First Tikid, Hour of the Mage, 250/310 (9/7/10)
OK, about to run a network test, while that copies, I'm updating this. Am loading levels successfully, and have enough of the TCP out working to get the sun cycle set to the network server. The textures are holding their detail up close to the camera, for like the first time ever since I started working with Torque, so the levels look good, the textures colors are very rich all the way across the geometry. I guess that's progress.

(later) Well, that's a pisser--can't find the master server. Thought I'd caught the appearances, but must have missed one or some other flag needs to be flipped, but the game can't find the ghost server from off-LAN. Guess I'll have to look into that, or else I can't run obsessive tests all day.

Second Tikid, Hour of the Daystar, 249/310 (9/6/10)
Hacking away at this "port" thing here. Took the weekend off for the first time in more than a month, specifically to work at this. Needless to say, I've destroyed my sleep cycle over the last few days...

This is a tough port. I've just got the missions loading at this point, with the Antran chrome on the loading screen. Which isn't much, and I've been at it for hours today, on my sixth attempt. Yesterday, I hacked in all the C++ needed for GreenEar and the compass. It's the script side of things that's holding me up.

Third Tikid, Hour of the Dagger, 240/310 (8/28/10)
W00t! Successfully ported the "official" web site to a new server!

This takes it from apache 1.3 to apache 2, from PPC to Intel, and from vertical stack to distributed stack. I've had a skunkworks site running on apache2 since January, but have been... trepidatious... about moving the old site. When I started working, I didn't know quite what I now do, and things were quite a bit more primitive with fewer tools to help. I had to write everything, from the site editor I'm typing this in to all the little callback routines now known as "AJAX". The integration between webserver and application was a lot tighter than it is in my current work, and as the site grew organically over several variants over several years, things got crufty.

They're still crufty... this was a straight port, not the full rewrite I've been promising myself for the last couple of years. However, it is a step back on the path rightwards, and I'm pleased with the speed, especially, so far.

Anybody notice anything funnier (than usual) please let me know via u2u or Forum!

First Tikid, Hour of the Priestess, 235/310 (8/23/10)
Worked through the weekend, but at "work" work, for the second week in the row. The AFX 2 build for Beta 2 was released, and I've downloaded it, but a port is going to be trying. I've taken one stab at it--thousands of files were touched between Betas 1 and 2, and straight diffing proved improbable. I've concluded I'm will have to reapply all the patches and resources one by one from their source texts--all--ALL--the console commands have been modified, so all the script classes I've written are going to have to be changed, as well. That's one hell of a lot of work, and I'll lose at least one more weekend to "work" work before I can accelerate in serious manner. I've got too many grading tables and work orders swimming around in my head at the moment to think about animation hooks and inventory systems--though I did get a pretty good insight into skill trees which I spent some time writing up in a spare moment last week.

It's fixin' to be over 100 degrees F here today, too. That does not help--maybe I'll get some documentation read (I have a foot-and-a-half of books to get through) if the evening cools off.

Third Tikid, Hour of the Jewel, 227/310 (8/15/10)
A name for a production company finally occurred to me, and--this is the modern cornerstone, I think--bought the domain name (if the URL is available the Universe is looking with favor on you, right?).

Our mission is to create something slightly tastier than a quest-loot-and-reward multiplayer game. Not that there's anything wrong with quest-loot-and-reward systems. They work very well, and they serve as a good model everybody knows how to play already. We just want to write above a third-grade level, and to implement branching quest chains, with different outcomes for characters, depending on choices a player makes. Thus, Consequence Games.

