Two mildly amusing quotes from Sci-Fi things this week...

Computer: Operation will cut life support and is not recommended. Beka: Override authorisation code "Shut up and do what I tell you" [ Override authorisation; Beka Valentine; Andromeda ]

Which I think it is a much better authorisation code than the silly numbers that they use in Star Trek. And speaking of Star Trek, Voyager had an interesting sequence with the Doctor day-dreaming...

Computer: Core breach much sooner than you might think. [ Core breach; Computer; Voyager ]

Rather than the regular 'Core breach in 20 seconds' of whatever.

Not a lot to write about in the diary this week really; lots of things been happening, not all of which I'm happy with but that's life.

I've got the universe level loading to work, but the test machine keeps running out of memory whilst doing so. To be fair it only has 32M in it, but still it's disappointing at this stage. Anyhow, I've given up on it for now. Sadly it doesn't look like it'll even fit in to the wimp slot due to the 28M limit - the AOF files that we have so far total 36M and that's not everything. It doesn't look like it's going to be practical.

I've moved around a load of the conditionally compiled sound code and I've now got a small amount if it working - it's not going to be the same as the original because of the way that I've been forced to do things. But it does at least do something. But, background music. Yay.

And I've had a quick play with some of the scrollbars code to ensure that that works now. Most of the problems here have been related to the original version expecting that 'x/0.0' being just fine, whereas we go and explode on such things. Whilst I could easily disable the FP exception, I don't think that I should - division by zero shouldn't really be silently ignored.

Anyhow, scrolly bars working, as far as I can tell. And you can scroll them with the mouse scroll wheel too. Funky chickens .

Mostly it's been a day of tidying up the input to include keyboard support, and adding some much needed bug fixes to the code. Some parts of the code rely on the memory library that they were built with having additional space at the end. My test memory library doesn't do that and so certain menus were exploding violently. I've added a little bit of checking to the memory routines as well as fixing the specific bug.

Fortunately this has been quite painless, unlike the keyboard support which was a pain because every time I added a chunk of code and tested it nothing happened. Actually getting the code to work in the test has been quite complex because you have to inject keys in the right way and my test harness wasn't quite doing that initially. It's much better now though.

The test harness itself is really quite advanced now. Probably not up to any professional standard, but there's quite a few test apps hanging off it now - there were 17 different tests when I looked a moment ago, each trying out different components, or combinations of components.

I got to see the first disc of Press Gang series 2 last night, too. Very fun; with not many downers in the entire disc. The sound quality is a bit variable at times, which may be the original recordings having degraded. The commentary on the first and sixth episodes are quite fun (hmm... a pyjama fetish ? that's a new one on me, but it is Julia that's wearing them, so it's reasonably understandable ), although it's strange to hear Julia talking about it and little bits coming back to them.

Although I don't want to put down the work they have put in to these things - because it is quite nice - it would be handy if the people involved watched the episodes once before they do the commentary. Then at least they have a chance to remember what happened, particularly when it's a long time ago. No offence, Julia, but it does seem a little odd to hear you asking what the episode's about. Yeah, I know that it was just a job for you, but we lived and breathed those episodes and ended up being able to quote sections even after only seeing it once or twice. Remember that we were quite impressionable kids .

I say 'we'. I mean 'me'. Yes, I thought that Lynda was wonderful and it really wasn't until long, long after the series finished (final series) that I recognised how manipulative and driven she was. And that's even having seen the the final episode.

Anyhow, back to the DVD; it's really quite fun and you forget some of the lines that had you in stitches first time around. The writing is mostly very sharp, and the characters are pretty great in themselves. I can't fault it mainly because it was something that I loved from childhood. Too many days spent wasting classes just waiting to get home to watch it.

It's now a few hours later and I've had a double episode about child abuse, a building dropping on spike and a girl bleeding to death, Spike and Lynda breaking up and the closure of the Junior Gazette until series three. Plus a whole private sequence reminiscing about Caroline (provoked by Colin's comment about the things Cindi said not making sense, which was how I often felt about Caroline because I'm dim like that), and a whole slew of more generalised anger directed at CTL. All in all I'm angry, tired and depressed, and still awake.

Oh, and hungry too.

On the plus side, I did try to think back to what I was doing 15 years ago and it's very hard. Partly because that places me at Adwick which I didn't really like all that much and so don't really remember. Unless I've misadded up, in which case that places me at Grove in which case I don't really remember much and the bits I do remember involve Angela who, for reasons I forget, I was besotted with for about two years. So, Julia, I'm very sorry for my earlier comments - it was a little rash of my to suggest that you'd have to remember everything. Forgive me ?

