Another Major Milestone and Great Leap Forward today. Hopefully. The game has 29 main puzzles (each made of sub-puzzles). I had planned to spend two days on each: one day coding, one day testing/refining. But the first two puzzles have taken four days each. In the words of Sypernanny, that is "not asseptable." Even though it I allowed for a learning curve, and extra time for the unexpected, this is too much. Yesterday, instead of starting on puzzle three, I spent the whole day bug fixing instead. No fun. So while swimming with the children at night I worked out a way to streamline the process (swimming? streamlining?) through using a standardized puzzle template. Standardized from a programming point of view that is - the puzzles will still be different! Today I will see if it works.

Today I'm adding the part where Valjean steals the candlesticks. Most adventure games (.e. Monkey Island and others) are built on stealing , so this would be a perfect puzzle: how to find the candlesticks when the bishop isn't looking! But I'm unusual among adventure gamers, I think stealing is wrong. So in this game we approach it from a different angle.

DreamHost was back up within a few hours yesterday, and I also added moving water to the river scenes. Tonight I'll add clouds to the sky.

I just checked out dreamhoststatus.com and everything is explained. The building that holds their servers had a planned power outage. Yes, I know that other more expensive companies have backups (and backups for their backups) but that is the point: other hosts are a lot more expensive! And it is reassuring to know exactly what is wrong and what is being done to fix it. With other hosts you are often left in the dark.

If I had a big healthy cashflow then I would use a more expensive host like LiquidWeb. But then if I had a big healthy cashflow my life would be very different in many ways. When compared with other providers who charge similar fees, DreamHost still gives better service (on average) than anyone else. And no, I am not getting any money for saying this!

It's 8:43 am GMT and I can't see the game web site! This si the first "down time" I've had since moving to Dreamhost. It was always a risk, since I'm operating on such a tight budget, and you pay for what you get, but I'll watch this closely.

Tolworthy's first law of game developing:

"If a job doesn't get done in one day, it probably won't get done in two or three."

I set myself tasks for each day, and I have discovered that if something doesn't get done in one day it usually takes a LOT longer. This is because developing an adventure game is potentially very complex. Even a simple task is more complex than it appears. Take something like "give object X to character Y." When playing the game it might take just five seconds., but when designing the game it involves two characters (you and the recipient), several "if...then" statements (is it possible to do that acion where you are, does your character even have the object to give, etc.?), several lines of code (remove object from A, add object to B, change the status of the story, etc.), possibly an animation, two or more lines of dialog, and so on. And then you have to think "what if soemthign unexpected happens" so you have to considerevery possible scenerio (what if the user gave the object to the wrong person, what if they try to do it at the wrong part of the game, what if the user does two actions in an order you don't expect, etc.). Sure, you can force the user to only be able to do certain things in a certain order, but then that's not a very fun game. Anyway, even the simplest task is very complex from a developer's point of view.

Because of this complexity, it is easy for a day's work to become overwhelming. As a developer I may be juggling a hundred variables in my head for that day, and if I don't get every detail tid up in one day it's easy to forget something, and then there will be a tiny error hidden deep in thegame design, which would leas to unwanted events that are not noticed til six months later and then take two days to trace and fix. Which then adds new bugs, since by that time I have forgotten what it is I was trying to achieve and why certain things were done in a certain order. Good grief!

This is a long way of explaining why, instead of continuing on the story, I spent yesterday redesigning how I organize the story code, to make it much easier to understand and follow. I should now be able to complete even the most complex parts of the story in just one day each. Today we shall see, as I try to finish stage two of Les Miserables, where Jean Valjean finds refuge at the bishop's house.

Today I'm adding the conversations where Valjean is allowed to stay with the bishop. At last! A part of the book that most people recognize!

Most of the puzzles in the game are based in some way on the book, but sometimes only the most hard core readers would recognize the references. Like one puzzle is based on an event or character trait that is usually forgotten, but is briefly implied in the book. And a few of the puzzles are just made up, just to confuse people. I'm evil like that. :)

Bug fixing is not my favorite task! Last night, for no apparent reason, when I clicked on 'talk to someone' the character then walked to the complete opposite side of the screen. After hours of searching, I found the reason: I had made a very tiny change in one obscurer part of the code to fix another minor problem, and this change fundamentally altered how everything else worked. Another bug resulted in whole streets with no houses. I just couldn't see what was wrong with the code - it worked perfectly on most streets. It turns out that when I told the program "there is a house here" I added the house to the BEGINNING of a list instead of the END of that list. Most times that didn't make any difference, but sometimes it was enough for the game to think there was already a house where no house really existed. The point is that a very tiny error can go unnoticed for weeks or even months, then suddenly have catastrophic results. This kind of bug can take a whole day away from the real work, so it isn't my favorite thing to discover! But it's fixed now.

