makes it difficult to understand what each service means when it only consists of one operation.
kgb1001001 at 5:01:40 PM EDT
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
Finally, a new article!
It seems like it's been forever (but in fact it's only been a few months) but I finally have a new
Article on WebSphere XD out on developerworks. This is the long awaited followup to my first XD article, and I still hope it's only part 2 of a 3 part series. Hopefully it won't take as long to write the third part...
kgb1001001 at 9:50:42 AM EDT
Permalink
|
Blog about this entry
| Add to del.icio.us | digg this
This entry has comments:
Add your own
