Wrote a first draft research proposal and gave it to Richard and Geoff.

Alison Finnigan got in touch via Kaye and told me that my scholarship will start in July rather than June so that I can get 6 semesters worth of scholarship instead of 5 semesters and June. This seems like a good deal to me :)

Created my BibTeX bibliography file and added all the papers from the back of my book. Looks good although I am still unsure how to reference Maya and Mirage and 3D Studio Max.

Considered getting 3D Studio Max from Jayson and Peter's Russian friend, decided that I should just go to ARL and have a play with it there sometime, if possible.

Examined the QWD format and concluded that I need to know what the format of the first connect packet is in order to do anything... should be in the mikeBot code.

Tried to get Tracey to install G3 Ethernet 2.0.4 on the 300 Macs so that Oli and I can play Myth this weekend.

Such a data structure might well be able to update polygon-graphs for CSG in realtime. I hope so, as then we can do interactive CSG like I want to!

I'm thinking of playing with POVRay's object description language to see what it's like.

Did some more serious thinking about the octree structures. It looks as though the algorithm might work reasonably fast, hopefully it can be optimised to realtime speeds for medium-complexity objects. Also, looks to be O(n) with respect to primitives in object!

I'm busy setting up the vote-for-finalists competition, that is going well. Ideally the automatic generator thingy will be done before Friday, when Round 2 starts.

Tested Mathew's CD-R device by plugging it into my primary IDE chain (as slave). De nada, the **** thing crashed after 226 MB written on the test CD. Looks like the CD-R unit will have to be replaced, which is bad.

Steven gave me an idea as to how I can speed up the octree real-time polymesh creation.

Played some Quake with Julian on Xtra, got frustrated by lag patches and Aussie LPBs so left.

Ideas for the proposed system are solidifying nicely in my head. Should be a possible project!

Incorporated Geoff and Richard's comments into the second draft of my research proposal.

Chased down Geoff about being paid for Mirage tutoring... apparently the departmental secretary is handling it.

Mirage tidyup, got rid of the ****ing RCS stuff that I haven't used in the last 18 months. Also moved the example code into the examples directory where it should be! Created an ASSERT macro.

Checked out OGR's E3 coverage, very nice.

Wrote the 3 CDs for the 3 copies of my thesis.

Handed the final bound thesis into R&I, but I will have to tidy it up tomorrow still. Not done yet!

I played a lot of Myth over the weekend, great fun!

Strategy film analysis didn't go ahead as planned, rescheduled for Tuesday 9pm PDT (Wednesday 4pm NZST).

Went to the Science Library and played with the Science Citation Index. It's an amusing tool, but it doesn't index SIGGRAPH, Computer Graphics, Computers & Graphics, The Visual Computer, Computer Graphics Forum or a whole swag of publications that I wanted to look up articles from. Grr. It does have IEEE CG&A, though, which is better than nothing.

Went to ARL to talk to George (who's getting married) about research and sources of publications. He had a few good ideas but they all basically come down to 'Go to the library and stay there until you've read all the graphics journals'.

The new memory for Othello and Desdemona will be arriving as part of the next equipment grant, whenever that is.

Hunted down Mike Muuss of the US Army Research Laboratory. He's written a program called BRL-CAD that uses n-Manifold Geometry techniques to create and maintain polygonal tesselations of CSG objects for interactive manipulation purposes. It looks nice... I've filled out their license agreement and will fax it to them on Monday.

Results are in for Bracket 15 of Round 2. No real surprises there.

I'm still trying to organise Bracket 5... aaaargh, scheduling hell.

Gave my thesis (2 copies) to Brian, who will post it to the external examiner.

Found a very important paper in SIGGRAPH 97 that basically implements a system very, very similar to the one that I want to create. Hmm.... well, I'll chase up its references and then decide if I still have a PhD.

I've read a large number of papers today... eeek. I'm currently suffering from poor scholarly writing overload.... but it should pass.

This paper by Rappoport and Spitz scares me more the more I read it. Oh well, I still have some interesting ideas that I can work on.

I don't know if I want to join Bloodrain (the NZ Quake CTF league) or not. It's a large time investment, and I don't find team CTF that much fun any more. Well, I'll see when we get QW 2.21 installed on these computers.

The Shakedown vote-for-finalists competition is over now. 71 entries... not bad, not bad. On that note, the Shakedown website passed 2000 hits last night, woo.

In fact, there's nothing really relevant to my work in any of the SIGGRAPH conference proceedings all the way back to 1981. Time to start looking at other journals.

Reviewed The Visual Computer, 1990-1998. A few useful papers, including one excellent one by this same van Emmerik guy... he seems to be a real guru in geometric constraint modelling. I must look into those references later.

Shakedown stuff continues. Bungie guy Jim says that I can have a Myth soundtrack CD to give to the winner of the pick-the-finalists competition. Oh joy.

Reviewed more of The Visual Computer, 1990-1985. Found a useful CSG reference from Geoff Wyvill in 1985, and a few UI papers.

Reviewed Communications of the ACM, for several years back. It seems to be mostly related to IT stuff and I couldn't find any computer graphics articles so I stopped at about 1992.

Processed another 2 brackets worth of Shakedown results. Meme is missing films for games 3 and 5, so I guess I'll just post an incomplete filmset.

Geoff pushed the meeting back until after the examiners' meeting.

Reviewed most of Computer Graphics back until 1991. Some interesting papers.

Had a good meeting with Geoff, very thought-provoking. Must investigate normal-form CSG graphs and disjoint set theory... ick!

My butcher@corporation.org email address has stopped working... grr. This does not bode well.

Created a small alias loop that kills people's Netscape processes when they log out - this should be happening anyway, but Netscape is buggy and sometimes it hangs around after logout. Hopefully no more rampant Netscape processes!

Dave setuid qwcl2.21 on all the other machines, which means that Julian and I can play QW 2.21 with AWOL. Oh joy. They have a practice this evening, I think.

AWOL practice was pretty brutal. We got thoroughly hammered, although the defence stood up reasonably okay on E4M3, which is the important map. I think we can take Void this weekend, I really do.

Myth is a lot of fun when played over a LAN. Unfortunately our Mac sysadmin is a buffoon, so 3 out of the 6 G3 macintoshes wouldn't even boot up. The other 3 were rigorously playtested, along with 2 of the slower Macs.

Hey Jimmus, you mentioned that you have got the System Shock CD back from Ed. Could you try and convince Peter to bring it down to Dunedin next time he's around? I'd quite like to have another bash at it.

Shakedown Round 2 is pretty much finished now. I've posted Round 3 brackets (but hosts aren't finalised yet, neither are two of the brackets).

Transcribed all of my references from my scratch working at the front of the book into the bibliography at the back of my book. Confirmed pages, volume info, etc. Also tidied up loose ends - I now have 3 avenues of litreview to complete.

Thinking about research proposal and modifications that should be done.

Dang ICQ is crashing continually. It's really frustrating, maybe I should just use email or IRC instead.

The Shakedown web host appears to have died. Aaaargh! What do I do now? I've spoken to a few people who might be able to help me rehost it.

Reviewed a good fraction of IEEE CG&A from 1985-1995. Quite a few useful articles, which will make nice additions to my bibliography.

Updated my BibTeX file with the bibliography entries from the back of my book. Now I have about 23 references, which isn't bad for an initial literature review.

Briefly rewrote and clarified a section of my research proposal which wasn't completely accurate in the light of further literature review. I plan tomorrow to go back through my proposal and find places to cite all of my new references.

Posted quite a bit to the forums on the Myth Codex ... somewhat trivial things, but amusing nevertheless.

My one to two page research proposal is up to 6 pages now. Oh well.

Emailed Richard O'Keefe about the research proposal, hopefully I will get some feedback on my final draft sometime soon.

Did some SERIOUS thinking about my PhD, data structures, algorithms. I think I've got most of Phase 1 mapped out in my head... fingers crossed.

Moved the Shakedown website back to www.wwwebservers.com. Still no word from Spry as to what caused the network outage.

Apparently www.wwwebservers.com went down briefly due to a minor hardware fault, and will stay up for good now.

Moved the Shakedown to Round 3; everything should be set shortly.

Shakedown Round 3 Bracket 6 is shaping up nicely, we should play this weekend some time. Hopefully. If people can get their acts together...

Wrote a nice wee script to automatically get my PhD, compress it, and back it up to a remote machine under the appropriate name, e.g. "phd.220698.tar.gz"

Coded up a good fraction of the classes that will represent my data structures. This has helped me see more clearly all the problems that I'm going to have in actually implementing this... oh well, that's why it's research.

I'm still not sure how I'm going to integrate the new system into Mirage. I think it may end up being standalone for a while yet.

Bungie have asked me to help be an administrator on bungie.net, so of course I said yes. Should be fun helping to keep the roving packs of foul-mouthed teenagers in line :)

Shakedown Bracket 4 played, some really cool films. (I give points in one game for artistic impression due to blowing up own troops).

Coding up various neat algorithms for tracking object between the CSGgraph and the Primtree, at least in my head. Now I just have to write them down.

People challenging my impartiality as a bungie.net Admin because I was supposedly selected because I knew some of the other admins. Well, duh... I don't really know what to say to these people.

I played in the Myth tournament organised by Bungie, the Platform Wars. I managed to win my first round bracket against some serious competition (woohoo!).

Started coding routines up for the CSGgraph and Primtree structures, in addition to the skeleton that I've already put up.

A significant amount of coding done today, lots of sketching of algorithms. I should have a working CSG / primtree system up and running by the middle of next week.

One of my friends (Peter Isotalo) got kicked out of the bungie.net admin circle... basically, he lost his temper and started abusing several Bungie employees in a public forum. Not a good idea.

Coding up algorithms for CSGgraph to Primtree conversion.

Major changes in the Primtree graph structure. Glad I made them, though, because it's now or never.

Rewrote a whole swag of primtree.cc in order to cope with the new Primtree data structures.

Some testing problems due to incremental templated constructors. I managed to sort them out with Mathew's help, and now I just have to address the real problems. Sigh..

Tried to fill in the PhD registration form. I require our Head of Department's signature so I'm going to go see him this afternoon.

Aaargh, the virtual methods in my CSG tree are making my program crash for no apparent reason. It's probably a C++ thing, grr.

The PhD registration form is submitted and done. Woohoo!

Organising what I'm going to do for my birthday; I think I'm going to go out for dinner on Sunday evening, probably to Cena.

Going to go and se Godzilla tonight... graaaargh!

