“Donkey Kong – Conversion and Classic”
The following is a copy of one of the discussion threads archived by Paul Dean (spyhunter007) on his Steve Wiebe Donkey Kong Timeline page. The FunSpot discussion forums no longer exist and there are no cached versions — this copy of the discussion thread seems to be the only remaining public archive. For purposes of research and commentary, it is reproduced here in a more easily referenced format.
Formatting changes:
- basic HTML
- intra-paragraph line breaks removed
- author and date placed on the same line
- the day of the week removed
- anchor tags added
Robert T Mruczek — October 29, 2003 – 09:04 pm
Hello all:
Analysis is virtually complete. Except for additional minute details of differences that I expect to receive shortly from Brian Kuh, we have completed the analysis of “Double Donkey Kong” and what is regarded as classic “Donkey Kong”.
Experts Bill Mitchell and Chris Ayra contributed to the analysis of the tape itself. Fellow referee Brien King managed to discuss the issue with Scott Brasington, one of the people who designed the “Double Donkey Kong” conversion kit.
As per Brien, Scott has agreed to let us post his OPINIONS on the matter below, as well as a few other interesting tidbits he provided. Afterwards, the official TG reply shall follow.
The summary is in MY OPINION these games (DKjr and DK) are identical to the original for purposes of world records. Details (opinions) below.
a) the EPROM speed has no affect as the CPU clock speed is the determining factor. the EPROM just has to be fast enough to keep up with the CPU. If too slow the game probably would not work at all.
b) I have patches in the attract mode sequence, power up sequence, initialization sequence, high score save sequence, etc.. These patches do not execute during game play. There is one patch in the interrupt handler that is executed all the time even during game play. For DDK I took very special attention just for this reason and was able to make the P1+P2 check in the same amount of CPU cycles as the original ISR (it took a clever combination of checking for coin input and p1+p2 at same time).
Game play timing on this game is really determined by the frame interrupt. Which runs at 30 (or 60) interrupts per second (I can’t remember which it is). This HW interrupt is the main timing to drive the game play state machines etc… This is a very common implementation for classic raster games. In the ‘background’ DK/J takes care of dealing with music and a few other misc housekeeping activities. Each interrupt is when the game software checks for players inputs, moves the graphics, checks for points to be awarded etc. So really the hardware has dictated that this occurs for each video frame and the length of how many instructions executed is not critical as long as you don’t overrun your budget in which case I suppose you could lose a frame. Again, that is why I was careful in making the interrupt path of code have the same cycle timing as the original.
c) The sounds that are different are handled by external analog hardware. The game software simply writes to a latch and the external analog hardware produces the sound independent of the game software timing. DDK since it is built on DKjr hardware so the sounds are those of DKjr. The sounds are mapped as follows.
DK pound is mapped to DKjr crash
DK walk is mapped to Dkjr walk
DK jump is mapped to Dkjr jump.None of these affect game play in terms of being able to achieve a high score.
So my opinion is for DK or DKJ you can accept a score on DDK.
That being said, lets talk about a couple other of my kits….
Asteroids: My asteroids HS kit has extended scoring. In order to implement this I had to add extra instructions (and vector instructions) in the normal game play sequence. This extra code is to check to see if extra digits need to be displayed and display them if needed. This causes extra CPU cycles and extra vector engine cycles if the additional digits are displayed. Vector games are a little different in raster games in timings. They don’t really have a frame interrupt at a constant speed, instead they setup a list of vectors to draw, tell the hardware to draw them, then in the background setup the next list of vectors to display. The ‘timing’ is more a function of how fast it can draw all the vectors. The more vectors the slower. the less the faster. I don’t know for sure, but I would guess this can cause very slight differences in the game play from a timing perspective. Not sure if that makes it easier or harder?? Additionally with the extended scoring, when the game rolls to 100K, it continues to execute with hard difficulty (the original game when it rolled back over to 0 would have reverted back to easy mode for the first 30k). So in that sense asteroids HS game is harder, as the player does not get the break of going back to easy every time it is rolled (since it does not roll over).
Multipede: Centipede/Multipede multigame: Since this game is built millipede hardware, I think a millipede score on a multipede would be acceptable for similar reasons as above about how raster games do timings and how the patches were applied. However, for centipede on multipede the patches are more complicated. They are more complicated due to how the trackballs are handled. So the interrupt path for centipede on multipede is longer. I don’t think this affects gameplay, as again, for most raster games gameplay timing is controlled by the frame interrupt. But I can’t be for certain it is easier or harder or same.
Hope that helps. Let me know if you have any more questions.
At this time, the current TG world record will be declassed accordingly to a separate category. We believe the score to be an incredible accomplishment in its own right, and will always treat it as such, however for sake of purity, the reclass is necessary once the new category is added.
Additionally, Steve Wiebe contacted me this weekend and understands the situation and our position, and is in the process of acquiring a classic DK so as to compete on the real deal.
That being said, I trust that the matter is now put to rest to everyone’s satisfaction. I just wanted to tie up the sole loose end and provide closure to the site where the information on the DDK conversion kit originated.
Robert
D_Harris – October 30, 2003 – 12:54 pm
Does that mean a completely different category, or the same with a “flag” denoting the specific(T.G. approved) hack?(Since these are now being tracked, completely separate categories are totally unnecessary in my opinion).
And since T.G. will now track another major hack(outside of Turbo Ms.Pac-man), what about certain Pac-man, and Williams’ game(Robotron, StarGate, ect.) hack/s?
Darren Harris
Staten Island, New York.
Robert T Mruczek — October 30, 2003 – 06:24 pm
Hello Darren:
The platform will still be regarded as “Arcade”, yet the title is not being considered a “Hack”. Rather, it is being considered as a TG-approved variant in the arcade section of the scoreboard and upcoming Book.
