Translations/Epopee — различия между версиями

Версия 10:36, 16 ноября 2014 (просмотреть исходный код) Alone (обсуждение | вклад) (→‎13 Sep 97 (todo)) ← Предыдущая правка		Версия 10:45, 16 ноября 2014 (просмотреть исходный код) Alone (обсуждение | вклад) (→‎23 Nov 97 (todo)) Следующая правка →
Строка 503:		Строка 503:
	По прежнему нет защиты (некуда разместить), но будет :)		По прежнему нет защиты (некуда разместить), но будет :)
−	=== 23 Nov 97 (todo) ===	+	=== 23 Nov 97 ===
−	Ruslan Scherbatyuk> Hикто не слышал - уже a'la War-Craft сделан?	+	Ruslan Scherbatyuk> Does anybody know - is a'la War-Craft complete?
−	Выход (в Питере) назначен на 6 декабря. Запасной вариант - неделей позже. Сейчас идёт авторское бета-тестирование. Пока засел на девятом уровне (за людей) - не могу освободить крестьян.	+	Launch date (in Petersburg) is set at 6th of December. Reserve plan is a week after. Now author's beta-testing is held. I stuck at 9th level (for humans) by now - can't free the peasants.
−	Приятно фактически 100%-е отсутствие глюков - пока обошлись без зависаний. Даже страшный глюк с хождением воинов по лесу и домам обезврежен (вроде бы :)	+	I like to see practically 100% bug free game - no hangs so far. Even the evil bug where soldiers walked on top of the forest and buildings was fixed (is seems :)
−	Ruslan Scherbatyuk> А UFO-3 не собираются делать.	+	Ruslan Scherbatyuk> UFO-3 won't be done.
−	Про это никто не слышал...	+	Nobody knows about that...
	=== 26 Nov 97 ===		=== 26 Nov 97 ===

---

Версия 10:45, 16 ноября 2014

Epopee by Vyacheslav Mednonogov (ZX-Format #8[1][2])

Part 1

When we hear sometimes that another cool thing for ZX Spectrum was postponed for unknown time we ask - why? Why authors who've set the launch date themselves couldn't make it in time? It is really so hard to write programs? Just learn assembly language and write one kilobyte after another...

I think now you have rare opportunity to understand this yourself. Here I want to present a diary of making one program, collected from messages to ZX.SPECTRUM conference in FIDO. No comments, in purpose. You should know though that not everything that was said more than year ago is still true. What program is it? Guess yourself :)

12 Jul 96

...to be serious, half a year ago there was a Doom-demo from Kharkov.

Now I'm thinking about this idea too. You can't feed people with games like Durak, Buratino and UFO forever. Waiting for Enlight'96 to see works from my competitors. If I see that I can't do better, I better start something like Warcraft.

17 Jul 96

That's how I see Doom on Spectrum:

1. Screen - 256х64 to 16х64 (width can be adjusted by 16 pixels to make the player comfortable enough with his CPU's frequency).
2. View angle - 60 degrees.
3. Texture size - 64х64, 20 base textures, 20 additional ones (additional textures to be kept in high memory, so they will be drawn slower and used rarely).
4. Maximum scale - 2:1. Two 16K pages for fast scale subroutines from 2:1 to 1:32.
5. Floor and ceiling are at fixed height, not textured. Walls almost certainly should be at 90 degrees to each other. So, it's more Wolf than Doom.
6. Enemies are 32х64 sprites (8 top halves 32х32 for any position and 16 low halves to show moving legs). Masked. I plan to rip them from Doom. Maybe someone tells how? If not, I'll draw them myself.
7. Objects are 32х32 sprites, masked. The same for bullets, projectiles, explosions.

So that's the plan.

One pixel will be drawn in 25 (!) t-states in average. This means the screen, say 128х64, will be 200000 t-states. Add extra work to switch pages, select texture line, move buffer to actual screen etc (around 50 percent overhead as always). My calculations give 12 FPS or so. In short, there's no problem with drawing walls.

There is a problem with adding sprites of characters and objects. Especially how to find quickly, what line of what texture is drawn on given screen vertical line and in what scale. I picked a book on this topic, "Secrets of game programming" published by Piter, and my hair stood on end - four multiplications/divisions per line. There is also a problem how to keep the maze in memory to calculate less on each step, which wall is farther or nearer.

I'll begin, however. Even if it will be very-very slow, the screen width can be always decreased to 16 pixels. And then everything will be very-very fast :))

21 Jul 96

I haven't come up with the story yet, but at a wild guess it will look like this (if we see it as UFO-3):

People gained a lot of technologies after 2017 alien war, they made it to space and built a lot of colonies there (in Solar system or elsewhere). And someday all of this went wrong :(

Player will fly around these colonies (may be 10..1000 of them :) and clear them from evil, saving his ships by the way and destroying enemy ships. In the course of the game he must make scientific investigations, find some access codes, new weapons - I'm used to it :) Finally, he must find the prime cause.