Nooo! Found the bug, and it's an incredibly stupid mistake at a very fundamental level in my memory-handling code. Sigh, now I guess I should fix it...

Went to the Albert Arms for lunch today, very nice.

Finally, the initialisation / memory management code that I wrote a week and a half ago is tested and done. Now to check all the CSGgraph manipulation code is working.

After a few typo fixed, the CSGgraph topology manipulation code is all working, it seems. I've tested all of the clauses in the code, and it checks out fine. Now to write the matrix manipulation traversal routines.

Matrix routines written, I'll test them tomorrow.

Currently investigating getting Richard O'Keefe a usercode on the Graphics machines.

Brackets 1, 2, 5 and 9 are playing... no idea when they will happen, though.

Went shopping, bought food.

Network outage across a large section of the central United States due to tornadoes, apparently. Bungie.net is down, as are smart.net and theresistance.net. Hard luck Bungie, as they were trying to play a tournament on there... whoops.

Fixed several small system incompatibilities and tried to recompile BRL-CAD. It all compiles successfully, but there isn't enough disk space to have the whole thing installed at once! Aaargh!

Browbeat people until they moved stuff off Small. Now there's enough disk space to do a full compile, so I'll leave it going and go home.

MGED documentation is compiled (without pictures). I'm reading it now and playing with the system (it's 250 pages long!).

Chris Handley has come back from CGI, Computer Graphics International '98, with the Conference Proceedings. I'm going to grab it and peruse at my leisure...

A little more Shakedown stuff. Round 4 is coming together nicely.

I'm going to embark on another Myth Tournament when the Shakedown gets closer to completion - the Myth World Cup. It's going to have the same bracket structure as the FIFA World Cup, and hopefully be just as much fun :)

All the scores for the original Shakedown brackets are in now. Just the absentees' bracket to come.

Shakedown Round 3 is complete, Round 4 has begun (and has ten days to complete).

More generalised thinking about algorithms and data structures for the solid octree propagation upon manipulation.

Had a look at the ITS's network recommendations for the graphics lab. Implemented some measures to reduce network bandwidth (running Netscape with a cache directory located on a remote NFS filesystem is bad.)

My birthday dinner was very nice indeed; Cena is certainly an excellent place.

Begun coding the octree code.

Preliminary octree structure is in place. I'm just thinking about the actual CSG combination / primitive octree creation business.

Planning for our Quake game this weekend. Our team's looking reasonably solid.

Partially completed the SolidOctree recursive updating code. Stubs are implemented for all the SolidOctree code that I haven't written yet. It all compiles fine, but dies if I actually run it (well, duh...). Lunchtime!

Debugging SolidOctree, devising methods of visualising the 3D octree structure.

Wrote a dump-octree-to-text routine and it works! Woohoo.

The update-transform routine appears to be working. I have to check the CHANGED flag is being propagated correctly, which I can do once I get the incremental octree update going.

Designing images for the Myth World Cup.

More Shakedown scheduling, some people are irate that they've only got a 56k modem guy hosting their bracket. Plebeians.

Hmm. Surprise, surprise, it doesn't work. I'll look at it more this evening, but it should be fixable.

Round 4 Bracket 1 is complete! Time to review the films.

Aaargh, those klutzes in Bracket 1 played the wrong maps for the Shakedown. Sigh.

HTML prototyping for the Myth World Cup... I need to write a few applets, which should be fun.

Played a lot of Myth.

Played Quake with AWOL; we won, and beat DMC (the Dark Ministry Clan) into the ground. Good to see - our first win of the Bloodrain season.

Wrote a modular database interface in C++ for the Myth World Cup.

Played a lot of Myth. Shakedown Bracket 1 completed; cool.

• Friday: get up at 7am, go in to work.
• Friday night: Hmm, I have to be up at 5am tomorrow morning for a game of Quake. I should go home now. Oh, Julian is going to go see Mortal Kombat 2 at the movies? Stuff it, I'll stay up all night.
• Saturday: play Quake all morning. Go home at midday for a nap.
• Saturday night: get up at 5pm. Do Shakedown stuff all day and early Sunday, and early Monday morning.
• Monday: go to sleep at 6am. Wake up at 6pm.

Some extremely funny people have formed the Force Clan for Quake. They have a section on their web site called "I'm A Dork And This Is My Quake Clan". Very, very funny indeed. If you want a laugh, go check it out.

Allan still hasn't come over and moved the machines like he promised he'd do. Grr.

Shakedown Round 5 brackets are complete.

Finished the automatic news-item page generator and linked it up to the automatic FTP updater.

*child_changed++;

*(child_changed++);

(*child_changed)++;

Removed all the debugging messages from my code.

Arranged brackets, hosts and times for the Shakedown Round 5.

Gave a talk to the graphics lab about my PhD progress so far.

Scheduling the Shakedown semi-finals, it's a hassle. At least I've found two more T1-class hosts.

Added a counter to the Myth World Cup page.

More MWC graphics updating. Fun fun fun!

Bungie.net admin things, oh joy!

Wrote a neato Java applet that ticks up units killed on the MWC main page. I really hate some of Java's features, but at least it all works now! I'm not planning on touching that for a looong time.

Tried to play the Shakedown bracket 2, but we had 2 no-shows and had to call it off. Going to reschedule.

Took the opportunity of chaos and change to do a bit of tidyup. Picked up dirty cutlery from behind tables, threw out about a ream and a half of scrap paper, rationalised textbooks and manuals etc.

While doing the tidyup, found an important document that I should get Brian to sign and give to Registry. I'll do that this afternoon, as Brian is in a staff meeting right now.

Gave Brian the document in question. Woohoo, my scholarship money should start being paid now!

Didn't do any photocopying as it turns out. I read a few very useful papers on implicit surface polygonization, which gave me some ideas of my own that I'm going to rough out this afternoon.

I had some interesting ideas, but I think the first thing to do is to get a rough system going first. That's what I'm going to work on tomorrow.

Myth World Cup stuff; more images come in, the site takes shape. Should be ready to open for registration in a day or two.

Spoke to Bungie, and we worked out that the Myth tournament market is a bit flooded right now; once their Platform Wars finishes, I'll start the MWC. This work out to an approximate start date of August 10th, with registration opening on August 1st.

Finally got into work at 7pm, after tidying up non-work-related things in the late afternoon, and having dinner. Going to settle down and code up the polygonization routines (at least in sketch format) this evening.

Never managed to get settled down to work. I'm just going to go home. I managed to get a bit done on the auto-FTP program for the Myth World Cup.

Planning, thinking about the terminal structure in the SolidOctree.

Helped Leeann move the new chairs from Archway over to the Graphics Lab. They're nice... mmm, comfort.

Wrote some terminal-inout code this morning. It looks like I'm going to have to add new Primitive routines to Mirage that have the right interface to mesh well with my new project.

Shakedown semi-final A completed, with a few hitches (namely, host not showing up). Now I just have to wait for semi-final B, which takes place this weekend.

Well, I've added 3200 lines of code to Mirage today (most of it cut-and paste). I hope there are no bugs (hmm...). Testing tomorrow.

People flaming the bungie.net admins on a Myth message board, so we descended upon them and smote them with our righteous fury (heh heh).

Some very nice new articles on www.tomshardware.com ... I especially like S3's response to his post (heh).

Went out for dinner to the Little India. Very nice, and very hot :)

Playing Myth, Myth is gooooood.

Made some screenshots of a fake 'soccer' game for the Myth World Cup website and its first feature. It's pretty classic :)

Shakedown semi-finals are complete now - I'll schedule the final tomorrow.

Made a few changes to the Myth World Cup site, can't FTP them because the external network is gone.

The network performance is fixed! Now I can work!

Oops, it seems it was one of the Graphics machines that was storming the campus net.

Debugged and tested the cellinfo creation. Woohoo, it works!

Working on the PolygonObject data structure, a bizarre segfault though. :(

The reference tracker is done and appears to be working perfectly.

Writing a topology generator on a cell-by-cell basis for the polygonization, a la Lorenson's paper in SIGGRAPH 87.

The polygonizer is mostly finished now; I just have to tidy up a few things tomorrow and then I can work on putting the terminal-node polygon subsets into a single objectset that can be previewed (say, in OpenGL).

The polygonizer is linked together now and setting up coherent polygon lists. Now I just need to link it to an OpenGL / Togl application with buttons that you can push.

Dammit, the Mesa OpenGL distribution isn't set up consistently across machines here. They're all different versions, aaargh! I will probably just design a system-specific application that runs on Small.

Scheduled the Shakedown final for this Sunday at 8pm CDT (around Monday lunchtime NZST).

Myth World Cup progresses well. I learnt how to write Perl scripts and created some rudimentary cgi-bin programs to handle the registration forms.

Writing an automatic-emailer program for Myth World Cup publicity purposes.

Myth World Cup is almost ready! The Feature absolutely rocks and Ben managed to fix some bugs with my HTML code as well!

Ouch, one of my Myth Admin friends had his computer broken into while he was away on work and his Admin account was stolen + used to maliciously impersonate him. Not good.

Myth v1.3 is out. Party time!

Myth World Cup opens for registration. The hit counter has kind of hit the roof.

Woohoo, AWOL won the relegation game versus NbN. That means we don't drop down to Pool B in the Bloodrain CTF League!

Myth World Cup is up to 22 orders so far.

Tried to play the Shakedown grand final today. God, what an abortion that was! One guy's ISP died as we were setting up the first game, and another's died after 6 games. We're going to try and reschedule, probably.

Case of the Myth Codex has offered me a position helping to edit their tournament news. Should be interesting, but I'm going to hang fire until the 13th of August at least, as I have to give a seminar on my research to the department, and that will require quite a bit of preparation.

Myth World Cup stuff... fiddling with the automatic HTML generator so that it only uploads the pages that have changed.

Profiling my polygonizer to see what areas I can speed up.

It seems a good 30% of the time is spent creating the cellinfo structure for the polygonizer to do its work on. I'm looking at this routine and seeing if I can speed it up at all.

Having a good long play with the polygonizer, trying to cut down the number of point_intersect calls. Should be possible to reuse all the other cells' inout values.

Well, I cut the point_intersect calls by a factor of four. Woohoo!

My 'improvements' have made my program continually crash. Hmm, I will work on this more tomorrow.

More MWC stuff, 41 orders now.

Fixed the ray-intersect problem, at least partially. Started working on another problem, which is that profile edges are remarkably jagged. Have to sit down and think about whether this is a limitation of the method, or a bug.

AWOL are preparing for our game this Saturday, vs DMC (the only Pool A clan we've beaten). We're playing on one of the wombles' favourite levels, which should be good.