The major hacks that TG tracks (Turbo Ms Pac, Super Galaxian, Super Missile Attack, can’t be sure if there are others) are also listed in the arcade section, and are generally regarded as TG-approved “hacks” but are listed in the arcade category for simplicity sake.
Just like “Jungle King” and “Jungle Hunt”, we will treat the two DK scores separate in the arcade category.
The DKJr score, however, will remain as is…there is sufficient evidence to substantiate that there is virtually no difference, including timing, between the two, as the conversion kit is DKJr based to begin with.
Hope that helps.
Robert
mspaeth — Sunday, November 02, 2003 – 07:50 pm
That’s just bizarre.
The only hardware differences between dk and dkjr is addition sprites, and modified analog sounds.
Almost no code changes are needed to make DK run on DKJr hardware, so if the changes needed to add the game switching the DKjr are considered negligible enough to not differentiate between DDK DKJr and real DKJr, the same code changes to DK would seem to be just as negligible.
Robert T Mruczek — Sunday, November 02, 2003 – 11:58 pm
Hello Mark:
Well, after the investigation and timing differences were noted, the prudent course of action was to track the two separately. The differences, unfortunately, are not as glaring as, say, the known ROMsets for “Marble Madness” or “Astro Fighter”, but across Level 22 and 117 stages they do add up.
There were gamers that said the two HAD to be tracked separately, those that said “only if” the differences were significant, and those that said “not at all”…just so you know.
Robert
Rick — November 09, 2003 – 02:11 pm
Robert,
I can understand your interest in researching the differences in DK and DDK but isn’t there some loss of objectivity when one of the experts analyzing the game for you is a current record holder on that title (Billy M)? After looking at the current book of records I see Billy M and Chris Ayra have a relationship that goes back to the original Twin Galaxies arcade so that appears to bring the objectivity level even lower, doesn’t it? You know, buddies will stick together and back each other up.
After reading several postings from mspaeth, it would appear he is very knowledgeable with computer hardware and software and probably would have been a better choice as analyst since he appears to have no stake in the outcome. What did the creator of DDK have to say about timing issues? Just my 2 cents.
Rick
Barry — November 09, 2003 – 05:39 pm
An explanation of how the test was conducted is in order. Two guys sitting on a couch watching a videotape can hardly be called scientific. Rick, it seems the creator’s input you were looking for is in the first posting of this thread. From what he has written, I doubt any human could tell the difference.
Robert T Mruczek — November 11, 2003 – 09:49 pm
Barry:
Thanks for the reply. I agree…human players watching a tape is not absolutely scientific. However, consider that watch has 22+ years of experience at the title, more so than the originating programmers. If anyone can visually spot timing differences, it’s these guys.
Second, we have the extremely detailed opinions of one of the gamer designers which pretty much states that the games, in his opinion, play the same for record and scoring purposes. Third, I have data from a third source, Brian Kuh, who plays this one more so than most gamers and who has recently achieved a kill screen proficiency.
Add up all the empirical data and we at TG believe that we’ve covered all the bases, and with the right sources. For sake of purity, the scores are split.
As for myself, I place a great deal of implicit trust and respect in the integrity of fellow TG Board of Director member, Bill Mitchell, and know that he places what’s good for the gaming community first and foremost in all of his gaming contributions.
Robert
Anonymous — November 12, 2003 – 07:44 am
“..As for myself, I place a great deal of implicit trust and respect in the integrity of fellow TG Board of Director member, Bill Mitchell, and know that he places what’s good for the gaming community first and foremost in all of his gaming contributions.”
The employees of Enron thought the same way.
Richard M. — November 12, 2003 – 08:20 am
Anonymous’ skepticism is understood and felt to a degree. I am willing to trust almost anybody at least a little. But I am unwilling to trust anyone absolutely.
Maybe the games do play the same so that an expert on the “conversion kit” version is an expert on the original. I think the proof of that would be for Steve Wiebe to play the original Donkey Kong to establish the record. His existing game should fall under a separate category.
And I see from Robert’s post opening his thread that Steve is getting his own DK machine for this purpose. 🙂
— Richard M.
Robert T Mruczek — November 12, 2003 – 07:01 pm
Hello all:
The famous “Anonymous” posts yet again. This time, I do respect his/her opinion. No debate from me. However, I have to mention the following, which might be apparent to some but not to all…
-> Regardless of whom you rely on for your scoring information and verification, know that the staff involved is likely small, less than 10 people (if even that many), none of which get compensated for their efforts
-> We do this in addition to full-time jobs and family concerns
-> In the precious few hours left after transit to/from work and what little eating/sleeping we enjoy, we dedicate that to the gaming community
-> It is in those few hours per week, which in some cases is almost a full-time job in itself, that we verify scores, watch performances, write articles, answer gamer inquiries, and tend to media requests
That being said, we do try our best to work with the resources at hand. Unfortunately, we do not have a multi-tiered group that can check the checkers and so forth…it’s just us…we rely on ourselves and fellow gamers for the matters at hand.
Your “Enron” comment, though well understood, is an entirely different matter, as no monies are involved, only integrity. Say what you will about “Arthur Anderson” as well, but there too was monies involved. We are a free service provider for the time being, one which shares (more so than most) results of investigations, extreme details of verifications, and more. Others do not share our enthusiasm for openness and full disclosure.
As for the trust I place in my colleague, know that there are five (5) Board Members, so a certain amount of check-and-balance is in place, for the sake of further maintaining impartiality within decisions affecting the gaming community. I believe that it is only prudent to have a check-and-balance system in place, trust aside, so that different perspectives are always consulted.
Again, your comments are well appreciated and understood, though I do not necessarily agree with the analogy in full.
Robert