And the cleaning of bases and UFOs will be made with Doom engine. There can be any plot - for example, the remaining aliens have dug in, so you must run though sewerage and underground tunnels :)))))

The main thing is to make it all good-looking and somehow FAST (8 FPS by Western standards, but even 6-7 are OK for me). It won't be boring then.

27 Jul 96

I spent some 3 weeks thinking how to make the Doom. I'm stuck at sprites. If we are close enough, their time to draw (scaling in the first time) is comparable to redrawing the full screen.

We can't keep pre-scaled sprites - even 128K is too little for this. So, I dally off this idea until I obtain sufficient experience or someone helps.

Switch to plan B - UFO-3 à la Warcraft :) I.e. a lot of characters (walking robots this time) fight by themselves, you just give them commands. Maybe I'll add Doom-like part as a sub-level - walking in a maze to find something (without any enemies).

30 Jul 96

Max Iwamoto called me on the phone.

He doesn't recommend start something Doom-like. They drew 128x64 screen for 6 frames (6/50 of a second) - in the best case. 10-12 in the worst. Without any characters! So, down with this Doom :) Until everybody will have turbo mode :))

09 Aug 96

UFO-3. BLACK RAVEN.

Preliminary advertisement.

After many estimates I started making the game. This will be full-scale Warcraft. I'd write the actual Warcraft but, because some G-Soft are making it in St. Petersburg too, I changed the plot. What if G-Soft will finish their version, why we need two Warcrafts? :))

Playfield is 64x64 cells, viewfield is 20x20 CHR$, one cell is 2x2 CHR$, buildings are 4x4 CHR$, the graphics is black and white. And so on...

Launch date unknown (even approximate). I'll try not to lengthen out.

18 Aug 96

Ilya Fomin> Slava, write it, WARCRAFT is RULEZZZ! One of my friends tries this idea for two years but never wrote anything but demo.

There are two big difficulties in the game: ultra-fast (not exactly optimal) routing and computer strategy. Luckily enough, I think myself a big pro in exact two these fields - I know how to deceive people so that everyone will say "how clever is the program" :)

16 Sep 96

Work in progress. Remember, this will be almost WC-1. By the way the "UFO-3" prefix was thrown out, so it will be just "Black Raven" (can't glue aliens to Warcraft :)

I realized at last, how the characters find quite suitable routes so fast. This was the main trouble in the game. I can make it too (or better ;)

Still no idea how to adapt fully mouse-controlled game to keyboard or joystick: the interface will be very different from the original. Just in case, provide yourselves with Kempston mice :)))

28 Sep 96

UFO-2 was a mistake - too many man-hours wasted. Make it simpler, less piling up. If we had a lot of software houses and big competition we'd exert ourselves and tweak the whole game.

Nevertheless, I made more that $200 worth of advance orders for "Black Raven":

Intro and outro demos - $120 - contractor is X-TRADE.

Two tunes - 2х$10 - contractor is Dangerous.

24 colour pictures - 24х20000 roubles - contractor is Denis Barkov.

Of course I could better ask people from Moscow (Scorpiоn Club and KSA) for two latter items, but there is 800 km to Moscow from here :((

08 Nov 96

So how it will look. Viewfield is 24х24 CHR$, monochrome. As Ilya requested there will be background music (although it was not so easy to add having stack-based drawing :)

17 Nov 96

Sprites are drawn with stack-based routine again, and interrupts can occur any time - even when the stack points to sprite image!

Artyom Kharitonov> And when it comes, the sprite writhes as in hell :)))

The thing is, the sprite remains live and untouched.

Briefly, how it's done (in case that the interrupt service routine has its own stack):

Method "A" (suggested by Stalker, for example): keep copies of all sprites in memory, and if an interrupt occurs, look if the stack was on sprite image. If so, restore the broken part of the sprite (2 bytes) from its copy. Drawback: you have to double the memory used for sprites.

Method "B" (mine): write all the sprite drawing procedures so that they read words from sprite image into certain register pair (POP DE for example). In this case, if an interrupt occurs, the two broken bytes are guaranteed to be in DE. The rest is simple. And no copies.

Of course, this is not important for Dizzy-like games - they have so few moving sprites so they can can be drawn ten times in one frame. However this is very useful for Dunes, Dooms, and Warcrafts :)))

22 Nov 96

It seems the little men will be already running tomorrow, without any purpose though :)

Difficulties in Black Raven came from unexpected side. I gave the original tracks from Warcraft to Dangerous to remake them for Spectrum, and he said that this set of noises is not to be remade. Here you are, professional musicians :(

03 Dec 96

At least the version with General Sound (if it will be) will load all the songs and samples third disk that will be fully made and distributed by X-Trade.

19 Dec 96

Sergey Novikov> How's the crow?

Goes by itself :))

Approximate launch date may be given - in February (that means April :). The funny thing is, the most hard parts are initially written in C on PC, then I rewrite them in assembly language (by hand, of course). So it appears faster and more reliable. The next time I plan to write the full game on IBM, then port it to ZX :))

There are a lot of things done already, but that's little for playable demo - they can only walk, not fight. I wonder how fast it will be, because 96 characters can be processed in the same time (however there will be less in average - 15-20 in the first levels to 50-60 in the end).