Well, I guess, that is forgive the strange written word on this diary if you should ever find that you're reading a random strange diary of a software engineer. Which, let's face it, ain't going to happen this side of hell freezing over.

Still, it's strange to think back about that time. There's no diary to remind me. Which is a good thing I guess. I believe there were some things... what I mean to say is that Claire has told me some things that I'm very embarassed about from that sort of period, but which have been erased from my memory. Probably because of that embarassment, to be perfectly honest.

Just as a large portion of Sixth Form has gone - not having been thought about for years has just left it to rot and it has finally been discarded.

Yesterday was only a part-day on the project, if you could even call it that; only about an hour and a half making the mouse work. The rest of the day was taken up seeing "Everything I Know About Men", a recording of a TV episode at Pinewood studios. It's quite cool to see, to be honest. I'm not convinced that it'll be a smash comedy, but it was nice to see how they did things.

Anyhow, today's been getting the remainder of the mouse working. I have a pointer that moves about in time with the mouse, which is kinda useful.

And now I have a mouse and pointer that lets you click on menu entries to select options.

And now I have a mouse and pointer that lets me rotate and zoom in on models.

This is looking quite funky, really .

And now I have a working options window. At least, working to the point at which I'm happy to use it for testing. It's slow on page transitions, but that's not such a huge deal.

And now I have a scrollwheel input to perform zooming.

Not a lot of screenshot-able things today, really, 'cos it's been interactivity I've been aiming for. But still, there's stuff done which is cool.

Press Gang series 2 and Dreamland arrived today as well, although I've been too caught up in doing this stuff to look at them :-|

I've integrated the 3 component tests today into one huge test in order to demonstrate parts working together. One of the larger files - the 'universe', 'cos it controls so much is still causing me problems. It just pulls in too many other components and makes life harder because of that. I spent a long time trying to rationalise it before deciding to focus on tidying up the tests and pulling in a few other components that hadn't been tested together.

As a result, the progress isn't as shocking as previously but we have almost working sanely manager display now. I need to get some interactivity working soon, which will probably take a little while to function.

Coo; I got cited twice in David Chess' log today. Utterly weird given that I remember posting those only vaguely - I know I did and from the style that it's me, but did I maybe invent the memory of those posts ? Maybe his question of gods has confused me.

It seems I've got some of the tools necessary to deal in sound, so I've spent a lot of today moving sounds around to make them usable and trying to ensure that we have a system that we can use for them. Not entirely complete yet, but effective so far.

The other thing I've been looking at has been the clouds. I don't know why they weren't transparent last night; I've changed a conditional and found I needed to initialise a math library, but it does now work better. And in brown, rather than blue - endianness of colours biting me there.

I've dropped an email to one of the guys who has been very active on this, so hopefully they'll be able to help me with the problems I'm having in extracting samples.

I've been out today, to the cinema to see 'Around the world in 80 days'. Surprisingly fun, and silly. Obviously it's a retelling of the story we know and love, but it's deviances from the main story are not bad for the story and work well. I don't like Steve Coogan in many things, but he was actually quite likeable in this; Jackie Chan is his usual self .

And today is day 11, and I've made parts of clouds work in a test. I'm not happen with them; they're not alpha-blended by depth. But it's possible that that's actually a feature of the later versions which I don't have the source to. Or maybe the software engine just doesn't do it.

After a rather annoying two days, I've now got things being drawn in 3D from the real code, both within menu screens and manually. And more importantly, they're being coloured too. Last night I was cursing for about 8 hours because I couldn't get anything to appear when not in wireframe mode. It turned out that I'd just forgotten to turn the lights on. Without the lighting, the whole thing appeared very very very dark. So much so that it was near invisible on my monitor.

Anyhow, yay for stuff rendering properly and even doing rotations. We've got static PNGs, sequences of PNGs for an animation and even a simple animation extractor so that I can show off the fact that they do rotate without requiring huge binaries. Even still it's a 2M animation. A little more than you'd expect with, say, MPEG compression, but this is Squash I'm using and we're not using differences between frames even slightly.

Otherwise today, not a lot happening. Quiet really. Except I've got a cute little cat on my arm who wants some food and thinks if he stays there long enough he'll get it. He's probably right.

We now have little lights on the model. That's taken about 6 hours to make work, unfortunately. A lot of that time was left hunting for why the lights were the wrong colour, not alpha-blended and in the wrong place. These were (for the first two) because a lump of MMX-specific code had been replaced by a simpler chunk to make my life easier much earlier, and (for the last) because I'd forgotten I shouldn't translate the model for the tests.