MWC up to 44 orders.

Wrote a pool-page generator for the Myth World Cup pages. Now I just need the images for it, and I also need to write a generator that outputs the page for each pool, and a generator that creates match pages, and I have to make links from the main news page to the most recent match pages, and I need to make links from each team's page to its match pages. I also have to put together a 'match reports' page. All by the 10th of August.

Went to see Steven and Mathew's PhD presentations today; they gave half-hour seminars to the department. I have to do the same thing a week from now, so this kind of spurred me into action. I'm at a stopping point for my research right now, so it's time to stop active research and move into seminar production mode.

Wrote the generator for individual pool pages for the Myth World Cup; code reorganization and tidyup. I really need to hustle if I want to have this all finished in time for the start of hostilities on the 10th of August.

More MWC orders... 61 now.

Argh, there's a feature film being made based on the Wing Commander series of computer games. Sob.

The Corporation has disbanded.

Must sleep well, clan match at 6am tomorrow morning.

We got pounded into the ground. So damn aggravating too; I do my job, but it seems like the rest of the team just falls to pieces around me.

Did some MWC stuff, it's all going quite well.

Wrote most of the C++ stuff for the Myth World Cup. Nearly ready to launch the tournament, probably on the 12th or 13th.

Mirage stuff... fixing the configure script and the final v1.0.6.1 update.

More storyboarding of my presentation, the text is somewhat complete now.

I found a place that will sell me a PII-400 with 128 MB of RAM, a Diamond Viper V550, a Riva TNT-based 3D accelerator, a 7 GB hard drive, a 2x DVD-ROM drive for approximately $3300 inc GST. Excellent! Going to wait a while, though, certainly at least until Intel's next price drop on the 26th of August.

MWC now at 75 orders. I hope to break 80 before registration closes.

Juliet seems to be mostly fixed. We defragmented all the disks, reinstalled Office 98, and fixed the problems with the SCSI disks.

Spent most of the rest of the day preparing my presentation's slides. Just the text so far.

It's now midnight, and I just finished preparing several images and writing a complex diagram on a portable whiteboard. Tomorrow, I have to prepare two more images (the toughest...), write the words that I'm going to say, and practice my seminar! Oh, and hire a projector from the ITS, and arrange for the Department's Powerbook.

Doing lots and lots of Myth World Cup stuff!

MWC starts today. Qualifying Round has two weeks to complete.

On top of the computers being reworked, the network is absolutely un-usable! My work has basically ground to a standstill, and I don't know if I can finish my seminar before tomorrow morning.

Finished generating pictures for my seminar. Had to fix a few bugs in the system to do so... whoops. :)

Finished preparing my seminar... it's now 2am. I'm going home to sleep now, I will get the chance to practice it tomorrow morning, several times.

Lots of MWC scheduling stuff.

Seminar went very well. Some moderately tricky questions, Kevin seemed pretty impressed with my presentation so I'm quite happy.

Results are arriving for the Myth World Cup... 2 matches so far.

The equipment talk was pretty much a non-issue. It seems that a consensus isn't going to happen; Kevin is strongly blocking any purchase of NT workstations, I think he's worried that I only want the machine to play games on. I really hope I'm not forced to use Linux for my PhD.

Went to Ray's lecture and got bamboozled with inorganic chemistry... actually, it wasn't too bad. I wasn't sufficiently sure of myself to ask questions during the seminar, however :)

If Peter is reading this to you, Ray, it really was nice to catch up with you again. Didn't get to show you my flat, but that's not really important anyway.

Myth World Cup is ticking along well. More matches!

I've sorted out what directions my research is going to take in the next couple of months. Also did some planning as to the new ray-casting algorithm for the pointinfo determination in modified terminal nodes.

5 more MWC matches processed. The qualifying round is now 10% done!

It's now 6pm and I haven't had a chance to do any work because Geoff promised he'd meet with me this afternoon, and has been stalling me repeatedly. He just told me that he can't meet me at all today, it will have to be tomorrow.

MWC is going quite well on the whole. Many problems with out-of-sync errors in Myth.

I've decided not to play with AWOL any more; I don't have the time at the moment.

Partially implemented the new ray-intersect technique for pointinfo calculation. No real problems, and the code is probably about half-done now.

Had a good long talk with Geoff about directions for my PhD, which resulted in some dramatic hatcheting on my research proposal. Chopped the length in two and it's now much more readable.

Julian found an interesting web page...

Processed + updated large numbers of MWC games.

Finished implementing the ray-intersect routine, it has sped up the CSG to polygon conversion significantly. Next stage of my PhD: integration of triangle viewing + triangle output.

Went to McDonalds for Steve's girlfriend Tracy's birthday. We told her to get dressed up for dinner, didn't tell her where we were going... heh, heh. Also amusing: the look on the girl's face at the counter when I ordered $70 worth of food at once.

Finished going through the emails and tagging+sorting all the MWC ones. Now it's back through for a second pass.

Processed many, many MWC films. Didn't see Geoff at all today (even though I was looking very hard), so no work either. :(

Massive amounts of trouble with libC versions on Small. I'm getting thoroughly sick of GCC and OpenGL and Tcl and Tk problems and version conflicts.

Managed to hack around the problems with libc in order to get my OpenGL viewer up and running.

Debugged an extensive portion of the CSG triangle generator. Still some problems in the CSGTerminal creation, it looks like.

Fixed all the CSGTerminal bugs. I just have to try and deal with aliasing artifacts now.

No dice. Geoff wants us to accomodate our timetable, because he customarily goes to afternoon tea at that time. Grr.

MWC Qualifying Round is finished, all films processed, teams are calculated. Now I'm going to go home and sleep, as it is 12:55am. :)

Played a bit of Myth.

Set the Myth World Cup first round in motion! Excellent! :)

Thinking about normal calculation, adaptive subdivision and memory requirements.

Keyboard on my main work machine died, so it was out of action for most of the afternoon. Drat. It's back in action now, though.

Started looking at a normal-storage representation scheme. It's going to require a major rewrite of most of the cellinfo stuff, though, so I'm reluctant to do it. :/

Myth World Cup stuff goes well. I'm thinking about modifying a few existing third-party maps for the next MWC round.

Argh. Interruption time; I've found a major problem with Mirage's internals. Oh, bother.

The Mirage problem is deeper-rooted than I thought; it's a big one. Erk. I think I'll put it aside for now, and go back to working on my PhD for the meantime.

Note to self: abovecluster_transform stuff done for prims, intlist. raytrace.c no idea re: hitpoint translation. investigate light.c, texture prototyping. versioning with old texture code?

Finished the normal-sampling code and hooked it into the display subsystem. Gouraud shading improves the visual effect significantly. Next up: tinkering with the UI for my demonstration to Richard on Thursday. Also have to come up with a way of fixing the CSG-cell errors.

The view manipulation interface is finished now; 6 DOF controller is pretty much complete.

Argh. Richard said he can't make it; he forgot and double-booked himself. I guess I don't have any feedback from my supervisors this week, either. Sigh.

Writing an automatic dummy-finding program.

Found a small problem with my edge-encoding scheme that was causing division by zero in some cases.

Added some features for tracking the odd CSGpoint-polygon errors which I'm getting. Hopefully this will let me track them down.

Went to the library and started boning up on adaptive polygonization techniques that I can use for my CSG modeller. However, I have quite a few constraints that aren't satisfied by many methods: they should be separated into distinct data-gathering/subdivision and polygonization phases; they should be fast-ish; it should be possible to combine their data structures via CSG easily; they should be able to deal gracefully with high-frequency data.

Changed my current polygonization from center-OUT disambiguation to center-IN, which is a better tradeoff for the kind of objects that I'm likely to be polygonizing (connected ones). This produces a much nicer visual effect. However, in the process I seem to have introduced an intermittent minor bug, which I'm now trying to hunt down.

Fixed said bug.

Finally got my research proposal back from Geoff, and handed it to Anthony! It's good to get that out of the way... now 3 months into my PhD...

Zorn, the warren, zorn. The men came and clogged the entrance holes, then they gassed the warren. I barely escaped; Buckthorn fought valiantly, but he disappeared. I could find neither Hazel, Hawkbit, nor Strawberry. I saw the carcass of Bigwig though, it was awful. I fled as fast as I could, Campion was right behind me, but he must have fallen pray to the men and their guns, for when I looked back he was gone as well. I ran until I fainted, and now I am here to tell you the terrible news. Frith-rah save us!

(They lost to Aleph Naught, quite badly.)

Scoring and analyzing MWC games. Lots of them.

More MWC reviewing, pretty much done now. Trying to decide the games for the next round.

Paper reading, reference chasing.

Learnt that Geoff is on sabbatical next year; he says that he plans to be in NZ for most of the year, doing his own research and supervising students full-time. Hurrah! I guess that makes Year One of my PhD the 'Chris Year', Year Two the 'Geoff Year' and Year Three the 'Richard Year'.

I'm going to try and implement Laidlaw and Trumbore's paper on a cell-by-cell basis for the polygonizer. This will produce exact edges, but it might take a LONG time.

After a lot of thinking, I know what direction I want my program's polygonizer to take. Let the coding frenzy begin!

Surprise, surprise, it's not working first time. I'm trying to track down the problems now.

Nasty, evil problems will have to be fixed tomorrow.

The rewritten polygonization code is totally fixed now. Spent a while looking at how I'm going to implement the Laidlaw and Trumbore algorithm.

After, of course, I get the code I've just written up and running. That will probably take most of Monday morning.

Well, we played the match. What an abomination. Repercussions that I had to deal with for all of Monday, ended up wasting a lot of my time. :(

Debugging code that I wrote on Friday. Problems with face/face intersections, it appears.

Maybe I spoke too soon. The intervals seem to be mysteriously disjoint even though it's geometrically impossible for them to be so. More investigation is required.

Urk. I am a fool. Failure to understand the problem doesn't lend itself to the creation of a good solution, Chris!

The OEM version of Half-Life has been released and from all accounts it is a superb piece of work. Looks like I might be forced to buy it at some stage.

Heh, a very amusing link: Mr. Slander's thoughts on Quake.

Found and fixed one problem with error tolerances.

The polygonization code is pretty much complete; I've tested it with a slightly more complex object (some tori intersecting) and there don't seem to be any new problems. Next step is getting Gouraud-shading working again.

Chatted to Geoff about my method, he seems to think it's reasonable, though of course he thought of it about a decade ago but never got around to researching it.

Went to the library and got some textbooks on long-term loan.