Scrolling turned out to be fast - a lot faster than it was in UFO1/2, because the graphics is flat and all the moves stop while scrolling (that was not before). The water is animated (switchable, but doesn't affect drawing time).

Routing seems fast too, although far from optimal. However it searches fairly what side of the obstacle to pass around. Average time is no more than 5000-15000 t-states. It won't affect game speed because it is called rarely.

Game map is not always on screen but appears by keystroke. Maybe not that good but all the objects are clearly visible, shown with 3х3 pixel sprites. Of course the redrawing speed is not so high because of this. The user couldn't see this though because the map will redraw once per 4-5 game iterations.

All the buttons at the control panel have their hotkeys for those who has no mouse :) Half of the keyboard is used for this (joking).

The enemies are called kungs, not orcs.

21 Dec 96

Sergey Novikov> ...kangs :) Couldn't you invent something better?

You are late, my friend... You should guess yourself, invent yourself, and mail it to me yourself... Now they are called this way in the announcement :))))

If you had known how the others are called... Ж-[]

ps: kangs aren't bad too, I had to thought of that :))

pps: I worried about the arrow for a long time. The arrow is a challenge because screen update is not synchnized with interrupts. First I wanted to make as in my previous games - to redraw the cursor when screen updates. However the time to recalculate BR game state may vary badly, so the arrow would move with different speed in levels with different complexity (I even added adjustable cursor speed, that's terrible). Recently I found how to draw the cursor on interrupts irrespective to screen redraw at that moment.

And this whole trouble was because of people not having mice :)))

17 Jan 97

By the way, the demo of BR for ZXF #6 is ready - you can take it - following the traditions I kept a lot of bugs in it (joke).

2All: There are no enemies, you can't fight, you can't build/produce/mine, I put around 20 people there what is funny itself (they walk stupidly sometimes but always make it to their target). The map is present, invisible cells can open. Surprisingly all of that isn't so slow - equal to FAST speed in Warcraft 1 (it's even cooler on computers with turbo mode). In short, that's OK for an unplayable demo. Full version is guaranteed to slow down no more than by 5-10%. That's certain that the next game will be "Black Raven - 2" - an attempt to squeeze Warcraft 2 in 128 (maybe 256?) KB.

30 Jan 97

About "Black Raven":

The work is at the final stage - the end is near.

In short, number of characters on screen affects the speed strongly - every character/explosion/projectile etc slows it down by 1-1.5%. So in the given case everything is 1.5-2 times slower (I hope that situation will be rare in actual game :) Characters are selected with CAPS. If you hold it you can select characters within the frame (as in Art Studio). Selected characters are outlined. You can group up to 6 people. Using SYM you can add or remove a character in the group.

Logic calculations grow weakly with adding more characters in game - no more than 1% slowdown for ten characters. I feared mostly because of obstacle rounding procedure (that can be up to 15000 t-states depending of the obstacle size) but the tests have shown that even for such a big group it is called once per 4 frames in average (it looks very funny if you send them all somewhere - like a caravan :))

26 Feb 97

Dmitry Fedotov> I was told that Scorpion has Black Raven game on 2 disks.

No, they just made new catalogue (in colour!!!) of Spectrum games and added an announcement about BR in advance. But in the end there is a postscript - launch date is March to April.

Дмитрий Федотов> for 6 USD - this means it's ready??? Please alert to the situation.

As for the price, I think, in St. Petersburg it will be 30000 roubles exactly.

The last week I was occupied with making animations, and possibly the following week too. According to the plan, the first disk should contain the intro part, that consists of 16 frames of 14х18 CHR$ size. And two final animations of 48 frames of the same size. So that is ZX-multimedia :)) Little more than half of it is ready. I suppose, everybody will enjoy it :)) [I like it very much]

02 Mar 97

The static part of sequence converts to Spectrum slightly different in every frame (maybe you remember Welcome Press #1, it can be seen there) - it needs a lot of correction by hand. So I move the background in most scenes (a lot smoother than the rest though).

I've chosen this frame size to keep 8 frames per page. 14х18 is good to move it to screen with simple LDIRs without "tearing" (I'm not a demo maker, you know :))

05 Mar 97

All the animations, including two final ones, are made (it remained only to stick it all together). The most disgusting is the scene with creeping spiders - it gives the shivers...

01 Apr 97

GS support for BR will be made by Stinger (A. Ageev).

11 Apr 97

Especially nowadays, when user demands to makers arise. This strongly applies to real-time strategy. I don't know what is inside "Dune 2" and "WC" for Spectrum, but Black Raven is written in truly object oriented level, I dare to say :)) When one object acts onto another, this generally does not involve immediate effect but generates a sign of the event. Only when the target object is in turn, it recognizes the received impact and processes it.

This is very handy to debug single subroutines, this sets all the characters in equal conditions, this doesn't overload the program with many actions. However, this is troublesome when you debug interfaces between blocks (that's hard to find the source of the bug).

11 Apr 97

Black Raven - demo v0.02

Demo in .Z80 format.

Music by JAM from Moscow region.