On a more positive note, the first part of the game is playable now. It took four days to code and test instead of the planned for two, but the extra time was spent in bug fixing and in writing code and in masking future parts easier to write, so it isn't wasted. And I planned for some extra time for the unexpected, so we're still on schedule people!

Got my first feedback from a real tester today. Her computer didn't blow up and she's still talking to me. Which is a great relief.

Latest schedule:

Late May: main story is playable. Send to composer.

Late May to Late September: late alpha testing, spend two months making suggested changes. And one month adding the non-Les Mis story, and one month on barricade animations and crowd scenes.

Late September: Teaser video on YouTube. First beta testing.

October - November: serious beta testing and polishing.

November - December 15th: send for reviews, finalize download details, create demo version, upload, then hide under the bedcovers with fingers crossed.

Found a funny bug yesterday. When you stand in certain positions, and people walk past you, their heads pass behind you and their bodies pass in front. This bug is harder to fix than it sounds, since many characters have separate heads and bodies, and the walking and positioning code is hard wired into the engine. Besides, I'm not sure that most people will notice, so I may have to leave it. A similar thing happened in early LucasArts adventure games, where it was possible for one character to walk right through another one, like a ghost. The ideal solution is to build in "collision detection" but that would slow the game right down and create new bugs, because the engine would have to check the position of every single character many times a second and work out alternative paths which might result in unwanted routes being taken.

I was a little optimistic in completing the first section in one day, mainly because the first part of anything takes longest - no matter how ready I think it is, there is always custom code to be written, and tested, and tested again, and tested again... :) But the first part is eight tenths complete, so it should be finished this morning. I've allowed 40 days for all 29 parts, not including detailed conversations and new animations, so this is still on schedule.

After those 40 days (which is much faster than I originally expected) I'll spend about a month working on the non-Les Mis story. This means that by the start of May the entire game will be playable! Which leaves plenty of time before September to add in the special stuff. Realistically of course I will put in some of the animations as we go along, but still, the point is that for a while I was feeling highly pressured for time, but now the schedule is looking good.

Another milestone. Today I complete the first playable section of the game: from the beginning to just before Jean Valjean is released from jail.

I have divided the story into 29 sections, and hope to complete one per day. Realistically it may take a little longer, but this means that real testing can begin by the end of March or first week in April. It will also take a month to add a playable story to the non-Les Miserables part of the game, and another month to add major animations such as the barricades scene, but that leaves three months for major changes suggested by testers, and another three months for final polishing and promotion. I am determined that this game will be thoroughly tested and approved by real gamers before it's released for general use.

Only 299 days to go! Which means for me, just 199 days! Since the last 100 days will be spent testing, polishing, promoting and practising. Scary suff. After all these years, just 199 days to go.

Yesterday went well. The whole story now exists in primitive pseudo-code. Specifically, 29 main tasks, each comprising between 1 and 10 sub-tasks. Today I'm going back over it to organize all the surrounding notes ready to be made into functions and conversations. Then tomorrow (or more realistically, Monday) I'll make the first finished task, compile it, and see how it plays.

I'm trying to do this part (the story and puzzles) as quickly as possible, so I can then spend as long as possible adapting it in respose to user feedback.

Today is a very important day. I was doing some experiments last night and I reckon I can plan the skeleton for the entire story in one day. Yes, just one day (I have a day off my regular job.) I have already edited the book's text so I only have the parts that would make a good game, but it is still just text files. I think I can turn it into pseudo-code in just one day. It would then take a few weeks of course to write all the conversations, and add other details. But that would be flexible: once the pseudo-code is written, it can then be converted to real code as quickly or as slowly as you like. (For non-programmers, pseudo-ode is just the logic of the code without the code itself. Typically it is a long series of "if then" statements.

Why does this matter? Because this gets to the heart of what I dreamed of when I first started planing this game, all those years ago: FAST DEVELOPMENT. I dreamed of being able to add a complete story in just a few months. I dreamed of having fifty classic books PLUS the complete works of Shakespeare in one enormous game. And I had a gut feeling that if I stuck to yesterday's technology (simple point and click) using today's tools then I could create very quickly indeed. I figured that novels and adventure games are fundamentally similar, so most of the planning and puzzle design has already been done in the books themselves. Today I'm going to prove whether my theory was right. Now that the background work has been done (characters and backgrounds) I can see if coding the story itself really can be that quick.