Third Tikid, Hour of the Priestess, 224/310 (8/12/10)
Got a screenshot up yesterday, which is visible above. I didn't crop it to fit, instead adding the letterbox bars, just to get the panoramic camera angle in. It's very crude, but I think I am starting to get the hang of it... I started from zero 25 months ago with no 3D or C++ experience at all, but I have kept working at it in my spare hours. Heck, I spent the first four months after buying TGEA 1.7 crashing more than thirty times a day trying to make terrain, until Rene Damm turned up in the Torque forums and documented a fix which pointed out the problem was in the engine code, and not my ignorance! Then I worked in a misinstalled version of Constructor for six more months, not realizing that I had screwed up the install because I did not understand how copy worked under Windows--on the Mac, copying a folder of the same name to the same directory replaces the directory in toto--under Windows it only replaces files inside one of the directories of the same name and leaves other files in place. Constructor updates are supplied as patch folders, not installers. Once I twigged to that, I realized what the problem was and reinstalled (the symptom was that if you alt-tabbed, you would lose the menu bar, and be left in full screen with no way to save what you had up, or exit, even with key commands).

I'm still WAY behind in the learning curve. I've known for a long time now it takes a commitment of two years for me to become minimally functional in a new skill. At this point I'm pleased that I took the plunge--it was scary buying an extremely expensive, professional piece of software I didn't have a clue how to use! However, I was bound and determined not to get involved with Adobe platforms again, as well as realizing that it would take me fifteen years to develop my own engine again from scratch (I'd put ten into the 2D multiplayer engine). And today, even if for just a brief few moments until it all closes in on me again, I am pleased with the result.

Second Tikid, Hour of the Priestess, 219/310 (8/6/10)
Well, T3D finally made it to Beta 2. This morning will be the first I've had time this week to check out the substantive new docs delivered on release; at the moment I'm scanning the patch notes, which are insane. A very large number of longstanding problems with the engine got addressed, but it is the freshly-minted documentation I, and most of the Torque community I would say, are interested in this time.

Torque has long had a community attitude of documentation doesn't matter, just read the code, and it has cost them dearly in the last few years, I think. At large, the price of that manly smellin'-the-pits attitude was a generation of young (and not-so-young) would-be game builders--the problem being that this generation of would-be game builders is larger than ever before: due to the recession, there is a wave of everybody-in-their-garage-is-building-an-MMO similar to everybody's-grandma-needs-a-website in 1997. It's a great time to be an engine-builder if you have an engine people want to buy.

Not having accessible, documented, or well-kept code hurt them badly in the wake of the release of competing engines which made it easy for these people to see results without the Torque tradition of poring over scraps of aging code in the Byzantine ruins of the various community repositories on the Torque site.

If you could make a game good enough for you and your friends without having to deal with C++, why wouldn't you? Except for the fringe cases who understood why they would get with C++ and code, the amount of effort needed to learn the path of true power just doesn't appeal to the larger audience. The number of available charlatans is always larger than the number of available wizards, is it not so? And that last right there, folks, shows just how easy it is to slip into the mindset of the historical Torque community--you get to think yourself a wizard if you can do it, and they can't. Which works as long as you are the only game in town, but when someone puts up a lemonade stand next to yours and starts giving away an engine which abstracts away the C++ entirely--for free--that attitude needs to change, and it took the internal culture at Garage Games about a year and a half to react. T3D is a great engine--but it requires skill and skullsweat to make anything happen.

I'm glad they did, sincere kudos to Eric Priesz and Michael Perry, who led the documentation effort. I'm looking forward to examining what they have come up with--I'll have time to read the docs, I can't begin porting the WOA code until the parallel release containing the AFX framework used for special effects comes out. Soon, I hope!

Second Tikid, Hour of the Horse, 212/310 (7/31/10)
Got the compass mostly working, finally, after several days of work cobbling working code together from several old resources. Finally seem to have it nailed, except for minor problems if I go in and out of the editor. One of the AFX test routines draws into the same buffer as the compass face somehow, but it's actually pretty and seems harmless. It's more annoying that the compass face ghosts if one resizes the window larger before the load completes.

Second Tikid, Hour of the Priestess, 208/310 (7/27/10)
Last night I spent time again in the search for a better display face. I put in a couple last night, and then got up this morning and ripped them out and put in another set to see how it worked. Picture above is a screen cap of the sculpt test for the Barrens zone. I also set up a sculpt for the Crags-Under-Moon, but that's not lit at all yet so no screen captures. Kind of itching to do the model for Thale, though.