To have some fun, select the Wizard and press "scorpions" button. Then send them somewhere. Remember scattering cockroaches when you switch the light on in the kitchen at night? :)))

19 Apr 97

Dangerous>> The biggest bug in the world: Mednonogov wants to make music for Black Raven as a supplement on 3rd disk!!!

Сергей Hовиков> A novel trick. People will pay for a 3 disk game more than for a 2 disk game.

The game will be distributed on 2 disks. Tunes and sounds are completely up to Dangerous (X-Trade), and he decides how to distribute them (I'd post them in fileecho or BBS with pleasure). I doubt anybody remembers that GS was scheduled even before UFO-2, and they spoke about support even then (!). By the way, we had an agreement with X-Trade about making an elaborate demo for BR and they haven't made it. I don't want to burn my fingers for the third time.

The sound support will load once at the start of the game, and will be called from the game (if X-Trade won't forget their promise to supply bug-free GS support subroutines :))

21 Apr 97

Almost 500 KB on the first disk is taken by video.

Alex Letaev> Packed or not?

476 when packed, 794 when unpacked. However, if there will remain enough space on the second disk, I'll move two final animations there.

21 Apr 97

"Rounding the obstacles" - in UFO-1 it was quite wretched - the hero got stuck at every post. And the aliens simply moved around in pseudorandom directions - ofcourse they came somewhere sometimes :) [Brownian motion effect]

Routing in UFO-2 was described in a big article in ZXF #6 (with example in Basic :) Lee algorithm was used there.

Routing in BR... I can describe it in simple words. A character walks in the line of the target. He bumps into an obstacle, he inspects its outline clockwise and counter-clockwise, finds a spot on the outline which is the closest to the target and decides how to reach it faster - clockwise or counter-clockwise. He rounds the obstacle and continues the movement towards the target. And so with all the obstacles, until he reaches the target.

Theoretically it would be better to remember the route to optimize it (make it smoother), but Spectrum has too few memory. Because of that I store only the direction of bypassing the obstacle and the length of the bypass, one byte is enough for this. Practically this is disturbed by other moving objects. Although it works somehow (see the demo).

08 May 97

IMHO recently the coolest things for ZX are made with the help of IBM/Amiga. Sometimes almost without using actual Spectrum :)) Just with emulators. For example, STEP group does so for a long time (using UKV 1.2)

Indeed, if a hereditary musician LAV writes his tunes using modest sound of Shalaev's emulator, forgetting the delights of real AY.

Alexander Lebedev> Slava, when will you start to code for IBM? :)

The first step is already done - I bought a book "Secrets of game programming for WINDOWS 95" :))

Part 2

16 May 97

Black Raven - the last hurdle.

Work with BR slowed down a little. The cause the need to detect nearby enemies for each character. Especially in the case of computer AI, whose soldiers see quite far (how else you can do an AI? :) I tried to solve this in straightforward way (check all agaist all) but in vain. I hope the task can be solved. I wonder how the developers of Dune 2 did this (or they stuck here too?).

18 May 97

With the rules of Black Raven seeing a big sector is a fair game, because human player is able too see the whole open part of the map. I'd let this to the computer, too, but:

• processing all this information is too much for little Speccy;
• even it I obtain this information I hardly can find decent rules to use it (this reminds me the same problem in Durak where in "card-sharper" mode the computer knows all the cards in the deck and the hands but can't really use this great information :))

On the other hand, to have the game absolutely fair, the computer must be all like human - have its own map of visible space, send scouts to see more. This would make the program a lot "fatter".

Yuri Matveev> If the problem is to find the closest object to fight - use the usual method of spiral scanning until you hit the enemy. Then attack.

This method is not very good too. Even if you limit the radius of the spiral (N), the number of cells to see will be = (2*N+1)^2, that is very long time for computer's soldiers that have quite a big N, while the game is real-time. Even for our soldiers the N is 2..4 in average.

We need to hope that the task can be solved.

Yuri Matveev> Of course :) Look at warcrafts, dunes, and all the commands-and-conquers :)

Well, having 133 MHz, I'd make it in straightforward way and in Basic :))

21 May 97

If I decided right, in a regular game all the enemy soldiers must have visibility range a litle more than the Sight of catapult, for them to react to its approaching in advance. The rule is simple but it should be made - or a few slow catapults of the player will break everything in pieces. That's the whole trouble :((

Sergey Novikov> Well, we wait at the barrack to accumulate 4 units, then go fighting. 9 is even better - the computer will play very strong then - not anyone can deal with 9 units on once!

I haven't got so far, but in WC1 there was a special place (of places) to gather. That looked beter than in WC2, where the squad started from the barrack, but there was a vulnerable point - you can intercept enemy units while they walk from the barrack to the gathering place by one.

Sergey Novikov> Additional mode can be added - for computers with turbo mode ONLY - where the computer will think better, and more effects can be added (animations in game).