The last year or two has been frustrating for me, because I have had to spend so much time on the background stuff without actually adding the stories. That makes me nervous because no matter how much I try, I can't make my graphics or code look as good as a professional game: I am one guy, doing everything in my spare time, with almost zero budget, and competing with large teams of full time professionals. I had to spend that time because otherwise the game simply would not work even at a basic level (except as a text game, and I don't think there is a commercial market for text games). But I knew that I would never create anything wonderful or special that way, the best I could do would be to create something tolerable. But now I am past that stage. The background work is now complete. Now I can concentrate on the area where (hopefully) this game can stand out from the crowd and offer something genuinely new and different and (for people who like stories) better.

Of course, even if I succeed and create the first game very quickly, it will still be just one game. It will not look so different from other amateur games. But as I add more stories, and more and more and more, the real strength of this approach will become apparent. Or not.

Today is the day t find out if my great dream of rapid game development can become a reality.

I'm pleased to announce that today the final coding of the actual story begins in earnest. I find it hard to concentrate on more than one thing at once, so that's why I've been doing the work in chunks of similar tasks, instead of completing one scene all the way then starting from scratch on the next. Just for the record, here is the raod map that takes us up to December:

NOW: adding the story.

May: basic story complete, story is playable from start to finish.

June: add extra touches, especially crowds scenes and the barricades animations.

July: polish and bug fix the story in response to fedback.

August: add the non-Les Miserables story.

September: catch up on things that took longer than expected; release teaser video on YouTube

October: test, polish and bug fix.

November: Add the introduction dialogue that ties everything together. I'm leaving that til last because it has to respond to the whole look and feel of the final game. Send beta test versions to magazines and web sites for review.

December: make shorter demo, release game, promote it. Start work on version 2!

TODAY'S THE DAY! Well not actually THE day, but the biggest milestone from my point of view. I'm currently uploading the first alpha version of the game for testing. This is make or break!

Doing the "final" testing today. Testing is SO boring. Find a minor bug. Change the code. Wait twenty minutes for it to compile. Find the bug is still there. Change the code again. Wait another twenty minutes. Repeat five times then discover it's time to take the kids somewhere as promised. Return three hours later, fix the original bug and find another one. Spend another hour tracking that one down. Get tired and go to bed. One day closer to the final release date!

It's not quite as bad as it sounds. While the game is compiling I work on adding the Les Miserables story code.

A few days ago I wrote that I haven't even started the main story yet. It's not as terrible as it sounds: I already have the events and puzzles planned out, and all the backgrounds and characters are ready, it's just a matter of creating the detailed code that defines exactly who says what and when and what happens if... etc., etc.

Blogger wasn't working when I logged on ths morning (around 8:30 am GMT), so this is later than planned. Just an update to say all the backgrounds are in place, all the characters are in place, and now I'm testing the first version of the game (without the Les Mis story - that comes next) before sending it for the first off-site testing on Wednesday.

The new "walk to" code is much cleaner and neater than before. Which is a good job! In re-writing that code I realized that the old code completely ignored the fact that some characters have seprate heads and bodies. Under the old code, if such a character was walking and you clicked on them to talk to them, the part you clicked on would stop, and the other part would keep on moving! I don't think Madame Guillotine features quite that strongly in Les Miserables. :)

Adding the final non-Les Mis characters. Whenever I think "this will take one hour" it ends up taking two because it has to look right. Looking right is important! And I'll probably re-write the "walk to" code. Right now when you talk to a character you sometimes face the wrong direction. This is because the underlying code is too complicated and too easy to go wrong. So I need to fix that and make it more elegant and robust.

A later than usual blog today. Fingers crossed, all the non-Les Mis characters should be in position by the end of tonight. The Les Mis characters can be added quickly tomorrow. For the alpha version I don't need to worry about where they are, they can all just stand there on one scene. I don't need to worry about their final positions until I get to their part in the main story.

I also hope to have time to add a few final scenes to the introduction section on Sunday, then in the rest of Sunday through Tuesday it will be test, test, test, before sending it to an outsider to see, then starting work on adding the finished story, chapter by chapter.

This will be a major milestone. Finally! If you compare the game to a theatrical production, up until now I've been making business plans, building the theatre, painting scenery, making costumes, writing scripts, working on publicity, etc. This has taken literally years (part tiume and making plenty of mistakes along the way). But finally, within a few short days, the first ever "outsider" will get to look at my "theater" and then - at last! I finally begin "directing."

Well yesterday was far more successful than I expected! Up at 5:30, worked through to 10 pm, with just 3 hours of breaks for babysitting, dog walking and food. Made some real progress. When I only get an hour or two it can be frustratng, but a whole day like that is very satisfying, watching the game come together.

Long term my dream is to be able to give up my regular job and spend all my time on the game. I'm confident that I can produce something pretty special IF (and only if) I can put huge amounts of time into it. We shall see what the future brings.

