Well, busy, busy, busy. A flurry of activity is occuring at my end of the line. Between idling in #flipcode, idling at university and various studies at home my time has been a scarce, yet rarely precious commodity. Still, I have managed to do some coding.
One of the more interesting things I've been doing is preparing for my internship, which, much to my bank accounts distress, is in another country. Mitsubishi Electric has decided to let me toil away brining the people doing research coffee (and coding and stuff) in their Research Labs (www.merl.com). So, I should now be staying in the Boston/Cambridge area with my g/f Candace for 3 months (starting December).
Code wise, I've been doing various interesting framework experiments and implementations for the engine. I've been engineering the framework very carefully. I've recoded various bits numbers of times. I'm finally getting it to a point that I feel actually will work in a quick and flexible way. That is, I've algorithmically optimised the framework design at the start of the project. This is something I've missed doing in the past - Much to my later frustration. For me optimisation was generally a post thought. But, I've found that careful algorithmic optimisation of the framework can also cut down on code hassles.
The general advantage of doing this though, is that various hair pulling out at a later state is avoided. Careful design of things like interfaces, and underlying frame work avoids you having to change them later on - And the great amount of code you have based on them. Much to my chargrin in the past, I've decided to kludge things, rather than spend a while re-doing something that should've been done right.
While the above sounds obvious, often its not. I design the engine structures and often a bit of the architecture, but not its implementations. Now that I am, and I'm using a rigid enforcement of the rule, I'm feeling a lot more comfortable.
That doesn't mean I've been sitting idle on architecture though. I've done a few tests of different things, and I'm in the midst of finishing one of them now. Something I nearly finished, but put on hold for a deluge of school work, then the flurry of design coding was my bi-parametric composite engine. It is like a texture shading style thing, but takes surfaces defined by 2 parameters and combines them in odd ways. Blends etc. It also supports mono-parametric functions/waves etc. I've done all the simple blending functions (in all there mono/bi parametric permutations) and an interpreter of the binary codes used to generate the shapes (this is quicker than it sounds). I'be been able to make some simple shapes, but I've not finished all the parametric functions, or the more advanced sweep blends and the ability to use things like directions of a surface/curve/wave as parameters of other functions (for various reasons). Also, I've got to design a scripting language. (This will be extensible via a meta extension language - But thats hard to explain).
And now, for the cameo appearance. Bored on day in a lab at university, I decided to search in google, on various people, to see what came up. I found one, I felt obliged to share, about one of my ex-work mates:
Spice Girls Guestbook
... I love you "SPICE GILRS". Name:Tane Piper Email:email@example.com Location: Kinghorn, Fife Scotland Date: Monday, March 31, 1997 at 14:31:21 Comments: ... mesmdec.me.ic.ac.uk/~rwd/spice/Guestbook_3.0/mar97.html - 46k - Cached - Similar pages
Conor "DirtyPunk" Stokes