Well, just a little more than 4K remained, you can't get much going :(( Not even effects and animations - I think of removing the music (5K, after all). GS effects can also be fired, and game saves, some functions like building fences and roads. Troubles everywhere :(

21 May 97

Unlike UFO-1/2, a direction in BR is purely formal. A unit can turn in 180 degrees at once that is not as realistic. On the other hand, that makes sense because if you stand the units wrong in the line of defense you can lose a lot of health (because of time to turn around). Rotating your soldiers when you stand them in position will be unhandy, and you should look afer them, do they face the right way. Because of that, they have circular sight.

21 May 97

Dmitry Grigoryev> So let them stay and scan. With a ray - you can scan the whole map. After, say, 10 iterations all the directions will be scanned. If they find anybody, let them chase him.

Why 10? Maybe more? :)

I understand the idea - partial scan on every iteration. I think to do this, too, if I won't come with anything better. However I'd better divide all the soldiers in groups and at every iteration I'll work with just one group. This will worsen the reaction though :((

26 May 97

To see the full picture, here's the method suggested by Nickolay Mezencef.

(It was about scanning the map by each character of a real-time strategy, to find a target.)

"Why can't we, instead of scanning cells, check the coordinates of enemy objects. I don't think this will take more than one frame even with the maximum number of objects at both sides. If reaction time is 1 second (I mean the objects act once per second), such checking will take infinitely small resourses."

This poses a question what is better: scan the surrounding cells indeed or check the coordinates of all the enemy objects. The first method was discussed thoroughly, unlike the second. Nickolay proposes the second method with the following optimization:

"This can be optimized even further: if an object moves, it itself informs enemy objects about its coming in their range of view. In this case, if there was no move, the scanning is not needed at all."

Very (IMHO) good solution - all the work in fact will be held with moving characters. A little odd, i.e. not the units standing in defense will scan the area, but moving enemies will signal them about their approaching. It is known that most of time the sides gather their forces and don't move. I can mention just one downside of this method: all these calculations should be done with peasants/peons, who always move. There are not a few of them (usually 15-25% of total population).

Rescanning can be required in some other cases (a character killed an enemy and must search for another victim; enemies moving in attack can additionally scan the area to not miss player's soldiers that move in enemy's rear, although this was not done in WC2). But all of this is really nothing :)

To sensibly close the topic I'll describe the method that I've chosen finally.

All the map is divided in squares (in the case of BR, 8x8 cell squares appeared the most suitable). Each of these squares keeps a list of enemy units that reside there. If a square contains no enemies, the list is empty.

Scanning for each player's soldier is checking 9 squares (8 adjacent squares and his own square). If enemies were found in a square, the coordinates are checked to find if an enemy is close enough.

The same procedure is held with enemy soldiers.

The idea is based on the fact that most of the time all the characters of confronting sides reside very far from each other, so the lists in scanned squares will be mostly empty.

As an alternative, we can use 4x4 squares but in this case all characters having Sight>4 would scan 25 squares instead of 9 :(

A little difficulty is in making lists but this is just a small technical task :))

12 Jun 97

Now I can promise firmly - Black Raven will be no dumber than Warcraft :))

13 Jun 97

Unit descriptor at Spectrum will fit in 20 bytes with ease (even 19), unlike, for example, UFO-2, where 32 bytes were needed. 2K given is enough for 102 characters.

Trying to copy WC2 seems to be indeed hard though, because for 128х128 map (instead of 64х64 in WC1) 32К is needed (with 2 bytes per cell).

In this case the map + characters data + bullet/spell data + buildings data + most of variables + 6К shadow screen must always reside in the memory accessible from all the pages, i.e. #4000 to #C000, that is just 32К :((

So, if WC2 will be on Spectrum, it will have smaller map (80х80 or 128х64) and work on 256K computers only (even without flying characters unless cell descriptor will be 3 bytes instead of 2).

21 Jun 97

Well, the memory problem seems solved (yet :). Sprites of three less used characters were cut down (this means their sprites facing right are made from left-facing sprites in realtime. This was done using tables in ~3000 t-states) - demon, walleater, bonebreaker. This gave additional 2.5K - this must be enough for the most sophisticated AI :))

It seems the game will have 3 levels of difficulty:

• normal (as in WC1), without a couple or three last levels;
• hard (more clever enemy + harder rules);
• very hard (the same as 2 but without saves in game).

The main rating of player's acting will be the time of passing the levels and the game in total - IMHO that's fair.

I haven't got close to the AI yet but I managed to clear almost all the bugs and finish everything unfinished :) This includes the system of alarm indicators (at bottom right corner) that signal the attack, finishing the building/production, destroying buildings etc.

24 Jun 97

Spectrum graphics takes 8 times less than on PC; 2-4 times less phases showing movement or attack, less landscape sprites - in short, everything fits in ~80K.

Misha Turnovsky> To have a normal strategy you need 1000-1500 characters IMHO.

That's too many. In WC1, a number of units on human side is usually no more than 30-35, on computer's side - 50-60 (in final levels). In WC2 this number is 1.5-2 times bigger (among other things, this is because of that computer's game is based on defense). Spectrum can do this easily (as measured at 3.5 MHz, it is between FAST and FASTER in WC1, and at 7 MHz - well... :))

26 Jun 97

Denis Parinov> How many levels there will be?