Nice review of Shogo at PlanetShogo.

Oooohhh... it's pretty. It all seems to be working, bar a few small bugs. WOOHOO!

This marks the movement of my PhD into its next phase, that of UI design. The task is to implement a user interface which makes manipulation of CSG objects not only possible, but easy to interact with.

I grabbed a swag of bookmarks from Steve Smithies, who's doing an M.Sc in user-interface design. I'm currently going through them looking for stuff; spent a very enjoyable few hours reading the Interface Hall of Shame.

The Myth World Cup is going very well, teams are sorting out their scheduling and it all looks like it will fall into place this coming weekend.

HCI browsing on the web; found some good online bibliographies, and the proceedings of SIGCHI '95, '96 and '97.

Said book is on close reserve, so I was only allowed to access it for two hours. But in that time, I managed to read quite a fair chunk of it, and assimilate it for lots of nutritious HCI facts.

I did some massive searches on www.hcibib.org and spent a few hours parsing the torrents of disgorged papers, looking for relevant publications. I found quite a few, but I have no idea how I'm going to get these papers; none are carried by our library!

Booked myself in for a training session on how to use the "Web of Science" system at the Science Library; on the 5th of October.

More web-based research; I'm currently picking the brains at Microsoft Research.

Lots of Deep Thought(tm) about a user interface to the CSG modeller. It's going to be quite a task.

I think I have most of the UI sorted out now. I've also been trying to think coherently about how to design transparent objects for primitive selection into the CSG-polygon structure, and I think I've got it.

Went to a seminar about Bob and Geoff's pet paper COSC 326; interesting.

Noooo - my program just dies now after the upgrade. No apparent reason.

Fixed the problem. Did I mention that Togl is a nasty, evil, dirty, ugly hack?

Solved a small bug with edgenormal calculation on vertex-vertex splits while doing polygon CSG. The gouraud-shading is all working very nicely now. I also implemented a wireframe mode for debugging purposes.

Busy debugging Bungie's Myth application for them, trying to sort out what causes it to go out-of-sync in network games.

After a great deal of coding, I believe that the selection structures are all complete. Time to debug it all.

Lots of debugging fun. It's all going rather well, on the whole, although I do wish that I had a decent debugger to use.

Well, that was simple. Just a few minor corrections, and he can enter play.

w00p! It's all working! Beautifully! It does pretty transparency things, with only a few shading artefacts caused by the incorrect draw cycle! WOOHOO!

Oh man, there's a lot of code to be written. Mostly just stuff that I can grab from Mirage, but it all requires a bit of data-format conversion, yuck.

The crash bug was remarkably easy to fix; however, I have a somewhat more insidious problem that requires solving now.

Solved the problem with pieces of ghosts not appearing. The RCSG ray-casting code appears to be working perfectly, with no debugging necessary; excellent.

Moved on to recursive up-selection upon multiple-click now. There seems to be a problem with primcsgnode selection through wrappers, so I'm going to work on that now.

Fixed; damn, I feel stupid.

Spent two hours at the library, looking through the library's collection on Computer-Human Interface literature. It's pretty pitiful. I also got a pile of interloaned articles back this morning, which I assimilated and filed. Still, I have one useful article (!!) ... unfortunately, it doesn't go into any detail as to exactly how their technique works. I might try to track down the author and get some hard information out of him.

Julian just finished building Enlightenment, which is a cool window manager. I'm going to take the afternoon off and play with it.

Well, Enlightenment is very pretty, but it doesn't have a whole bunch of features (menus, a pager) that I want to use. I'm going to stick with fvwm.

I tracked down Dan Venolia successfully by email, and he told me that most of his work has been lost in multiple moves since he published that paper. He gave me a reference to US patent number 5,463,722 and I sent off an Interloan request for it (after spending about an hour finding out the correct process).

It's time to rewrite most of my viewing code, as I know how the modeller is going to fit together now. It's become rather intertwined with the Tcl interface and it will be good to get them nicely separated out again.

Lots and lots of coding, rearranging, function mangling etc. The structure of the modeller is a lot nicer now, though. I also learnt about display lists in OpenGL at one point when I got sidetracked.

Well, all the coding is done, the modeller compiles quite happily... and dies as soon as I run it. Bugger. The problem defies an immediate solution (i.e. I couldn't debug it in under 30-seconds), so I'm going home.

Hooray, the display lists work. w00p! That was pretty easy, really. OpenGL can be very nice sometimes.

Fired off an Interloan request for another constraint-satisfaction object-matching paper.

Added a small ghost / no ghost check to speed up display. I wish I had some way of determining the framerate increase that this gives me.

Updated the Altus Praeses web page a little.

Mike Gleicher replied to my email, and I'm now downloading his PhD thesis from an FTP site in the US. Oh happy day!

Hmph. Most of his thesis is corrupted; he says that he's working on a new web mirror and will let me know once that's done. I guess it's time to work on something else..

Well, I basically got no work done today. How depressing. I did help Geoff and Kevin with file format conversion for the assignment that they're presently marking, but that's not really good enough.

Argh. I reviewed the Dwarves' films of their semi-final and they pulled some rather lame tricks that exploit a loophole in the rules. I'm going to close the loophole and get them to replay the games in question.

Shut down my 'Cash Draw Saver' with the BNZ, it was giving me a horrific 1.5% interest. I'm going to start up an 'Accelerator' account with ASB.

Done. I am now a happy little number in the machine that is ASB. I am also considering applying for a Visa card through them, as it could be useful for making purchases over the Internet.

I've borrowed the bottom drawer of Steven Mills' filing cabinet, and I'm busy organising all of my photocopied papers.

I got the CD that James sent me, and will burn Half-Life onto it once I get the current CD back from the guy that has it.

I am getting so sick of people that exploit the letter of my tournament rules in order to try and gain an advantage, then turn around and act wounded when I slap them down. This current situation is extremely aggravating.

I received my thesis back from the external examiner, and all the paperwork has been completed. I now have a piece of paper saying "You have been awarded a grade of A for your thesis and are eligible for the award of the degree of Master of Science with Distinction in Computer Science." ... woohoo! I will probably graduate in absentia in May 1999.

Oh god, I wasted most of the rest of the afternoon sorting out this asinine MWC problem. Sometimes I wish people would spend just a little more effort to look at the other party's point of view.

And lo, I didn't get anything done at all. We went to Ombrello's for lunch, which was very nice (and rather dear!). Pumpkin ravioli, mmm.

Went to see Saving Private Ryan at the movies. It's a long (3-hour) film, and one that I would recommend to virtually anyone with the maturity to actually comprehend what war is.

Organising a LAN gaming session for this Monday, in the graphics lab. Should be quite a lot of fun!

It looks as though my PR antics have succeeded rather well, again. I wish I didn't have to keep doing this, though.

MWC third-place playoff match worked rather well, Civil defeating Foundation to win narrowly.

Watched Goldeneye on video... man, James Bond movies are bad. Never again.

Hi ho, hi ho, it's off to photocopy I go.

Photocopied a whole stack of papers for my records. I'm now going through Dan Venolia's US Patent number 5,463,722 and boy, is it dense going.

Back to Rappoport's seminal 1997 paper on CDA modelling for interactive CSG manipulation. It's a good paper, but it has no references to geometric constraints; blast.

More Myth investigating. It looks like silent-OOS may be due to Mac/PC version conflicts.

Reminded Bungie about my Myth II beta-testing application (it's been four weeks, and not a peep).

Sifted through a whole pile of Myth metaserver logs looking for people padding their accounts.

Found the homepage of Ioannis Fudos, a guy who works with graph-constructive constraints in 3D. I emailed him about getting a copy of his PhD thesis, and we'll see how that goes.

Reading through Ioannis Fudos' MSc thesis, it's quite interesting but rather long.

Fascinating MSc thesis, basically developed many of the things that I'd like to look at implementing, but in a rather narrow, restrictive 2D case. I *must* get a look at his PhD thesis!

Found a cool thing called 3DMark99. It doesn't work on Mathew's Voodoo-1-based computer though. :(

Oy, it's a large thesis. Sigh.

Finished reading Chapters 1-7. Fascinating stuff, but another 3 chapters to go tomorrow.

The System Shock 2 website looks cool.

Going through Gleicher's references, looking for Nice Things.

Tweaked an expect script for easy access to the Library's computer.

Off to the library, to read UIST '90, '91 and '92...

Dammit. I forgot that I had already read (and dismissed most of) those publications. Oh well. At least I was able to read the University's (meagre) stocks of the Computer Aided Design journal.

Sent my big list of Bad Guys off to Bungie for them to lock.

Re-evaluated all the directions of literature-review that I have available to me. I've now got a short list of things to look at.

Woohoo, I found the Web of Science online citation searching tool. Joy! I can play here for the best part of a week... I've already found about 15 articles to interloan and a whole swag to look up locally.

Interloans ahoy! Off they go, and I'll see if I get anything back that makes this whole exercise worthwhile... I also have a few nice lit-reviews from circa 1994 that I will have to go through with a fine-tooth comb (270-odd references, oh joy).

Must go supervise the MWC Grand Final. Shame that our Internet connection is saturated, again.

Woo! A superlative Grand Final! The teams in question rock, totally and completely. I am really impressed. Results to be posted later on tonight.

MWC review of the final and the trophy page are almost done. I'm just going to get comments from the participants, then post it tomorrow morning. WOOHOO! IT'S DONE!

Reading the Electronic Primer on Constraint Solving on Purdue University's web page.

Ye gods, getting information out of Microsoft's website is like getting blood out of a stone.

Our Internet connection basically died, so I can't continue with any of my web-based litreviewing. Argh. I spent a while sorting out specs for the new NT boxes, but I'm about to give up and go do something else.

Civil Order asked me if I wanted to join their merry little club, but I think that at the moment I shall respectfully decline. I'm having a lot of fun with Altus Praeses at the moment.

Investigated 3D acceleration under NT, talked to people at ARL and posted a query to USENET.

I got diverted into examining security issues under UNIX while waiting for Geoff to come and talk to Mathew and I about system specs.

Resolved the final issues regarding the configuration of the new research machines. Geoff will talk to Brian tomorrow and they will hopefully be ordered by the end of the week.

Got a whole bunch of Interloan requests back today stamped "Not Available In NZ" ... blast.

Sent off international requests for the documents that aren't available in NZ.

Sent Bungie the details of the people who won the Myth World Cup, shipping info, that sort of thing.

Several of the papers are quite interesting, even useful. Particularly Sistare's paper, which introduces the idea of object creation with respect to some 'locus' that determines both creation positions and implicit constraints upon creation.