Third Tikid, Hour of the Priestess, 206/310 (7/25/10)
I implemented health and mana bars yestereve. They work a little differently than you'd think, because I want vertical bars which drop as their values lower, and the default widget drops right to left. So I instead reversed the modifier and am using the bar as a mask--as the value drops, the actual bar rises and masks the colored bar, which is a cosmetic graphic.

Then I ran into a problem which I think turned out to be caused by the tracer I had printing back to the console--it created such a flood of messages that the "good" packets were swamped out. Taking the debug print statement out seems to have fixed the issue where some status bar updates were getting lost and everything (spells, swords, player camera dollying) works flawlessly this morning.

AIPlayers are refusing to die, and seem to have lost connection to their datablocks, making them unkillable. That'll have to be what I look at--after I make a full backup, which takes forever, but everything seems to be working and I don't want another week(s) of hell putting it back together after I boggle the next thing up.

Second Tikid, Hour of the Dragon, 205/310 (7/24/10)
OK. That was a bit of a long stretch... I reimplemented the melee code again, because there were things happening I didn't understand and didn't like. This time I cooked it with a slightly different recipe, and it finally seems to be working in the same fashion on both server and client. Torque's scripts are hellishly confusing--for all the "power" of namespaces and classes, mostly what they seem to do is cause straightforward tasks to turn into incomprehensible spaghetti, in practice over any length of time. Tracing and identifying which of a dozen identically named routines is actually the one called in any given instance can take hours.

Even more oddly, I see different behavior out of different mission files, which are just text files loaded into a common engine. Since the engine is common, and the text files contain no scripts, etc., I do not understand why, for example, the horse steers completely differently when mission A is loaded than it does in mission B. There isn't any mission-specific code anywhere.

Now that the melee code seems to work for animations and damage, I need to investigate why when a spell is used, the puppet is staying dead, but when killed with a melee weapon they get right back up again after their dying animation finishes. Another delve into source follows...

Second Tikid, Hour of the Priestess, 202/310 (7/21/10)
Well, I was in some pretty deep despair before I dropped onto my couch about nine last night--I had managed to eliminate the display of the puppet, somehow. Footprints--the effed-up bad-blend footprints--were appearing when I drove the camera around, the avatar just refused to show.

This morning I got up, and retraced the implemented code, and this time, flipped two flags from what the "correct" position was supposed to be and the puppet reappeared. I had been considering a full reversion to the base AFX and reimplementing the various packages I've been porting, but suddenly I was pulling 39 frames a second in a mission where I normally see 18, and am going to take that as a sign to at least complete fully implementing this last resource before giving up. I think I may have figured out the player pack/unpack routines--I've certainly been putting a lot of hours sifting through the C++.

All of this is having a huge impact on my other gaming--once I open the various WOA build tools, I keep finding one more thing to work on, and I don't get around to opening City of Heroes. I'm not playing Star Trek Online or Champions at all, both of them leave me disinterested and disengaged about five minutes after login. CoH is in kind of a expectant lull, waiting for Going Rogue to drop, and Cryptic's games have left such a bad taste in my mouth I can't get into them at all; they're actually repellent.

Second Tikid, Hour of the Priestess, 200/310 (7/19/10)
A mixed bag this weekend. I got the groundcover replication in the Barrens test zone working nicely. Also resculpted the terrain to more properly indicate the line of the blowout.

Took the invulnerability flag out to create some spare bits, in an attempt to figure out why things like only the first attack animation used in a session was playing on the client. Like you would use "Thrust" and then "Slash" wouldn't work, only Thrust would play, while all of them worked on the server.