For now I think of 12 for either side (12+12). Although I suspect for some reason that there will be more. There are four landscapes so far - winter forest, regular forest, empty land, and underground (for special levels if you remember WC1). Maybe there will be two more - mountains (instead of rivers) and black forest (black land, fire rivers). In BR-2 (if it will be) I plan more landscapes: city, a port with ships.

01 Jul 97

Today I finished the last spell - enemy tornadoes, looking more like nazi swastikas - couldn't invent anything better :)

01 Jul 97

The biggest problem for Speccy in the future will be the absence of empty 5,25" disks in sale!!! This is even more frightening than evil hackers :))

01 Jul 97

In BR the computer will act according to the following plan:

• units will be divided in "constant" and "additional" staff;
• constant staff - units that were on the map in the beginning - guards and peasants. Peasants are divided in advance into those who mine gold or forest and maybe repairers. New peasants appear only in case of their death. Guards won't restore if dead (under normal difficulty), or will restore (hard difficulty);
• additional staff appears in the course of the game (when resources are collected or certain time passed or ...) and form attack squads. After gathering a certain number of additional units all they set off to the campaign (the target is usually "town-hall" but not always).

Each enemy unit has a special field in its desriptor that shows what he does (for example, in case of wizard it shows the type of spell he specializes in). Another field shows what he is actually doing now. The computer doesn't process this unit until this field is 0 (not 0 means the task is in process). Once the task is done or interrupted by external forces, the computer tries to "find job" for the unit taking into account his professional abilities. The guards are an exception (they find their job themselves, scanning around), as well as additional units until they gathered in number sufficient for a campaign.

Something like this. Variants are possible though...

10 Aug 97

John Stunner> How's the BR?

Lives his mysterious life. You know, it's summer, there is 25°С in Petersburg :) So, the last month it goes somehow weak :(

15 Aug 97

Black Raven - Demo v0.04

Well, that's it, in .Z80 format. As planned, this is the last demo version, only the full game can follow...

One hypothetical level, not so complicated (no magic, no catapults; enemies are not so clever, not idiots though). Allegedly passable. If you win, shows "Well Done", or "R.I.P." otherwise.

That's symbolistic, I started working on BR precisely one year ago...

Special for Еnlight'97! ;)

P.S.: Warning, if you play without mouse, you'll have hand paralysis in 30 minutes! Ж-)

30 Aug 97

Alex Koval'> I made it! Yes! I got it through! I finished that demo!!! In the end there is WELL DONE in attibutes.

Exactly!!!

Alex Koval'> Cool! If it will the the last level you can keep the logic.

In fact, it was the first ;)

My brother couldn't finish it :( While he said he'd be my beta tester. Here they are, PC Warcraft lovers :))

04 Sep 97 (todo)

About BR: в последние дни была проведена капитальная доработка трассировщика согласно сообщённым глюкам:

• воины противной стороны при трассировке считаются проходимыми, поэтому, если перегородить мост или другое узкое место своими солдатами, оно не будет считаться непроходимым для врага (как в деме);
• проходимыми считаются теперь и движущиеся воины, что даёт возможность проходить большими группами через узкие проходы без разбегания отряда. С помощью дополнительных ухищрений удалось достичь 100% проходимости отряда;
• исключены закручивания (когда воин долго ходит вокруг пустого места);
• гораздо лучше теперь обстоит дело с выходом из узких тупиковых проходов.

Кроме того, по просьбам трудящихся:

• чтобы атаковать своего воина, надо нажать специальную клавишу. Таким образом, исключена возможность случайно убить своего воина в пылу борьбы. Введена фиксация цели - выбрав цель, нельзя случайно выбрать другую до отпускания SPACE. Захват цели индицируется миганием курсора;
• запрещено строить здания на каменных дорожках и тропинках, чтобы нельзя было построить непроходимую преграду из зданий (как это возможно в деме). То же относится и к заборам (в уровнях, где возможно их строить);
• другие мелочи.

Видимо, к конце недели будет завершена стратегическая часть (наконец-то ;) и начнётся работа над уровнями.

11 Sep 97

Denis Dmitriev> I don't understand why the group is no more than 6 soldiers?

I don't know myself... However, if you select a group and click on "G" icon, on the shield below you'll see all the members of the group shown - without parameters though.

13 Sep 97

Paul Falcon> How it will go between levels?

Once you completed the mission (usually "kill them all"), you progress to the next level.

P.S.: I just remembered that graphic artist (D. Barkov) supplied 10 level pictures only and ran away to PC :( I had to made 22 more in fast pace (then it became clear than there will be 16+16 levels in total :)

Following the requests from the users I made all the graphics in colour.

Interesting thing was that scanning then colouring appeared less productive than converting in colour (with ZXC 1.1), then correcting by hand (correctinf takes 1.5 to 2.5 hours in average). All the work was done in five (!) days.

IMHO this shows that there won't be any problem with quality graphics on Spectrum - practically any picture can be converted in 4-5 colours very well (and fast).