Obtained a mystical code from the office, which allows me to interloan requests using the Department's fund to cover any expenses (from overseas document requests).

Processed and filed the interloaned papers. I can feel an idea nudging around the back of my head, I'm going to try and write it out properly this afternoon. Right now, it's lunchtime.

I was able to write out a little of the idea in the afternoon, but then the graphics meeting got in the way.

Lunch at Poppa's Pizza. Mmm, good.

Altus Praeses may be falling apart. Hmm.

Emailed Ioannis Fudos again, I hope that I get a response this time.

More interloan stuff. I interloaned a paper entitled "Solving Geometric Constraint Systems" by Glenn A. Kramer, and they sent me the book of the same title. Ah well, I don't mind. I'll take anything about the topic. :)

Found an online abstract-and-text source for the Computer-Aided Design periodical that is very useful.

Interloaned a few publications from this source.

Fallout 2 was released just recently, and there are a lot of people talking about it on comp.sys.ibm.pc.games.rpg ... there's been a lot of talk about its prequel (Fallout) and how good it was. It struck me that I might be able to find a second-hand copy somewhere, which would be great. So, the call is out!

Finished reading the Kramer text that I received yesterday. A very interesting book, but its work on mechanical linkages isn't as directly related to my research as I'd hoped. Has given me some good ideas (I hope).

I've managed to clear almost everything off the Onion, so it's time to go get the MacOS CD off Tracey. Fun fun fun. I can also return my interloan book.

Loooks like there might be some 'issues' with MacOS 8.5 trashing hard drives. I think I'll upgrade to MacOS 8.1 instead, as it provides HFS+ support anyway.

Ugh. The upgrade procedure found errors on the hard drive, so I'm going to have to run Norton Disk Doctor on it before upgrading. :\

Upgrade complete, HFS+ online. Cluster size reduced to within acceptable tolerances. End transmission.

Woo, found someone who wants to sell Fallout for $12 US (that's including a strategy guide). Huzzah!

Thinking deeply about the nature of the matching constraints that I want to implement. I think I've finally grasped an idea that will *work*, and work nicely.

Writing out a detailed plan of steps that I need to implement in order to build this system.

Ah, it feels good to have a plan again. Starting tomorrow, I can get back into hands-on coding, at least for a while. I wonder if I've forgotten anything?

Spent most of the morning thinking about exactly what needs coding and the various inter-dependencies between modules.

Implemented TransformList and hooked it back up to the primitive manipulation slider.

Attended an IRC chat about Fallout 2, quite interesting. It was raining, so I didn't want to leave the lab. :)

Did some coding, implementing repeatobjectnode... not finished, lunchtime.

Although I got a bit done after lunch, the graphics meeting and talks with Geoff interrupted me before I was quite finished. Oh well.

Shipping from the States for Fallout would have cost about $45 NZ. So, I bought it over the Internet from a local place for $35, inc. shipping.

Myth. Lots and lots of Myth.

Blast. The TransformList code that I wrote on (Thursday?) is crashing.

Uh-oh, memoryexplosion spiral. Ugh, I hate printf debugging.

Fixed the memory explosion, but I seem to be accessing memory at 0x00000018 which is a system register, if I'm not mistaken. Why the hell is it doing that? Sigh.

Fixed that problem, it was a rather subtle bug, but it's gone now.

Fixed an insidious transformation-order logic bug. Repeatobjectnodes appear to be fully functional now.

Trying to track down a problem with inout pointdata and spurious CSG polygons inside MINUS volumes that result from an addition (?).

Found and fixed the problem. Was to do with the ham-handed handling of the ghost polygons.

Done! Mmm. That wasn't too hard. Now to deliver it to Anthony.

Read and processed some more papers that arrived via interloan.

Apparently my PhD registration form is only just being processed now. Lousy bureaucrats.

Deep thinking about DMDs and integration of the skitter list.

Myth. Busting lamers. No games yet. :(
Well, some games, but the people didn't show up for our scheduled tournament game. (!!!). For that, they die.

Pre-ordered Half-Life. Mmmm, yummy.

The rendering code seems to be working, it's all nice. After some grappling with code that should have worked, I rewrote it, and it works fine now. Shrug.

Met up with Caron Lamb, who I haven't seen for about 18 months. She's been on an exchange program in the US, and it's nice to catch up with old friends.

Thinking about DMDs and means for selection.

Coding up the basis of the DMD selection / manipulation architecture for the interactive editor.

The selection architecture is mostly in place, but I just realised that I have to make a rather fundamental change to it. Tomorrow, I think.

Prototype DMD construction, coding in order to allow such.

Superb! I can now interactively move objects using the DMDs, in 3D space. The interactive-editor side of things is really coming together now.

Got axis-aligned translation working too.

Translation and scaling DMDs are finished. Tomorrow: rotation!

Gamezone reports that Half-Life has been delayed in Australia and will be hitting NZ shores tomorrow.

Axis-aligned rotations are working perfectly, I'll work on free rotations after lunch.

Free rotation is working! The rotation widget is virtually complete now; the only thing missing is a set of white 'guide lines' to indicate the current direction of rotation.

Trying to play Myth, but our Internet connection died totally.

Pricing computer systems for my new research machine. I can't believe that it still hasn't been ordered. This is abysmal.

After talking to Geoff and Allan this afternoon, we finally resolved to just order the original systems from Dell, no educational pricing and damn the consequences. Maybe the order will actually go through today!

Downloaded and played the demo for Thief, it's quite fun. So many cool games, so little time!

Well, that was a bureaucratic waste of time. As if I expected anything else...

The Dell machines have finally been ordered. Woo woo. Mathew and I put together pricing for Visual C++ and a Microsoft Developer Network subscription as well.

I found a better graph-layout tool, with a proper API and all the features I want, but the FTP server that it resides on is unbearably slow. Argh!

Downloaded daVinci, only to find that its API insists on starting up a complete application and drawing all the graphics itself. NOT good. I guess I have to look for other algorithms.

Downloaded a DAG graph viewer, I also have a few references on graph drawing algorithms that I can look up tomorrow at the library.

Thinking about running a tournament when Myth II arrives.

Going through the documentation and supporting papers for TclDG and DynaDAG.

I don't think any of these dynamic DAG drawing papers are going to be useful for my PhD. The nature of the objectgraph doesn't lend itself well to any of the standard solutions.

Library trouble, missing volumes of journals. I guess I didn't really need that paper anyway.

Scheming about creating a possible Myth II tournament in association with a guy called Angelus.

Lunchtime!

Signed up for the GTK-ICQ development project.

Much hunting around on the web looking for alternate computer and video card suppliers. It's very aggravating.

Got an offer from Shan at Tales of the Myrkridia to join their staff on the basis of some layout comments I had. I declined.

Ho ho, the staff of the Tales seem to be imploding. Alas, poor Atleus.

Colluding with a friend to set up the Myth Done Quick website.

Downloaded graphviz 1.3 source code, time to compile.

Much makefile hacking, X library searching, doing the #define shuffle! But it all compiles and installs. Now I just have to see if it's worth using.

Dell stuff, trying to find an alternative video card that will work with NT.

I can't find any sites that have proof that the TNT-based cards can do 3D in a window. I know they do, but I can't find proof for Mat. Argh!

Arcane LD_LIBRARY_PATH woes with the dynamic linker mean that I can't run graphviz and netscape from the same setup. Had to write a glue script to launch Netscape instead.

Downloaded lesstif 0.87 and installed it in my PhD directory, in order to run an ancient Motif-using application called dflo. It's a graph-viewing application that does flowcharts, and it might have some principles that I can apply to my DAG layout.

I have dflo up and running. Its graph-layout algorithms appear to be somewhat sub-optimal, so that's not necessarily the best avenue to explore. Damn.

MSVC++ 6.0 arrived. Unfortunately we don't have the machines for it yet, and are unlikely to before Christmas.

Some work on changing the way that new objectnodes are created, but now my program mysteriously crashes. And when I try to run the debugger on it, that crashes the debugger. Argh.

We decided to go with the Canopus Spectra 2500 cards at least initially, despite some worries with their image quality when running at high resolutions (like 1600x1200 high). Going to order two initially, with the option of another one for Mathew later on.

Fixed the crash bug, only took about a minute. Woo.

Allan has ordered the Dell machines. Here's hoping that they get here before Christmas.

Fixed a problem with add_repeat_around and add_wrapper_around that appears to have been present in my code for a very long time. Phew! That's a relief.

Implemented the ParentTracker and removed a whole passel of (now) useless code.

Lots of thinking about how I can plug / play the GraphLayout subclasses and what to do about that.

Lunchtime!

Working on implementing HorizGraphLayout and HorizLayoutNode algorithms. My modeller has now passed 20,000 lines of code.

My computer is back up again; on a hunch, I ripped out the L2 cache card and it now no longer crashes during the boot process. Oh, joyous rapture! It is a bit slower (20% or so) now, but I still spent several hours last night playing Half-Life (from 12 till 2...).

I think the graph layout and structure engine is complete. Now I will need to hook it up to some kind of drawing structure, probably text-based.

Heh, the Diamond cards have arrived today. Now we just have to wait about a fortnight so that we have computers to put them in. :(

I've compiled a lot of work on the layout engine, now I'm testing it. Some arcane error is causing it to crash inside the compiler internals. Argh!

Fixed, it was a minor stupidity.

Mmm, barbecued lamingtons.

D'oh! I was in the middle of a conference about the rules for the tournament and our University's gateway machine died. That sucks.

The network connection is back up again, and our nefarious plans continue.

Wrote draft rules for the tournament. Working title: "Trial by Combat".

Drafted HTML layout for the web page. If Jeff's image skills are up to it, this tournament will look superb. :)

The Layout code that I wrote before the weekend appears to be solid. Time to extend it and hook it into a rudimentary graph view.

Lots of Tcl code compartmentalisation and decoupling. Fun, fun, fun.

Show-stopper bug in the Tcl error handling. I have *no* idea what's going wrong, but my program just hangs whenever I get an error, instead of going into the Tcl error handler.

I managed to get rid of the error in question, but I still have no error handling. :(

Scrollbars and canvas are now added and working. I just have to write the actual node drawing engine and we will be away laughing. Mwahaha etc.

Designing the graph layout. Made me realise that some of the graph operations need more failsafes in order to ensure that they don't create invalid graph layouts.

Decided on names for the new research machines; the lab is going to move to Shakespearean characters. I have reserved the name Iago for my new machine.