It's day 7 of the project and we've now got the root menu screen, pop-up menus, and a rather strange animation background animation working. All just test code outside the main application, but still quite impressive in its own little way. However, I've hit a stumbling block on another of the rendering components in that a couple of functions are in x86 assembly. This, after trying to build one of the most complex tests I've tried yet, pulling in a load of new files from the pool of those that have compiled but not been tried.

The functions look like they should be pretty simple to work out, as they're mathematical operations. I just have to be able to work out what they do, and how they do it. My x86 isn't great, but I'm confident I can do something with it.

On the down side today I discovered that I probably wouldn't be able to make the sound work. There's a proprietary library used for it, so I don't think I'll be able to do anything with it. On the other hand I could probably knock up a tool to do the decoding of the sample data on windows using their library and then just use that as the output data. I've got the objects for them and the headers to say how to use it, so I could probably knock something together. Well, if I had the laptop I could. As it is I'm going to have to wait until it returns and even then it'll need a whole load of patches adding to it to get it to the 'safest' state that MS like it at. Many days of downloads I expect.

Well, it's now day 6 of the project and we have subtitles which can be rendered - both text and logos working properly. Which is reasonably cute, but there were a lot of places where function pointers and functions were intermixed because of the old DLL way of working. Whilst it's entirely possible that a future version might use DLLs, for now I'm directly linking everything. It's safer that way.

Main menu now renders, albeit through a somewhat contrived 'HACK' harness. The number of 'HACK' elements is getting larger as more sources are subsumed into the test code. There's actually only about 8 source files uncompiled, but they're quite complex.

I've started recording a few of the screens that I've created so that there's a little record of where things are at, much as I did with the previous 'large' project.

Well, it's been a fun filled week or so.

Friday, myself, Mum, Dad and Julian went up to Center Parcs in the Lakes, dropping in on Pat and Steve's. Claire and Christopher were there which was nice to see, before the wedding. On Saturday was Claire and Justin's wedding which was very cute. I can't quite get over the fact that Claire's married now though. It just seems odd. Another odd part of it was trying to remember Justin's last name - he's always been just Justin. Plus now writing about 'Justin' as someone other than me is weird. 'Justin and Claire' Aww.

We had quite a nice time at Center Parcs, anyhow, although the rain that had held off on the Saturday did try to spoil things a little whilst we were playing Mini Golf.

Monday I was back home to find that the laptop's HD had died over the weekend and no email had been received since Thursday. The HD appears to have developed faults in the middle of part of the directory structure and left Windows unable to boot - Knoppix from CD was ok, and read the disc fine up to a point so I backed up what I can and have returned the laptop. NTL on the other hand have been pretty crap at sorting out the email problem. On Monday they said 'within 48 hours'. Today, with it still not fixed they said give it a few hours, and it's now 9 of those hours that should have been 'few' and still no email.

It's not that the account is locked, but that it actually says that 'Account is inactive'. I can dial up, use the internet, and the additional email addresses for the account are fine. It's just that justin.fletcher@... is failing. Anyone emailing me gets the unhelpful 'Invalid recipient' error message. I've just spoken to NTL again and they should definitely, definately have an answer for me tomorrow.

And it's day 5 of my project now. JPEG buffer saves are working, and I can print 'Hello World' to the buffer. In colour, too. And with different fonts. There's something rather nasty happening with the linker and weak symbols which means that it's trying to pull in the entire 20M AOF file to link small programs, but I'm not too worried at the moment; I can just build it with a limited AOF instead of the full blown version. Buttons can now be rendered, using a similar test harness, so this is looking more workable already .

Yesterday (day 4) was mostly taken up with tidying things sufficiently that a final binary would link without any problems. The problems were coming from tentatively declared variables in headers which Norcroft was thinking of as duplicated variables when linking. My reading of the C Specification was unclear on whether Norcroft is wrong or right in its handling of this as an error, so I'm not going to go either way on this one.

Since Tuesday I've been playing with a new project port. It's a biggy, but it's going to be fun. Last night I found that I had most of the main sources working, although the embedded assembler is something of a stumbling block. There's not much you can do but re-write it. Anyhow, this is day 3

However, today, I've got the 2D engine working. Colours are screwed because it's all set up for RGB555, not BGR555, but that's not a huge deal. I can just reverse the functions at some later date.

And the 3D engine works now, too. Which is quite exciting, although not really 'cos all it can do is draw a load of points so far. But it's better than a kick in the teeth. It's not like it's actually useful just yet, so it might be a while before I get a practical demo to show to anyone.

Oh, and in other news, Press Gang DVD is delayed until at least the 12th of July.