13 Oct 97 (todo)

Тут выяснилось, что ЧВ включён в новый каталог of Logros/Nemo (а в каталоге фирмы Scorpion он уже с апреля :), и что люди активно шлют деньги. Хоть я и не давал твёрдых обещаний, но на всякий случай спешу успокоить покупателей насчёт вложенных денег - всё будет о.к., причём относительно скоро ;)

К концу этой недели будут завершены все 32 уровня. Что останется:

• к каждому уровню надо задать массу параметров;
• убитые крестьяне пока не заменяются новыми;
• хотя все заклинания готовы, маги и колдуны компьютера не умеют пока их применять;
• всё ещё нет работы с диском;
• нет защиты (никак не решу, на каком способе остановится, всё больше мне нравится метод, применяемый в Turn и ZXFormat);
• Larkov must return me my Pentagon that stays since Enlight;
• вроде, всё!

В завершении я её немного потестирую - на предмет глюков, конечно, проходить не собираюсь - задолбало за год :) Кстати, проходимость некоторых уровней под сомнением, но продвинутый геймер наверняка справится.

Hа вторую часть (миссии на стороне кунгов), наверное, навешу пароль, который сообщу через месяц/полтора после начала продажи.

15 Nov 97 (todo)

Мы с Растером вызвали друг друга на соц. соревнование на предмет: "Кто раньше выйдет - ZXF #7 или ЧВ?" Проигравший отдаёт победителю 1 (одну) банку пы-ыва :)

ps: c ЧВ финишные проблемы - глобальная нехватка памяти :((( День за днём идёт битва за каждый байт. Hе в смысле оптимизации, а в смысле урезания спрайтов и выбрасывания кусков кода.

Зато уже есть вариант под TR-DOS (в формате FDI) с несколькими уровнями, Finаl-Cut'ами и возможностью отгрузок. Отгрузки получились очень большими (около 20 Кб), поэтому будут вестись на отдельный TR-DOS диск. Предусмотрена работа с двумя дисководами (спасибо эмулятору UKV), причём можно выбирать, какой дисковод подо что использовать. Второй диск - всегда 5.25" (чтобы поцарапать его :), а первый - 3.5/5.25 на вкус покупателя. Как и отгрузочный (тоже любой). Думаю, это нормально, т.к. проведённые опросы не выявили людей, пользующихся только 3.5- дисководами.

Как и говорилось, по тех. причинам поддержка GS не была включена в игру, в сетапе поэтому появилась пара новых пунктов - "цветная/нецветная игра" и "выбор диска для отгрузок"

Сразу после загрузки начинается бесконечный демо-режим для тех трейдеров, кто, по методе Логроса, держит на торговой точке рабочий компьютер. Сам могу смотреть часами, как в деме дерутся :))

Формат второго диска (где уровни) продуман в этот раз таким образом, что он абсолютно не зависит от 1-ого. Поэтому, вполне вероятно появление в дальнейшем дисков с доп. миссиями. Кроме новых уровней заранее предусмотрена возможность изменить графическое оформление ландшафтов, кнопок, музыкальное сопровождение и межуровнёвую графику. Может быть, за выпуск новых миссий возьмутся наши так называемые хакеры, однако к подобной возможности следует относится скептически :))

По прежнему нет защиты (некуда разместить), но будет :)

23 Nov 97

Ruslan Scherbatyuk> Does anybody know - is a'la War-Craft complete?

Launch date (in Petersburg) is set at 6th of December. Reserve plan is a week after. Now author's beta-testing is held. I stuck at 9th level (for humans) by now - can't free the peasants.

I like to see practically 100% bug free game - no hangs so far. Even the evil bug where soldiers walked on top of the forest and buildings was fixed (is seems :)

Ruslan Scherbatyuk> UFO-3 won't be done.

Nobody knows about that...

26 Nov 97

The protection is not added yet, but 26 levels out of 32 are tested.

Twelve of them I passed fairly, the rest seem to be passable too. The most hard level (yet) - "Humans #13". I even wanted to remove it but I suddenly finished it in 50th try :-) My brother can testify.

05 Dec 97 (todo)

Black Raven won't launch tomorrow!

Hу, во-первых, потому что вышел ZX-Format 7. Грех не почитать свежий номер.

Во-вторых, возможность этого не исключалась.

В-третьих, просто не получилось :( Hеустойчиво читает/пишет отгрузки, защита глючит.

Так всегда бывает, когда что-то из кусков лепится в одну кучу...

Значит - выход назначается окончательно на 13-е декабря (ну и число, однако. Хорошо, хоть не пятница :)

В связи с этим - о распространении по другим городам:

Распространение начнётся неделей позже, чтобы успеть обкатать игру в Питере. Как всегда, этим занимаются Логрос, Зонов, Hемо.

Со мной можно связаться по тел: (812)- 143-33-03.

12 Dec 97 (todo)

Black Raven - v1.00.

Как и предполагалось, завтра первые 30 комплектов поступят в продажу. Это, скорее, бета версия. Уровни за кунгов - под паролем.

В течении недели надеюсь получить сообщения о возможных глюках, чтобы при необходимости доработать программу. Поэтому окончательная версия (1.01) выйдет через неделю. Как минимум, по причине того, что не удалось закончить описание и подготовить наклейки/вкладыши/конверты.