Frantically trying to get my modeller back into a working state before my PhD meeting with Richard and Geoff at 2pm. Of course, it's being recalcitrant, and it's not pleasant.

Another bloody stupid bug. Sometimes I really am a moron.

Hooked up the selection mechanism into a colouring for the layout engine.

Spent most (if not all) of the day pursuing and finally catching a rather nasty oversight in my CSGgraph manipulation code. The manipulation code isn't complete yet, and there are several more issues that need fixing, but it's getting much better.

Wrote some CGI scripts and email parsers for the TBC registration phase.

Hunted out a few more inconsistencies in the CSGgraph building routines.

I think the CSGgraph linking routines are all self-consistent now. I just have to add an interface for adding primitives... and add deletion...

Hooray, now I can delete stuff. How constructive. Only trouble is, if you delete everything in the scene, there's no way to add anything new. Hmmm.

Fixed that annoying 'no-more-toys' problem. Now the only thing that needs adding is the ability to link in primitives... and some more debugging, too.

Learnt today that the Dell machines were shipped from Auckland yesterday, and should arrive here on either Monday or Tuesday. w00!

Convinced Carch to join my cunning scheme.

More story writing for the TBC.

Website design. Lots of HTML hacking and Photoshop fun.

Had fun playing Myth Done Quick with Ruiner and a few friends.

Trinity has come back with the images for the TBC website, so I am stitching them into my HTML code.

The prototype TBC website is up on mythii.com if you want to look at it. Please don't publicise the URL, it isn't ready for public consumption quite yet.

Ah, what the heck. The TBC is open and accepting entries.

Found and fixed a problem with octree pointers not chaining up the graph correctly. This has let me get past a crash bug, and start debugging the real underlying problem that's causing anomalous selected patches on objects.

Unfortunately this has revealed that the problem exists in the TerminalTracker, code that I thought was final and bulletproof. I guess I have somehow designed a silver bullet that can kill it anyway. Argh. Oh well, lunchtime.

Diverted into NT machine administration concerns, didn't get around to debugging anything quite yet. It's a very complex issue and it's making my head hurt.

Once I actually understood the issue and made the required change, it all worked first time. Ah, that feels good. A shame that I didn't predict the situation occurring in the first place, though... it was a natural extension of my modeller and I really should have anticipated it.

Added an add-new-primitives button to my modeller. Easy! Seems to have trouble with axis-aligned surfaces, though, and I will have to fix that. Also managed to find a crash bug related to deleting objects.

Fixed the crash bug just as the clock struck five. Woohoo!

Struck an unusual problem whereby segments of the polygonization are hanging around after they should be disappearing. Time to write some in-depth debugging routines.

Woohoo, it's fixed now. Was a problem with the recursive propagation of CHANGED flags inside create_csg_octree, through multiple levels of octree wrappers.

Lint removal and code tidyup, cut the codesize down by 1,000 lines through removing commented-out debugging statements. Fixed a few nasty global hacks that are now outmoded due to better design. Finalised codebase for the port to NT (assuming the machines eventually get here).

Catching up on Myth-related stuff. Worryingly, there are a lot of people whom I respect ragging on the Myth II netmaps. Apparently they aren't very good. A bit disturbing, considering that I have committed myself to organising a tournament using them.

Downloaded demos for Myth II and a Civilization-like game called Alpha Centauri.

Decided, after talking with Kevin, that it might be a good idea to try for a paper in SIGGRAPH 99, just a short one talking about my cell-based CSG engine.

Decided to push back the TBC start date by a week, due to Myth II's late distribution cycle.

Wrote a long commentary on the Myth II demo.

Played a lot of the Myth II demo, it's a lot of fun.

Thinking about publishing my work. I chatted to Geoff and he recommended that I target a technical sketch in SIGGRAPH 99. One page, 500 or so words plus some figures. Submission date in April some time.

Played some Myth, won some, lost some.

Started a large rewrite of most of the triangle-splitting and CSG code.

Reworked and rewrote about 4000 lines of code. Time to compile and test it now.

Lots of bungie.net adminning.

Trying to select good games for the first round of the Trial by Combat, even though I don't have Myth II yet... ugh.

Played Myth. Quite a bit of Myth.

University Internet gateway went down on Sunday. Fortunately it was a beautiful sunny day! Yay!

The official high for Dunedin today was 30 degrees. Yow.

Got an OpenGL test application off Mathew and it runs fine, 3D-accelerated on my machine. The Riva TNT card is just working transparently and seamlessly, it's absolutely marvellous. I can't wait to see what it looks like when I hook up my modeller to it.

Everything is nearly perfect on my machine, apart from the fact that I don't have a nice interface to the Linux machines. I'm going to investigate X-windows extensions for Win32.

The ITS aren't going to enter the machines into the DNS until early next week, apparently. This means that I can't send email from my machine until this happens. Blast.

Discovered that my Myth II pre-order hasn't even left Bungie Software yet, and will be another two to three weeks. That's rather frustrating; see the Codex / b.net forum for more detailed comments.

My GL-CSG code is up and compiling under NT, linking versus Mirage and everything. I've eliminated all the memory leaks that are tied to Mirage's global constructors... next up: create a WGLView class and link the modeller up to it.

Hmm. Maybe not. I've hooked up the Modeller class and it is causing memory leaks _somewhere_ ... time to find out where! Oh, MS Debugger, how we love you.

It is so hot today that I just cannot concentrate. Pfaugh!

Well, maybe I can if I drink enough Mountain Dew. I think I've found and fixed all of the memory leaks in the modeller. Now I can begin the GL-view porting process.

Went and bought tickets to see the movie 'Enemy of the State' tonight.

I think I've fixed the final memory allocation problems in Mirage. At last.

Started hooking up the GL windows.

Oh god. It was all working fine, then suddenly it all stopped for no apparent reason. It's not even getting to my initialisation code before my application dies with an Access Violation of type 0xC0000005 in NV4OGLNT.DLL (the Riva TNT drivers). Argh argh argh argh.

Looks like the problem may have been that I wasn't releasing the rendering context back to the OS, and this was causing a bug to show up in the TNT OpenGL drivers. It doesn't seem to be coming back.

More complicated and hard-to-find bugs. But at least I am getting a picture in a window, even if it doesn't update.

Fixed all the crash bugs I can find. Now I just have to work out why the picture isn't updating when I click on something to select it.

Hrm. Looks like a tricky one. Time to go home, I think, and leave it for tomorrow.

Enemy of the State was a very enjoyable film, excellent.

The port is (for all intents and purposes) complete and is at the same level of functionality as the UNIX version was when I started. But now I have 3D-acceleration working.

Took a picture of Nick Meek for the departmental web page.

Hooked up all of the view-manipulation operations (rotate, shift, zoom) to the mouse events.

Finalised the MDI interface and removed all globals that were implicitly linked to Modeller. There are no memory leaks associated with my code now, as far as I can tell... creation of 6 modellers, manipulation, deletion all work with no memory leaks at all. Yay!

Changed the MDI child window so that the default application frames are 3-way split frames. This gives me a place to put the graph view and transformation view once I create them.

VC++ is being moronic in its syntax parsing and failing on a simple constructor. I have no idea why. It's very aggravating. I don't know how to solve it, so I'm just going to leave it until after the weekend, probably.

Trying to find enough hosts for the Trial by Combat, and get them organised.

The parse error in fact turned out to be that in the constructor of my class, one of the parameters was CGraphview * instead of CGraphView *. And Visual C++ died with the error "unexpected token class Mirage::WObjectGraphView(" instead. That is the least intuitive error message yet, I have to say.

Working on hooking up the GraphView to my modeller's code.

Myth II arrived today, woohoo. That takes several large, heavy weights off my mind.

Once again, someone has already got the login 'butcher'. Oh well, I guess I can live with 'butcher2' once more. :\

Trying to work out how I can get a drop-down menu in my toolbar. I'm sure the mystic code TBSTYLE_DROPDOWN has something to do with it... sigh.

Testing Myth II v1.2 with the admins.

Went to see the film 'Elizabeth' last night. It was real nice.

Continued trying to sort out and fix problems with TBC Round 1 results. Placed a results page up on the TBC website.

Fixed a nasty bug related to deleting objects from the modeller.

Had a good discussion with Geoff and Richard about targeting of my technical sketch in SIGGRAPH. I'm going to start writing tomorrow.

More TBC results tidying.

Amazing, I finally have all the Round 1 results collated. Now I just need to put them into .zip files.

Helped the fish-film people track down a bug with different versions of the motion-path code, hardcoded constant array sizes and memory corruption (I'm sure you can guess how these fit together, Puzza).

Installed NT-Emacs, a LaTeX environment and gnuserv, which lets me write nice TeX documents on my machine. Took some fiddling though.

Draft text complete, spent the afternoon drawing a figure. A very complicated figure. With lots of lines and points and things.

Finished compiling and collating TBC results for Round 2. www.mythii.com's connection is totally saturated though, so it took a while to upload them.

I may have to flag away trying to keep up with the volume of information on the Codex bungie.net forum. It's insane.

Spent most of the afternoon setting up iago as an FTP and X server so that I can work from it more easily.

Much arguing with our system programmers over whether or not we should purchase X-server software for our NT machines.

Played a lot of "The Legend of Zelda" on the Nintendo this afternoon; the games were going to go back to the video store at 2pm and I wanted to evaluate them before they did. Oh well. :)

With the return of VC++, I'm working on finalising the triangle-adjacency code that I sketched out before starting work on my technical-sketch.

Played a bit of Myth II this afternoon.

Ran out of room in my PhD workbook. D'oh! I need a break anyway, so I'm off to town to buy a new one. :)

The movie 'Ronin' is worth seeing IMHO, especially if you've ever had fun imagining a contemporary shadowrun.

Lovely afternoon -> went to read a book under a tree in the gardens.

All that freaking debugging, and the problem was that I had cut and pasted some code, and replaced 1 with 2 in every case EXCEPT ONE. For want of a character, the statement failed. For want of a statement, the assertion failed. For want of an assertion, the adjacency search failed. For want of an adjacency search, the triangle-CSG routine failed. For want of a triangle-CSG routine, the CSGTerminal polygonizer failed. For want of a CSGTerminal polygonizer, the TerminalTracker failed. For want of a TerminalTracker, the polygonization update failed. For want of a polygonization, the OpenGL display list failed. For want of a display list, the update routine failed. For want of an update routine, the entire goddamn application went belly-up. Sigh.

The adjacency graph appears to be as stable as can be expected. On to the next part, which is the creation of a transformation list window.