Then, while that was compiling, started working on the interface GUI, because having gotten login working so the player can be uniquely identified, the next step is creating/fetching/saving characters. This requires creation of several new GUIs, and linking them all together with client/server commands. Some of it looks ok--the buttons are sharper-looking, and in testing remotely it looks like the UFT symbols were properly compiled and saved so the font chosen shows up on clients (where it isn't installed).

Also adapted the "thorn" frame from the website to serve as cosmetics for the interface. I'm getting a little better at getting things done in the GUI editor.

Looks like the skin routines are broken--I can see the server's skin on the client, but not on the server and vice versa. Last night I found I couldn't kill NPCs suddenly after successfully hitting them once with the sword. I may have to consider rolling back to the shipping AFX, then reapplying the various kits again. After a while Torque starts to resemble a rabbit hole, with the overriding and inheritance.

I'd prefer to wait until the next code release, but that's a ways off yet, and looks to contain a massive set of changes, and AFX will be some weeks after that--probably looking at September earliest, so I don't want to wait. Applying the patches and resources gets harder and harder, as many of them were written for older versions of the code themselves. Translating forward and then intermingling the parts which touch network communication is complicated, and since each test requires a compile and launch, slow.

Third Tikid, Hour of the Daystar, 196/310 (7/15/10)
Take that, you fiend! Boy, there's something just wrong about that animation, but there you have it--a new axe hitting something for damage!

Well, it's exciting to me, given there's over a thousand hours of work over the last two years getting to that point. Darn learning curve.

First Tikid, Hour of the Moon, 192/310 (7/11/10)
Still polishing the login. Seems solid, but needs to be tested from off-net. Code is unbelievably complicated, with overrides, client and server commands, etc. Had to go through the scripts (all 727 of them!) with WinMerge to locate a line I commented out, thinking it didn't matter, before it stabilized. However, now you can type a bad login, the client will quit after warning you. The server survives bad logins, and still lets in good ones. And the logins themselves are checked against the new player database.

First Tikid, Hour of the Rogue, 191/310 (7/9/10)
Got up at 4:30 am this morning. It's 10:45, and I just got the login function to work.

Time for a celebratory beverage!

Second Tikid, Hour of the Mage, 189/310 (7/8/10)
I'm convinced after testing yesterday that the horse animation issue is related to the observance the client player only ever sees one attack animation, while the server sees all three. The question is if the code is buggy, why does the client see any animations?

Nonetheless, I proceeded to work on implementing the login system last night. It's funny how games can ignore really important bugs forever so the work can proceed. The kind of concentration I'd need to raise to think through the networking issue right now has to wait for a weekend anyway.

Third Tikid, Hour of the Dagger, 187/310 (7/6/10)
Yesterday began in frustrating fashion, with work continuing on the horse, and then took a sharp turn into the bizarre when I created a new zone and the horses became convinced they were dune buggies... the sporadic animations and sounds continue in the main mission, but are completely absent in new missions... missions don't contain any character-specific code at all.

On another front, however, awesomeness happened--I purchased some new weapons models from 3DRT. Now, stock models will not work with the server-based melee system I've currently got implemented. I asked 3DRT to quote me a price for modifying the models to add the required tweaks, and was very surprised--and thrilled--to receive a modified set of swords and axes via email yesterday afternoon.

Debate is still raging for me over whether to go with a "roll-to-hit" melee system, or use the "actual player skill" mechanism in-place at the moment. The "actual skill" system was looking undoable due to the model requirements, but suddenly having these beautiful assets drop out of the sky makes it contender again--not looking to make a world where every class plays alike, after all!

Roll-to-hit systems are very predictable, very "fair", very flavor-of-the-month. Actually having to hit something with a stick takes a little more doing in-game than selecting a target, and for a melee class feels very different than targeting and pushing a button to fire.

First Tikid, Hour of the Jewel, 185/310 (7/4/10)
OK. Got up at 2:30 yesterday am. Worked on the C++ for the horse in Torque. Finally was able to get the player model to mount the horse, suddenly, about noon. Cleaned it up a little, but it is so crude... the galloping animations aren't appearing at all this morning, they were only appearing sporadically yesterday. The "hoofbeat" synchronization, when the animation fires, leads one to think the animal has had more than a little to drink!

