Oooh, it's been ages since I did this. This has been a week of starting things and not getting them completed - never mind, they've all been useful learning experiences...and I am meant to be on holiday you know ;-)
In the last week I've been dramatically reminded about what a difference only having a modem makes. I'm so used to 'broadband' Internet connections both at the office and at home that I forget what it is not to have resources like the CPAN, mail and even IRC at my fingertips. Here's a tip for all the middle management of the world: let your employees sit on IRC all day - I can't tell you how unproductive it is not to be able to ask a quick question and get a pointer in the right direction from knowledgeable people.
As a consequence of this I've had to change tack on what I aimed to get done with NMS. I've instead been playing around with getting the structure tied down a bit. I spent a while investigating writing a DTD, even going as far as buying and reading XML in a Nutshell (dead tree is very useful when you suddenly don't have a 24/7 net connection.) On inspection, I don't think it's possible to specify things like "any well formed elements are allowed within this tag" with a DTD - which is a bit of a problem as the ttxpath stuff NMS relies on declarations like that to allow passing of any XML that the template system's VIEW directive can render. I think I'm going to have to take a look at doing a schema instead now that acme tells me it's been made a standard and all that.
In conjunction with this I've also been doing a some stuff with The Template Toolkit. this week. Beside being overly pedantic over a patch on virtual methods on the TT mailing list at silly times in the morning, I've been thinking about way I've done the whole aforementioned Template Toolkit and XPath thing and reworking it from scratch.
Leo's been getting the new London Perl Mongers Website up and running using some of my TT code. His questions this week have helped me see a lot of 'holes' in the approach I've suggesting; Not least that it has too much in Template code and requires too many steps before it's rendered. I had a bit of a bash at solving it....and now I've given up: I've posted to the list with a few suggestions. Hopefully someone can come up with a better way of doing it or they're somehow work out how to make the changes I suggested to the VIEW directive.
Finally, unimatrix got me interested in playing around with the Klingon Language. This has inspired me with a new silly project: Acme::Klingon::Number. A::K::N is a number class that produces a Klingon translation when it undergoes string conversion. It's mainly an excuse to experiment with perl's overload feature. It's currently passing all but Soch tests out of cha'vatlh wejmaH Hut (7 out of 239.)
I'll elaborate a little more on these points in the next week or so, but for now that will do, and I'll get some sleep.