Testing betalands maps, had a play around with Red Sands fc01. Bellendaine is trying for a different game mechanic with his Stampede! mesh, and IMHO it's failing dismally.

The TransView is up and working, allowing me to select individual transformations, and detecting double-clicks (but not yet popping up dialog boxes). To do: adding new transformations, deleting existing transformations, and reordering existing transformations.

It is very, very hot this afternoon in the graphics lab, but more importantly it's humid. I am having a lot of trouble concentrating.

You can now add transformations to the transview window. This brings the modeller up to something approaching useful functionality, and I can now start working on optimization and preparation of the SIGGRAPH 99 sketch.

Time to sit down and prepare a position paper on the graphics lab's current direction with regards to computer systems and OSes, and possible future directions.

Excerpt from a conversation I had with the ITS helpdesk:

> Can I get mail redirected from my student account
>(butch980@student.otago.ac.nz) to my departmental email account
>(cbutcher@atlas.otago.ac.nz), please?
>

Hi Chris,

Are you a staff member Chris?

Regards
Mike

I was so tempted to reply "Would I have a student account if I was, dumbass?" but managed to stop at the last second. Remember, Chris, don't insult the menials and they won't get all surly on you.

Another very hot day this afternoon in the graphics lab. All I can do is flake out, eat ice-cream and drink bad Fanta imitation soda from the Duck. It's a hard life.

Most of the afternoon also went towards aforementioned position paper.

Hmm. It seems that one needs to turn on the ability to profile code individually for each user, under NT. Now I have to kill everything and log in as Administrator in order to try and figure out which combination of permissions I need. Sigh.

Profiling my code (as Administrator)... a lot of time being spent in re-polygonizing the primitive solids, before the CSG splitting is done.

Planning for MWC99.

The people at MS tech support don't have any incidents of this nature in their knowledge-base. The guy I spoke to didn't have the problem profiling on his machine there, but he's going to run some tests on it and see if he can reproduce it. I'm going to try a workaround or two of my own in the meantime.

By un-installing VC++ and reinstalling it as the Chris user, now Chris can profile code but nobody else can. So that problem isn't too bad now, but it's still an issue that needs resolving. I'll see what the MS Technician says when he gets in touch with me.

More fleshing out of the ideas that I had yesterday.

I seem to have been subscribed by email to the Myth Codex bungie.net forum, with the result that I had about 90 email messages this morning, all filtered off into a little side-folder. :)

Well, I received a Myth II T-shirt in the post. Now that's weird. But I'm not complaining.

Myth II v1.2 has been released.

I basically didn't get much accomplished at work today.

Web design stuff for MWC '99. Also tournament organization stuff.

It is cold, and wet, and I got rained on going to lunch. Ah-choo.

I think that I have finally managed to sort out nicely all of the changes that I want to make to my PhD program's architecture. Spent the morning thinking about it, and the afternoon writing the ideas up in detail so that I can begin implementation tomorrow. I hope I got it right first time.

Decided to do a complete backup of my entire PhD directory.

Entered Bungie's Seven Phoenix Rising tournament.

Wore in my new boots; they work really well!

Got a reply from Terry Cole re: X-Win Pro and we should be up and running sometime later today (fingers crossed).

Went to the meeting for the Paradise trip this weekend, and met Dean Hall there! Apparently he is an experienced tramper and is looking to have some fun in Dunedin as well. He also reminded me that I should ring James some time. :)

Checking results from the TBC semi-finals that were played this weekend.

Finally sorted out the X-Win Pro software that I've spent weeks trying to get installed. Sigh.

Maya licenses are apparently on their way from Alias|Wavefront right now.

Arranged for Juliet to get an OS reinstall soon, hopefully will fix some of the incessant crashing that it's had lately.

Wrote up an MWC99 site map and sent it off for perusal by web guys.

Coding, reworking old code into new shape.

Got a little bit of work done today, but the coding hasn't really worked out quite as well as I would have liked. There are still some fundamental issues that I have to sort out.

Oli came down from Wellington to visit friends, so I spent a few hours chatting with him this afternoon.

The tramping club's river-crossing tutorial has been re-scheduled for Wednesday, so I can help Case with his tournament on Sunday. Hooray.

Had some long discussions with Wolf and Karan about their work in polygonization and possible impact upon my project. Interesting stuff, and I'll have to think about how it might integrate into my work.

SIGGRAPH sketch deadline is April 7th 1999. That is three weeks away today.

Macintoshes are the spawn of the devil. I spent about an hour trying to install Adobe Illustrator, and when I did eventually install it, it overwrote a newer version of the QuickTime utilities that were already present. This caused just about everything else to stop working. So now we have to download the new version of QT 3.0 again. ARRGH!

Set up the MWC99 mailing lists.

MWC99 web design stuff.

QT 3.0 installed, after about 2 hours of pissing about trying to deal with stupid MacOS 8.5 bugs.

I've managed to tidy up most of the loose ends involved in the changeover, and deal with all of the design issues. Now all that's left is a few files, which contain the real meat of the issue; csggraph, octree, raytrace, split, view and wglview.cpp.

Wrote Collection display code; wglview.cpp is finished. Next up is implementing the CSGgraph's Collection tracker, so that the primitives may compute their octree quickly.

Beating people up on various Myth forums.

MWC99 rules design.

The MWC99 hotline server is up and running. Yay.

Finished all the octree coding for level 1 of the changeover, only incremental_update to go.

Argh, realised something rather fundamental about how ghosts / inactive Collections have to operate. It's going to require some recoding in quite a few places...

octree.cpp is now finished, apart from the changes necessary for the new ghost mechanism. Shouldn't be too taxing and I should be able to do it later on this evening. Now it's off to the gearbuy.

Slow Myth day has let me get a lot of work done!

Returned the pack that I borrowed from Rec. Services. Tonight is the gear-buy at the Wilderness Shop so I should be able to equip myself with all manner of expensive toys.

I seem to have gotten lumbered with supporting Mirage again this year. And to make my life extra fun, Codewarrior refuses to compile the code and crashes for no apparent reason when compiling prim_cube.cpp...

Hopefully, that's the last I have to see of Mirage for a while. And the last I have to see of third-years.

Karan gave us a great demo of Maya this afternoon.

MWC99 planning, lots of it. Making up for the slow Myth day yesterday, sigh...

Played a M2 tournament game with my order; God, it was appalling. The tournament organisers picked a map that's not designed for 2 teams, but for 6; so the starting spots aren't arranged well around the map. We started right next to the other team and they made it to the center first -> nothing we could do but lose. Suck.

Helped Case and a few others do final testing on a new netmap for his tournament.

Changed the ray-tracing code to work with the new RCSGgraph structure.

Fixed the display code. Time for the big compile and Phase 1 check.

Fixing small errors, typos, code inconsistencies.

The code compiles and links fine; and now, the task of making it all work together begins.

Causing havoc on the Marathon's Story page.

Debugging, fixing small problems and logic errors.

Oh My God. Geoff just told me that because of the SIGGRAPH review process, my sketch needs to be completed by the 29th of March, far from the 5th of April that I was shooting for. Now it's time to really start working; this is going to be a tight one.

I've finished debugging and testing out all of the changes that I've made for Phase One. In the meantime I've uncovered a problem with the design architecture that I built, which will need to be resolved. I'll come in later on this evening and fix this if I can.

Every time I hit 'kill program' in the debugger it doesn't close off the OpenGL context correctly, and eventually it screws up the OpenGL driver enough to require a system restart. So down we go!

It lives...

Phase Two (the voxel intersection octree generation) is complete. On to the next phase, which will probably only get tougher; actually reworking the triangle-splitting code to fit into the new Collection strategy.

Finished coding for Phase Three, it all compiles. Time to run and debug now; this should be the toughest stage, because it involves the triangle-splitting code that was such an S.O.B. to get working in the first place.

Fixed all the memory leaks, but it's totally non-functional. Sigh...

Woohoo! A little pain clears the mind... after practicing arm bars at Hapkido this evening, I came back in here and literally solved all the problems within ten minutes work. I had forgotten to quite finish linking up all the code, but when I did so it all started working. Straight out of the box. On to Phase Four!

Well, maybe not. The split code is all working, but there's something fishy about my incremental update stuff.

Coding for Phase Four.

After quite a lot of coding, I decided to discard my original idea for Phase Four's CSG operations, which involved a lot of complex graphics coding.

Phase Four is complete, but disturbingly it's not showing the speedup that I was hoping to achieve by now. I'll be able to evaluate it better once the last piece of code is in place, but it's not looking good.

Phase Five is now complete, but not totally debugged... there are occasional spurious cells showing up empty (eek!). It's not fast at all... 5-10 fps on my simple test scene. Damn. I will profile it tomorrow morning (it's now 10:40pm and I'm going to get some sleep) but I was hoping for much, much more.

Fixed the problem with this work log that was compressing my 'work' section up into the left-hand 20% of the page. Surprise, surprise, it was that transcripted ITS email up in February. Sigh.

Signed up for some tramping stuff.

Helped some poor third-year with some basic programming concepts that his lecturers really should be teaching him. Sigh.

I just had (7:23pm) the first debugging-triggered endorphin rush I've ever got. After about 90 minutes of tracking, isolating and analyzing test cases, I found the problem deep within the heart of my triangle-adding code for cell checks. I discovered that the triangle-buffer wasn't being reset between cells. Not only does this solve a large number of the mysterious triangle errors, it also means that my technique ISN'T cripplingly slow, it was just doing 64 times as much work as it had to. It felt so good to finally fix that problem, I almost wept. Of course, the fact that I'm seriously sleep-deprived helps as well; I've been acting irrationally all day, as George will tell you.

There's something seriously wrong with either my ray-tracing code, or the way that I'm handling inverted surfaces. I think it's the inverted surfaces code, as this is completely new whereas the ray-tracing has been working for months. I hope the ray-tracing code is working. If it isn't I don't know what I'll do.

Time to get some dinner.

In the end, I decided not to come back in to work after grabbing something to eat, because it was 11pm already and I was tired.

Spent several hours helping Case by hosting the final of the Seven Phoenix Rising tournament. I need to get this sketch done, but helping Case is probably better for my long-term career goals right now. Plus, it's a lot more fun. :)

When I tried to put together a test scene for the SIGGRAPH sketch (so I have some results) I found a few more bugs. I've tracked down the root cause of one and will fix it tomorrow morning; my target is to get this thing totally debugged tomorrow, and get results out of it too. Then Tuesday is sketch writing day.