But the mounting is reliable, if crude (charge at the horse, give it a second, reappear in an awkward position atop the animal. I figured out that it didn't walk like an alt; instead, you roll the mouse forward and back to speed up to a gallop, or slow down to a trot. Whatever speed you "set" then works with the normal forward-and-turn keys until you change the speed with the mouse again.

Still, I can get on my bad motorhorsie and ride.

Third Tikid, Hour of the Sword, 183/310 (7/2/10)
Got up at 5 all week, but unfortunately not to work on the game. Hellish week at work, trying to finish the unfinishable module, which turned out to be a LOT more work than planned. Horrible.

I did just get horses in, although there are substantive problems with the collision mesh. I'm not sure if the issues are the same for vehicles and players. The horse idle animation, though, is the best I've seen for Torque. There don't appear to be any start and stop points in the cycle, it's absolutely smooth. Whoever animated it must have spent a year staring at it and tweaking it!

I'm going to have to work this weekend (working on 4th of July is my personal curse, I'm convinced), but I am going to try to take a very careful look at the horse collision code--the symptom is the horse doesn't stop at a .DIF collision, only with terrain... but then it slides off the terrain until it lands on a .DIF, at which point the collision is normal and the horse comes to stop at the correct height.

Bizarre? Bizarre. Kind of neat to watch, though. I'll get a screen shot up as soon as opportunity presents itself.

Third Tikid, Hour of the Jewel, 179/310 (6/28/10)
Woke up yesterday at dawn to find the arc of the sun well underway, so I had to revisit the math behind the time/elevation thing again.

I believe I have the math now. The sun set precisely at the Hour of the Rune. I don't know if it rose this morning at precisely the Hour of the Jewel, because although I was up, I was too fuzzy at the time to think to log in and check. Several cups of coffee later, it's too late.

Also did quite a bit of model work yesterday, hacking the animal models I'd bought o-so-long-ago and had no idea at all how to use so they would display the proper skins finally. Then I whacked at the web tools being developed to support the game complex--that was a bent mess, so I straightened some of it--as I need to start with the Account Creator/Character Store/Login Mechanism and it's all of a piece, really, isn't it?

Thanks to Ted Coffman at Green-Ear for all his work getting everything set right the other week!

Third Tikid, Hour of the Dragon, 177/310 (6/26/10)
OK, that fixes that problem. I had rigged up the server to transmit the hour to the client, at which event the client synced up its sun with the hour and then time was allowed to proceed. Noticed the other day that if the client logged on late in the hour, they would get light as of the beginning of the hour, while the server may be off by an 90 minutes in lighting. This means someone might be in shadow, while another has daylight.

Spent the morning putting in a much finer sync, which now keeps clients and server in perfect shadow sync, so noone is peering through darkness looking for other players unless everyone is.

Second Tikid, Hour of the Rune, 172/310 (6/21/10)
This setup is ridiculously stable. At least for a server population of two.

I've been logging into the server and "staying on" all day from work for two weeks now. It's not quite the same thing as "active gaming" as it's just running on a laptop quietly while I work on my desktop, but no disconnects and no crashes, with a 100% connection rate isn't horrible for a first round of testing, is it?

First Tikid, Hour of the Mage, 172/310 (6/21/10)
Happy Highsummer, everyone!

Modeled on Saturday. Got myself into some kind of mode in Constructor, and had a tough time in the afternoon with things going "soft", and collision disappearing. Wound up deleting what I had started on in Rough Zirkot and building it again after I figured out a way to turn the mode off.

Sunday I managed to successfully reskin models as different NPCs. I spent all morning working on doing it the "right" way, with commands in the engine, but couldn't work that out--changes don't propagate to clients, so the engine command is useless--so finally I whipped out a sector editor and went after the models themselves. Changing their skin names internally on a per-model basis did it, and I was able to introduce "new" races based on the same models. I played around with scaling the Adam model down to Dwarv-size--certainly not perfect, but I didn't have any passable Dwarv stand-ins, so I'll do a bit of skin painting and use them, too.