Things take a long time because unexpected tasks are always cropping up. For example, I have one scene where a man plays the mandolin. I had planned to just splap in the picture and forget it, but it just didn't sound right without mandolin music. And I could not find any mandolin music that was within my zero budget. So I took an hour working with ModPlug, finding out how to change public domain MIDI files so they sound like different insruments. The finished result is not fantastic, but it's perfectly suitable for its purpose.

Today I am determined to get a whole ton of stuff done without interruptions! A day off my regular job, no taxiing, swimming, visits from friends, no forgotten obligations that involve several hours somewhere else (I hope), just work work work. Or at least that's the plan. It's never happened yet, but I can dream... :0

Continuing adding secondary characters. Last night was professor Cavor. Tonight is some aliens.

That's the "inferno" characters in place, today I start installing the "other dimension" characters. It's my daughter's birthday (AND my mother's birthday) today, plus a full day at my regular job, so I probably won't get much added. But all the people should be in place by Sunday, when I send the game for its first ever test.

Remember how I originally planned for the second version to be War and Peace? Then I realized that was too ambitious at this early inexperienced stage, so planned for it to be Shakespeare's Julius Caesar? As the deadline rapidly approaches that plan is being revised again. For deadline reasons the second version will almost certainly be Dante's Inferno, simply because half the work has already been done, but I don't have time to polish it in time for the first release, so I've removed most of those scenes for now, but they can easily be reinstated when I start work on "release 2" in December.

Regarding deadlines, this is all part of the learning process. But my long term plan is the same as ever: crazy as it sounds, in about ten years I want to be so quick at adding content that I can add a new story in maybe three times what it takes to read the book thoroughly. That may sound insane, but most of the development time goes into backgrounds, characters, animations, etc. I have to put some time into these for the first release, or the game will be just too different for people to accept. But long term I do not want to compete on graphics, how could I? I do not have the budget of the big studios. Long term, I want to have a stock collection of characters and scenes, and reuse them for every story.

Long term, I see this as a theater that reuses its actors and scenery. And its selling point is depth, not eye candy. So long term I still plan to add the complete works of Shakespeare.

Yesterday I added and animated the characters for the "fairy tale" section of the game. Today I'm adding the characters for the "Dante's Inferno" section of the game. In the first release there isn't much happening in these sections, but I plan to expand on them in later releases.

The plan is still to have the first pre-alpha version sent out to be tested in about a week, then the cast and scenes and everything else are ready, so it's time to start on the most important part, the Les Miserables story!

Steadily working through the 130 main characters (30 Les Miserables, 100 other, but the Les Mis characters of course do far more), adding them to the game in their starting positions. Kind of like in the Truman Show where everyone is in place ready for direction to start.

That's better! I'm now adding and improving the character code. It still takes a long time, but the good thing is that everything I type will be (potentially) seen on the screen. This is much more satisfying than endlessly tracking bugs in background images.

The funny thing is, programming affects my weight. I'm on a diet, and when programming is fun I don't eat much because I'm too busy on the computer. But when programming is a frustrating then I spend more time in the kitchen!

In other news, I did some overtime last month so have a tiny bit extra money, so I'm buying a 4 gig memory stick, so I can carry the important files around with me for safety. Also, my son is very hap since he'll be getting his Nintendo Wii on Saturday!

It's February already! I need to speed up my progress. Luckily (actually, it was the plan all along) the programing parts are now finished.

I find programing very tedious since you have to understand all that complex code, make tiny changes, recompile (which takes over ten minutes), test, and start again. Fixing a relatively simple problem can easily take a whole day. Then that can easily uncover some other minor problem that takes another day (I have a family and a day job, so "a whole day" can just mean a few hours). It can be very frustrating when something that was "finished" a long time ago returns and takes up a whole week or two. Then a new month begins and I look at the project schedule and get a fright.

That's happened a few times over the last six months, regarding the space code, the countryside code, the caves code, the climbing code, joining up scenes, and all those background things that don't add anything to the story. It is inevitable with any programing, since it is fundamentally and irreducibly complex, so the tiniest detail can have unexpected results in other places.

Fingers crossed, the hardest code is finished and the rest of the work should progress much faster. When I have an evening I want to spend it in adding fun new things, not in fixing things that should already work. Tonight I will be adding the (already built) characters to various scenes, and we shall see if things move along faster than before.

None of this will affect the release date: the game will still be released (fully tested) on December 15th. But the worst case scenario is that I have to spend less time on the good stuff, the conversations, and I don't want to miss that out. In future versions of the game I will only add a few dozen new scenes each time (fewer and fewer new scenes with each new release) and no complex new code, so I can concentrate on the story and dialog.