I am definitely going on the aforementioned trip, up to the Heaphy track in Nelson. We'd be leaving Thursday evening and getting back the following Wednesday, after a 5-day tramp along the west coast of the upper South Island. Groovy, but before I can do that I need a jacket that takes up less space; my pack isn't designed for such long tramps and I'm going to need every cubic centimeter of space.

Fixed a stray-triangle issue. It's only 11am and it's already approaching my maximum working temperature in here. Damn damn damn.

Fixed a problem with stored inout results and calloc() zeroing a buffer.

Fortunately, the temperature cooled down a great deal this afternoon, and it's actually a pleasant working temperature.

After a great deal of considering, I've realised that the way that my SplitTracker is used to keep track of where Collections are... it isn't working. It's not going to work. I need a new method. Said method will probably take quite some time to code up and be hard to debug too. It's now 4pm on Monday afternoon; I have this evening to do it in, and that's all.

Coded up the SplitTracker and guess what... no change in the observed bug that I was trying to fix. Hmmm. It is now 11:30pm and the rendering DLL has started crashing again, so it's reboot time.

Oh dear lord. The whole point of the new SplitTracker code was so that I could have last iteration's data available to me. So what do I do? In the piece of code that gets the splitnode, it checks to see if there's a previous iteration and ZEROES THE NODE so that it won't be uninitialised. Thereby negating the entire point of the whole section of code. Arrrarrararararrgh. 12:15am and that problem's solved. There are still more though.

It's now 2am and there are still bugs that need to be worked out. I've accomplished an awful lot, but there's still more that needs to be done. My goal of having the code completely debugged by this evening just isn't going to happen; I'm going to get some sleep.

Fixed a really stupid bug with ghost insertion.

Fixed a bug that was causing terminals to mis-report each others' full/empty status due to return value overwriting.

After about four hours' work, tracked down and killed an extremely subtle bug that was causing child octrees to have some of their parent's data written into them. Oops.

I think there's one last bug left. It seems to be related to incremental-update and the splittracker code; I've written a whole suite of test routines and after dinner I will have to fix it ONCE AND FOR ALL.

Fixed the problems with the splittracker code, I think... however, when I go to deallocate it it crashes now. So there's still something wrong.

I wish NV4OGLNT.DLL wouldn't crash when my program did something wrong. This rebooting is a real chore.

Fixed the outlier-ghost problem again. Now the piston-head runs up to create_csg and I should be able to manipulate it okay... fingers crossed.

Nope. Crashing again. After about an hour's debugging I fixed the problem (incorrect reference count).

Fixed some problems with the rotation DMD.

It's not perfect, but the system is in a demonstrable form. Should be good enough for me to write the paper tomorrow morning. It's now 12:40am and I am going to go home and get an early night so that I can be in at 7:30 tomorrow morning.

Writing the sketch! Three pages can be pretty hard.

The sketch is in final draft form, and I've done one of the figures. Tomorrow I need to get a good figure illustrating the system in action.

Massive debugging efforts this morning when I discovered that a critical part of the system wasn't working in my test case. Argh!!

Oh dear oh dear... far too tight for my liking. I've just uploaded the sketch, and we meet up for the tramp in 6 minutes time. Phew! I'm off to the wilds near Nelson for a well-deserved rest. Not an alarm clock for kilometers.

Sorting out the MWC99 design teams, who mostly haven't done much in my absence. I can see that I'm going to have my work cut out for me this week.

Went off to lunch and thence to a tramping-club meeting. Interesting stuff was discussed and much sifty behaviour ensued.

Tried to enrol in an evening class through Rec Services and was shot down! It's full.

Played in a Myth tournament match with my order, which we were winning when I decided to go to lunch.

MWC99 writing stuff.

Finished reading David Brin's "Brightness Reef" and I'm hooked. Bad. I'm now off to the public library before it shuts, to see if I can pick up the next in the series.

Woo, we won the match yesterday after I left, 4-1. Hooray for Civil Order!

Found a book on the Internet that looks very, very interesting... Freeman Dyson's latest work.

Lots of MWC99 work on story and rules issues.

Going to see 'The Matrix' tonight at the movies.

The movie was full, so I started working on the MWC99 server-side page updating software instead. About half of the data formats are done and being read in correctly. Still to go: reading the rest of the data, and page generation, finally image generation.

Spent most (hell, all) of the day creating the MWC99 server software. It's now reading pools, brackets, match results and basically all of the tournament structure. This evening I have to debug it and write a text output engine so that I can check it.

All debugged and the pool-match reading appears to be working, but it's 22:30 and I need to get some sleep. I haven't written an output engine yet, that will come tomorrow. Oh, and I also need to test the elimination ladder tomorrow.

Lots of MWC99 work, finished off the reading side and put in place the skeleton for the page output engine.

Went to see 'The Matrix' with Simon (since Jessica cancelled). A fun film, and pretty damn good. But it could have been so much better if they had a real actor in the main role instead of Keanu Reeves! Dammit...

Infinity's Shore is a great book! I started reading it late last night (whoops) and couldn't put it down until... er, late.

Did a bit of MWC99 web stuff, practiced a few games with my order for our match vs The Companions on Monday, but mostly read Infinity's Shore all day. Oh well.

Hoder and I spent most of the day working on little details about the MWC99 website's layout and it looks GOOD. Very good.

Wrote a couple of news items for MWC99.

Played in a Myth match with my order, which we won 3-0. Yay us!

Played the finals of the Forge and we won! Best of five, we trashed our opponents 3-0. Woo woo; I wonder if there are any prizes.

Put more of the MWC99 website together, including the first actual HTML. Currently working on automatic image generation.

MWC99 playtesting, picking out good game and map combos for the tournament.

Continued writing the automatic text generation stuff for the MWC99 page.

Spoke to Tavis about the possibility of getting the OUTC web page up and running.

Lots and lots of work on the MWC99 website, it's going very well.

Picked up Heaven's Reach from the Public Library where it had been held for me, and read it all that night. Damn fine trilogy, well worth reading.

MWC99: working on team pages and general structure. Set up the About and Story sections, finished off player pages, added headings. Created threat and rating icons. The next thing is converting story info across to the website, then comes the big challenge: match results.

Changed the split-triangle buffers over from fixed-size to dynamic allocation, thereby avoiding a debug-assertion and hopefully making the code more stable. Checked that my code is running so that I have a showable demonstration for the school visits tomorrow afternoon.

Withdrew all the money from my savings account in order to start paying back my student loan. Walking across town with $2500 cash in your pocket certainly does make you feel alive... and paranoid. Especially when you walk past Dunedin's resident crazyman on the way. :)

Apparently Puzza is spell-checking my work log in an attempt to prove to me that spell-checkers can be useful. And I say to you, Puzz, that spell-checkers are for the illiterate, and those who like to be told that all of their place names and friends' surnames are spelt wrongly.

Proofread + revised more MWC99 story stuff.

Updated Mirage to v1.0.6.2 to fix an incompatibility with the new version of Red Hat Linux that our sysadmin's installed on some machines here. Specifically, ecgs-2.0.90.29 doesn't like some of my C++ constructs. Pffft.

MWC99 playtesting and writers' meeting.

Drafted out some code for the MWC99 results sections (match + game reviews).

Started reading "Godel, Escher, Bach: An Eternal, Golden Thread" by Douglas Hofstadter.

After writing all that code, I ran it and got out polygons that looked a lot like multicoloured snow. D'oh! But looking at the patterns of the snow, there were these pretty triangle-shaped artifacts. So I looked at the loop variables carefully again, and there was a j swapped for a k. Fixed it and it all works perfectly. I love it when a plan comes together!

Went to catch the movie 'TwentyFour Seven' at the Regent, as part of the Film Festival. Depressing, but very worthwhile regardless. Nice to see a film once in a while that leaves a few things unsaid for a change.

Tracked down and nailed that goddamn bug. It was a mix of things; a conceptual error was being masked by a pointer-increment inside a loop that was being bypassed by a 'continue' statement, which was causing the error to manifest in the next Collection that was split against the same object. Phew, now it's done.

Rewriting the way that the TriSplitter interacts with the Collections. Much more elegant now, and no longer crashing mysteriously (?).

Spoke too soon. Bugger.

Hunted and squashed a few more bugs.

Trying to work out how I'm going to get to LA, where I'm going to stay, how it's going to be paid for.

Oyez, oyez, it is fixed, and the cut edges are all nice and smooth and fixed. There are still a few bugs in the code, I believe, but all of the really big ones are gone. Phew!

Travel arrangements for SIGGRAPH. I need an ISIC card and a passport now, will see about getting application forms for those off tomorrow.

Had a meeting with Geoff and Richard, talked about my PhD in between digressions about compiler technology and related issues. Worthwhile, I guess.

Planning for MWC99 story stuff, and also how we're going to manage publicity.

I have to speak at the graphics meeting today, so it's time to plan out something for that.

Graphics meeting went well, helped me work out future ideas in my head well.

I'm going tramping for the weekend, down to Port Craig on the south coast. It may be a bit chilly, but I'm sure we'll survive.

Fortunately, the text for my sketch has been finished for months. But I have to get the figures into the format that they want now.

Eek. I had to render a PostScript figure at 900 dpi so that I could then compress it into a TIF... and it only renders into an uncompressed PPM format. That means I had to fling around some rather large files, viz.:

-rw-r--r--   1 cbutcher stage4   938297312 May 10 16:57 temp.ppm

Fired off the finished sketch materials for publication.

Spent an awfully long time catching up with issues raised over the weekend while I was out of contact.

Our Internet connection is absolutely unusable. But I have to keep trying, hanging on for every stray packet that leaks over the NZ -> US connection. Argh.

Went to a psychometric testing seminar. God, what an utter waste of time. Basically the speaker went over some sample questions from primary-school-level verbal and numeric aptitude tests. None of the insight into personality measurement that I was looking for.

MWC99 stuff.

Installed Quake III: Arena test on my research machine. It's pretty, but it's just Quake. Shrug. Had to update the display drivers to do so, fortunately nothing else seems to have broken.

An awful lot of MWC99 stuff. We're really getting close now.

Lots of MWC99 stuff, perl hacking, discovered that my main web designer has quit (argh!). So tomorrow is Photoshop day, I guess.

Spent almost all of Monday trying to get a birth certificate from the Department of Courts. Quite a runaround.

Opened the MWC99 website for business; this should ease my MWC workload off quite a bit, hopefully, and let me get back into my PhD.

Entered the first team into MWC99... woo.

Some paperwork for Geoff arrived today from SIGGRAPH and it looks like I'm going to have to deal with it (he's out of the country attending a conference). Sigh.