I can see myself burning through every available low-poly model rigged for Torque in the world in short order. Maybe I ought to "reimagine" things as a Human-only world and buy Poser! He he aaaaaauuuuugh.

First Tikid, Hour of the Dragon, 165/310 (6/14/10)
OK, I put up a fresh screenshot. This one is brand new and of a piece of modelwork I was working on yesterday. Yes, the proportions are a little off if you click on it to enlarge it--when I built the code for this page, the proportions of the images being enlarged were different, and they are a little off for screen captures. I'll fix it in the next site rev--this one has been solid for four years, and as I mentioned, I have a hankering to do a new site for the 3D world, just not an immediate need.

First Tikid, Hour of the Mage, 164/310 (6/13/10)
This VOIP is working surprisingly well... I got it working, then went out to run an errand. When I got back, the channel was still open, still working, and continued to be so during tests at hourly intervals all afternoon--well, at least until it got so hot that I melted. By the time I got back from my errand, there wasn't much to do owing to the sudden heat, and I'm afraid little programming got done--thinking was kind of impossible.

Today I was up at 6:00, showered, and out of the house to get groceries out of the way. Then I set two bottles of wine to chill. Now it's just into the Hour of the Mage, already over 70F, and looking to be another scorcher of a day. Good for reading documentation, bad day for trying to think! The AI kit looks interesting--on a clean install of T3D, I almost immediately had mobs chasing me. Certainly worth studying--also reading Matt Buckland's Programming Game AI by Example--load it all into my head, and let it gell.

First Tikid, Hour of the Dragon, 163/310 (6/12/10)
Well, surprisingly, I won't be reporting I ripped the VOIP code out... because I got it working this morning. Late last night I got another email from Ted at Ayalogic Tech Support and this time, I was able to log in. It was a little late--and I was a little deep in my cups lol--to start coding at that point, so I hit the sack and got up at 6:30 this morning and started in. Now it's working transparently--you start the server and it creates a zone chat, and when the client connects, they are automatically dropped into zone chat.

Now that's just stub code, and I'm going to have to back some of it out--people who aren't going to use it or don't have microphones don't need to keep an extra channel open, burning bandwidth--and add a connect/disconnect button explicitly, along with a "mute" button for those who don't have one on their headsets. That ought to keep me busy for a while--the GUI builder is a bit of a struggle for me--and then, I suppose, I'll have to put a corner of my mind to work thinking up how to let players create their own private chatrooms, for team use, etc.

Haven't gotten any modeling done yet, but I think VOIP is enough accomplished to earn stretching my legs (and another cup of coffee).

First Tikid, Hour of the Mage, 162/310 (6/11/10)
Goals for this weekend: Modeling on Saturday, clean install of AI Kit from Twisted Jenius for mobs and pathing study.

I'd say "finish installing Green-Ear code" but that has turned into such a debacle I'll probably just put in a coda on Sunday afternoon which reads, "ripped VOIP code out." A week of "it's fixed" emails daily with each successive attempt just to get into their website--which went down mid-week for a day--to set up test accounts, well, not succeeding. I'm sure they're struggling--I've never understood how their pricing model could be sustainable, although as a developer, I of course love it and it's probably necessary to even get small developers to consider it (flat, one-time fee vs. a subscription. Subscription pricing is a boat anchor to a tiny developer, who's usually paying for most things out of pocket)--and the near-invisibility of their product on Google searches convinces me of it.

I will put up with quite a bit. I've worked for some very small, very struggling companies in my time, and I know how it is. Still, the bare minimum bar of having usable product has to be met.