P.S. Может быть, стоило отложить выход ещё на недельку, но - раз обещал... :-)

13 Dec 97 (04:54 a.m.)

Alex Petrovsky> When DOOM and Black Raven will be out - who knows, people?

I don't know about Doom but Black Raven v1.0 already exists for 10 minutes in 34 units!!!

17 Dec 97 (todo)

За всё время эксплуатации найден пока один глюк сборки - невидимые поля после отгрузки из чёрных превращаются в белые (лечится входом в setup - пофиксено). Была ложная тревога от члена Омега-групп про неработоспособность лесопилки.

В целом реал-тайм стратегия для спектрумистов, конечно, в новинку. Самые продвинутые добрались только до пятого уровня (и то, подозреваю, добавив себе денег - а на пятом уровне деньги не помогают ;)

Игра не пошла на новом мегабайтном КАЙ'е (что поделать, ну не везёт мне на КАЙ и всё-тут). Глюк будет пофиксен.

По просьбам трудящихся пароль со второй части игры будет убран.

По просьбе Сергея Зонова будет вариант с двумя дисками под 3,5" - как только будут проведены научные изыскания по методам их царапанья :)

Соответственно, релизная версия 1.01 выйдет в эту субботу (20-ого декабря).

22 Dec 97 (todo)

Вышла. Работоспособность гарантирована. Пожалуй, один глюк пока остался - если вместо отгрузки игры дать игре "левую" информацию - может повиснуть. Мораль: не надо этого делать. Когда-нибудь поправлю...

Удивительно, что, хотя внутри программа представляет собой "маленький Виндоуз'95", работает без сбоев как часы :)

А в 1.00 грамотные люди (например, Саша из Omega group) дошли аж до 9-ого уровня (!), где пока и сидят. Причём по честному (!), без добавки денег (т.к. если там добавить всего 100 монет - уровень проходится без труда)

По просьбе трудящихся вставлена защита от Теневого Монитора. Защищает она или нет, не могу сказать (не имею Скорпиона), но зато ни Пентагон, ни Скорпион не вешаются при её работе :)

Первые 24 комплекта уехали в Ростов-на-Дону.

26 Dec 97

The first welcome news:

Dmitry from G-Soft finished all the 16 levels for humans (!) in v1.01.

In his opinion, the most hard were 15th and 16th levels because of demons and far away goldmines. He enjoyed the final animation :)

G-Soft's success is unsurprising, because they once made Warcraft demo published in ZX-Format, so they got into BR better than others.

01 Jan 98 (todo)

Владимир Спашко> Чёрные Вороны доехали ноpмально и начали свой путь в наpод :-) Писать о впечатлениях нет смысла- это надо видеть!!! Пеpвые успехи одного из фанатов - 9 уpовней за людей за двое суток с пеpеpывами на сон и еду.

Классно!

Сорри, что к тебе ушли версии, где нельзя отгружаться на диск "Б:" :((( Если кого в Ростове-на-Дону это сильно достаёт, потом перепишешь им первый диск теледиском или макдональдом (или я обменяю).

Кстати, про версии:

1.00 - тестовая бета-версия. Hа всякий случай я сделал её непроходимой (а то опять хакеры поломали бы и все свои глюки на меня свалили бы :))). Поразительно, что их обменяли всего шесть штук (хотя в инструкции явно сказано о необходимости обмена). Двое дошли аж до 12 уровня - и обломались :((

1.01 - релизный вариант, полностью проходимый (если уметь играть ;)

Поправлено всё, о чём сообщили пользователи.

1.01A - а) исправлен глюк с запретом отгрузок на "Б:" (просто долго никто не замечал); б) Талабрук случайно был назван Варлоком (пофиксено).

1.01B - а) иногда можно было строить заборы там, где это нельзя (пофиксено); б) G-Soft нашёл глюк, облегчающий разработку заклинаний (пофиксено).

Все 1.01х помечены как 1.01 (номер версии - в описании). Первые и вторые диски от любых версий 1.01х совместимы между собой. Однако отгрузки - не совместимы (в связи с тем, что в отгрузку входят физические адреса некоторых подпрограмм). Других версий не предвидится.

Игра за кунгов: пока не поступало информации о прошедших её дальше второго уровня. Впрочем, это старая информация. Уровни за кунгов отличаются, на мой взгляд, большим разнообразием - больше мест, где надо подумать.

02 Jan 98

When everyone celebrated New Year, guys from G-Soft played for kungs, four together :) They got to 9th level yet. Hard, they say. Level 7 ("islands") seemed the most hard, although their method of solving it is not the best IMHO. I think the following levels are tough too:

• 9th ("joining forces");
• 12th (the toughest of underground levels);
• 16th (maybe impassable).

Advanced gamer will manage it, though!

That's useful to see that every game faces quite hard problems that sometimes make the author change the plot!

Compared to this, a seasoned gamemaker cope with tasks like making a videoclip, converting graphics and making sound effects with ease, without slowing down, even if he never did this before :)

Always yours, Vyacheslav Mednonogov.