Gavin Terrill's Blog

Editing a Wikipedia Article is Exasperating

I am trying to edit a Wikipedia page. the one about "Business Process Automation" to be exact.

Specifically, there is a section in there that talks about tools. Now, Karora Technologies, a company I co-founded and still am involved with in an advisory capacity, has been developing one of these tools since 1998 and it has been quite successful with it. I'm thinking hmmm... Shouldn't I add a link to Karora's tool on this page? After all, there are already several other competing vendors and their products listed.

My first attempt was this:

Karora Technologies provides a product [http://www.karora.com/products/appconnector] that supports desktop mashups. Originally developed in 1998, it is now in use in many manufacturing and health care organizations helping automate organizational processes that typically span "silos" such as accounts payable or forms processing.

Mr Ollie subsequently rejected this with the reason "promotion of specific product".

Ok, fair enough. So I removed the link, and tried again:

[[Karora Technologies]] provides a product known as AppConnector that supports desktop mashups. Originally developed in 1998, it is now in use in many manufacturing and health care organizations helping automate organizational processes that typically span "silos" such as accounts payable or forms processing.

I even created a page for Karora. This was rejected ("promotional spam") and the page deleted. I would of thought that not linking to the product specifically, and providing additional information about where BPA is useful would have been Ok.

The thing that bugs me about this is that the ones that are already there are about specific products and are no more or less promtional than what I was attempting to put up there. Wikpedia Quality Control Fail. Me Fail.

ETag Article

My article on Using ETags to Reduce Bandwith & Workload with Spring & Hibernate went up on InfoQ today. I'm now waiting in trepidation to see what sort of feedback it gets. Hopefully, I haven't overlooked something obvious. It was fun pulling it together, as I don't get that much of an opportunity to get hands on keyboards these days. I am thinking that if there is enough interest, I make give this open source idea that I've been hearing about a shot and create a project around it.

Update 2007-07-20: There have been a couple of good follow ups on the article:

• Bill De Hora's compares the cultural design centrers of the Spring/Hibernate (Java), Djanjo and Rails communities: "I suspect designing for the Web instead of around it is at least as important as language choice"
• Dare Obasanjo gives a nod to Bill's article by taking a back handed shot at Spring/Hibernate: "It will use a Web application development framework that enables rapid Web development which is designed for the Web as opposed to having some HTTP related features being bolted on". Dare is fooling around with a Facebook application built on AWS. Hmm... Facebook... PHP right? I wonder if that qualifies as one of those "designed for the web" thingies?

InfoQ OSGi Interview

I was interviewed by InfoQ a couple of weeks ago on why we are choosing OSGi for our next generation architecture.

"I believe that in a couple of years time it (OSGi) will also be seen as the de facto standard when architecting IT friendly, service oriented enterprise-scale Java apps"

Read the article: OSGi for application modularity - one company's design choice.

Screen Scraping Web Service

Interesting discussion going on over here where eBay would like you to pay them so you can build a product that will help market their product. Here's a thought: Use AppConnector to screen scrape eBay, then access it via your own web service.

While we are it - I have to respond to this comment by Jeffrey McManus (head of the eBay dev team):

scraping the site is a really bad idea, not just because it’s “against the rules” but for technical and stability reasons

This is not strictly true anymore - Screen Scraping (or Presentation Integration as we like to call it) has progressed a bit since the 80's. AppConnector is smart enough to detect changes like this. Besides, I really doubt that eBay changes the names of form elements every two weeks - it would mean changing their own systems to cope with it which would be cost prohibitive, not too mention a waste of time for the QA guys.

AppConnector Ranking

Thanh notes we now get #1 billing with Google for our own product (AppConnector), and suggests we should be going for "Desktop Application Integration". I think "Presentation Integration" is a more disruptive descriptive term, but in the end it's all the same - integration done on the glass.