Second Tikid, Hour of the Warrior, 159/310 (6/8/10)
I'm convinced at this point the problem lies not with my code, but with a problem on Green-Ear's side. This conviction is in some way supported by numerous emails back and forth, but apparently I am lost in their database and tech support can't find me to fix a "known-bug" in their system involving channel passwords. Today, there was nothing--just silence--after the first exchange of emails Monday.

I am wondering at the wisdom of pursuing integration of this system--anything put into the codebase has to work, and pretty reliably, too, to make the cut. I'm keen to include the feature, and this seems on face as a decent way to do it. But that which should have been fairly simple--some C++, a link or two, a little script--is bogging down.

Anyway, I am off to vote in the primary. Later, I'm hitting the AI Kit and going to play with that while I wait out the G-E situation.

First Tikid, Hour of the Dragon, 156/310 (6/5/10)
Got up early this morning and got to work on integrating the voice framework. Wanted to see if I could pull it off, and before I had to drag myself away from the keyboard and come into work, I believe I managed to do it. Of course, the script side still has to be tackled--the code is compiled into the engine now, but nothing is calling it. I did grab the client off onto the laptop, and introducing the libraries and headers for voice support doesn't seem to have collided with any other function, as I've got it up and running a test off on the side of my desk at the moment.

Second Tikid, Hour of the Warrior, 155/310 (6/4/10)
The second test concludes successfully. The game client was able to connect. The connection was lost once, and after docking the client window for an hour the client slowed down so much I quit it and reconnected.

I was surprised last night at how little trouble I had fixing the chat connection issue from yesterday's test (now that I can actually say, having tested off-network, it is fixed). It is very important that clients connect only through the ghost server, however--they should not connect directly to the logic or store--those connections are much less secure, and it makes backend administration a headache as well. Let's face it--the potential for missing opening a port is directly proportional to the number of ports one has to remember must be open. Right now it's two, and I'd like to keep it that way.

Anyway, successful test, very happy with results--also got some bandwidth measured--looks like 3.5-4 Mb/hour bandwidth usage on the client, or ~500 bytes/second. That should work even on a dialup connection. I need to fire up CoH and see what it uses on the same laptop in an hour.

Second Tikid, Hour of the Priestess, 155/310 (6/4/10)
Chat's fixed--it's neat! Yeah, I had to test it and get it out of the way, or it would bug me all morning when I really need to trance out and code. I think I'll leave it up on the laptop screen like it was a vacation picture on my desk--except in my picture, the clouds are moving and the shadows are changing as time passes LOL.

Note to self: find out what the parameters are for triggering the lag icon, I think they are much too low for internet access. Also need to move the thing out from behind the chat box.

Third Tikid, Hour of the Jewel, 155/310 (6/4/10)
OK, I think I managed to get the chat running through the ghost server as intended with some work last night. I'll run the chat test again today if I get a minute and remember to take lunch--it's going to be a busy weekend at work work, unfortunately--several unrelated mishaps severely impacted a delivery schedule and I'm going to have to put in some hours to catch up. Which sucks balls.

Be neat if the chat's fixed though.

First Tikid, Hour of the Dragon, 154/310 (6/3/10)
Got summoned to the office very early this morning to fix something for someone. Right smack dab in the middle of my R.E.M. cycle, apparently, because at this point, having run home and showered and pounded some more coffee at 10:00, I'm still walking around stunned. More coffee is at hand, so I'm going to try that first.

Set up last night to run the first extra-net client test, which I just remembered to actually run. Sure enough, I can log in to the server remotely from the client. I'm a little surprised to find out that the client is actually sending chat directly to the logic server, rather than through the ghost server, as I had intended. This is, however, why we test--that behavior would have been completely invisible on the LAN, where the machines have broad permission to speak to one another; it is only exposed off-LAN, where access must be deliberately opened in the firewalls.

Ah! But getting it working will be a fairly simple proxy affair, I believe (but that's not the same as fixing it--I want clients to communicate through their ghost servers... although this might be advantageous--it's vastly less complicated, and could make channelization much, much easier). I'll see--when I'm a bit more coherent.

