Alistair Jones Blog

Agile 2009, Day 1

For the first session I went to Developing Agile Leaders and Teams: A Developmental & Transformational Path with Gilles Brouillette, because I've recently taken on a management-heavy role and my ideas on leadership are are, shall we say, unformed.  Gilles has a PhD in Transpersonal Psycology, and the session was heavy on what I can assume psychologists think about all day: questions about our perception of reality, and the models that we use to glean meaning from the world round us.  It was all rather deep.
His spent most of the 90 minutes building up a model of levels of physcological development.  The top three levels are apparently very rarely attained, but are where you need to be if you're going to be a successful leader.  So how do you get to these high levels of physiological development?  Apparently it's about changing the underlying way that you make sense of the world, from an "OR" model to an "AND" model.  It sounded to me like being able to hold more than possibility in your mind at the same time; a bit like understanding a smooth spectrum of probability rather than jumping to single binary outcome.

Compact syntax for Hypirinha tables

When you're designing a builder API, you have to decide what to return from each method call. In Hypirinha I made a choice early on that most methods should add a new child element, and return that new child element. This is optimal for trees that are deeply nested but where siblings are rare, and it's a reasonable compromise for most HTML. (If if I wanted to optimised for very wide trees, where nesting is rare, I would instead make methods return the parent node so that chained method calls create siblings.)

I had a dilemma about what the special text() method should return. It's slightly different from other methods on an element because in XML terms it creates a text node instead of an element node. I decided for the first release to have a Java type for text nodes and to return this type from the text method. This seemed logically consistent with all the other methods, which return the object they create. It would also allow me to add behaviour to the Text class if I found something useful for it to do.

I've been using Hypirinha in a real application for a few months now. Mostly it's used to produce simple status pages, for example showing how many messages are in a particular state. I found we were writing a lot of simple HTML tables, and they were more verbose than I would have liked. For example:

        Table table = table();
        Tr headerRow = table.tr();
        headerRow.th().text("State");
        headerRow.th().text("Number of messages");

        for (MessageState state : counts.keySet()) {
            Tr row = table.tr();
            row.td().text(state.name());
            row.td().text(valueOf(counts.get(state)));
        }

It seems a shame to have to store the Tr objects as variables. The problem was having to call text() on the Td and Th objects; because it returned Text I couldn't use the convenience contains() method that was designed for these occasional sibling element cases.

On reflection, I realised I still hadn't found anything useful for the Text type to do, so it was fair to drop it. I changed text() to return the parent element instead of the new text node, and the problem's solved:

        Table table = table();
        table.tr().contains(th().text("State"),
                th().text("Number of messages"));

        for (MessageState state : counts.keySet()) {
            table.tr().contains(td().text(state.name()),
                    td().text(valueOf(counts.get(state))));
        }

The change has found its way into Hypirinha 0.3. I'm sure there'll be a few more small tweaks as we get further along using the library in anger.

html2hypirinha Bookmarklet

I've been playing with the html2hypirinha app for a few weeks now, and it's been very convenient to have a bookmarklet. Try dragging this to your Bookmarks Toolbar: Convert to Hypirinha.

HTML to Hypirinha, via Google App Engine

Some weeks ago Felix sent me a small app he'd written to parse HTML and generate Java source code that would build the same HTML using Hypirinha.  I packaged it up so that it would run on the new Java flavour of Google App Engine and deployed it here:

http://hypirinha.appspot.com/html2hypirinha

The source code is here:

http://hypirinha.googlecode.com/svn/trunk/java-extras/html2hypirinha

I think it's a really nice facility because it quickly gives you a feel for what Hypirinha syntax looks like.  If you were actually writing an application using Hypirinha, it gives you an easy starting point - just point the app at some a page you like the look of, and then copy and paste the code into your IDE.  For example, check out the builder for the Guardian home page:

http://hypirinha.appspot.com/html2hypirinha?url=http%3A%2F%2Fwww.guardian.co.uk%2F

Once you've pasted the code, the real power begins - lean on your IDE's refactoring tools to remove duplication and introduce abstractions.

Hypirinha: HTML from Java

For the first time in my career I've been travelling to the office each day by train.  For some reason I find my time on the train extraordinarily productive, especially when I get my laptop out.  The first fruit of that productivity is a new open source project, hypirinha.  From the project home page:

Hypirinha is a Java library for generating HTML. Hypirinha follows the fluent interface style, whereby several method calls can be chained together as single statement to achieve the desired result.

I've found the exercise really fun, and I think the the library will actually be useful.  At my current client we're using it for some minor reports and find it works well.  Using it on a large-scale web application is still some way off, but I think it's useful to challenge the usual template-based approach we see if we can find something better.