First Tikid, Hour of the Dagger, 153/310 (6/2/10)
Figured it out just before I left. It's the river which crashes out loading clients. Nice to trace it after several days of iteration and find the offending object, which is sort of the key to preventing such mysterious issues from occurring again--not to mention building with confidence--but the problem is the terrain has a nice valley sculpted into it now where the river should go. I already knew the river editor was bad, and would crash the client if you attempted to edit the positions of the nodes, but had been working around that by editing the mission file by hand. Now, somehow, loading a river which works perfectly well on the server locks up the client and forces a control-alt-delete. Nice to know it wasn't the lights, which is what I've been thinking.

Third Tikid, Hour of the Jewel, 153/310 (6/2/10)
Tested up until I had to leave for work, then started testing again when I got home, slowly working to restore the level, the glow and the spheres. By about the Hour of the Houri, I achieved results, but it was agonizing and I'm still not sure what happened to cause client loads to take a dump.

This time, I not only got the glow to appear, but the textures to animate. In the borken version, the materials editor had somehow made the glow material a decal and stuck it in the road editor. Not being sure what happened includes "there's a serious bug in T3D's material editor, but I'm not quite sure how to describe it in words." "Augh?" might do.

I think I'll not use the materials editor for now, in any case.

At least I finally broke down and bought a portable Iomega firewire drive. It is red, which I find pleasing. I partitioned the heck out of it, and have three bootable volumes backing up Dispatch, Data and Delegate. Can't get it to mount on the PC via USB, though, but can "copy through" the laptop over the network to it, so that will do for now.

Third Tikid, Hour of the Jewel, 152/310 (6/1/10)
That wasn't fun. I decided I had better test some of the new modeling networked, only to discover that somewhere, somehow, I had broken the client load. Afterwards, well, the entire day was spent backtracing the week's coding, trying to figure out if it was in the C++, or the scripts. Takes forever to completely recompile the build, so patience is very much required. I've got a working client build again, but I still am not precisely sure what caused the issue--just that when the client reached the end of the object load, it would freeze and lockup requiring a force-quit. Could not get the console open during the process to see what was going on. I'm beginning to suspect it has something to do with the glow effects, although since they were entirely developed inside the editor's tools, I don't see how I could have created a fatal script--usually I have to do that longhand.

First Tikid, Hour of the Priestess, 150/310 (5/30/10)
Spent the bulk of the day--ok, ok, the whole day--working on game models. Ran into a problem while trying to fix an egregrious seam in a street join in Rough Zirkot and it appeared that I'd somehow introduced an invisible collision box. Running uphill would suddenly have me running up an invisible incline, even if I deleted the model! That slowed me down for hours until I discovered an invisible clone offset with collision turned on in the game editor. Not the model. Took me hours.

However, I finally succeeded in creating a 3D object in Milkshape and attaching a texture to it! Then I overrode the texture with the in-game materials editor and now there are properly luminescent streetlights!

Mostly, I continued roughing out buildings and streets. I intend to pull the buildings out later and model them individually, so LOD can be added for each to help keep the frame rate up, but I really, really need to do shape-fitting first.

The glowing globes look very cool--I'll get some screenshots tonight and put them up soon.

Third Tikid, Hour of the Warrior, 148/310 (5/28/10)
OK, that was fun. I just connected a Firewire cable to Data and Dispatch, and gave them their own private network at 400 Mbps. Beats 54 Mbps using Airport G. With pings in the quarter-millisecond range, I think this will be a reliable and quick way for the logic server to talk to the database. If I can scare up a longer cable, I'll put the web server on it too, as the website is a heavy consumer of data as well.

It's a compromise... I still can't get the machines all on one side of the house, due to space and power issues. However, I can put the machines which need to communicate quickly all on a subnet, without getting a router involved, using Firewire, and that ought to greatly improve reliability and speed.

And it's fun. Not as "cool" as look-no-cables-wireless, but far more reliable and infinitely geekier.