Duck! Me? – Development & Beta Testing

[EdwardianDuck]

Just a quick note to say that Beta 4 should be out later today. I'd avoid Beta 3 for now as the issues introduced by EXAMINE {Container} will almost certainly make it impossible to complete.

Jeremy

[ash73]

Finally sussed out the zvpebjnir, and more importantly it's opened my mind to the possibility... nabgure qrivpr zvtug or qbvat fbzrguvat fvzvyne, jura vg nccrnef abg gb or shapgvbavat!

[EdwardianDuck]

I've uploaded Beta 4 to the top post. This should fix all issues reported to date, plus a couple I found, one of which was so hideous I'm genuinely astounded that the game didn't just crash.

Regarding the "OK" confirmation message, the intention is that this only appears if there is no other response from the game. For example, when you LIGHT TORCH, you get a custom response and no "OK", but when you GET TORCH you just get the "OK".

I didn't get that letter from the Ministry of Adventure Games, perhaps they sent it second class due to budget cuts.

Jeremy

[ash73]

I've sent a message to the Ministry of Unfair Adventure Games and retracted my complaint, it would have helped if I tried the dial on 1 before spinal tap level 11.

Everyone needs to play this game, it's fab.

[ash73]

Making good progress now, I'm fixed!

Latest version is looking good, haven't spotted any new bugs.

Lots of positives, original, imaginative, tough puzzles, amusing in places... one of the best adventure games I've played recently. Thanks so much for taking the time to write it, and sharing it with us!

[EdwardianDuck]

Progress update.

I added a naive compression algorithm to the text, reducing ~37Kb to ~25Kb at a cost of about 2.5Kb for the dictionary (which unfortunately must exist in main RAM) and additional code, so I'm getting close to 30% reduction with negligible performance degradation.

Beta 5 to follow later in the week, after I've done some more testing.

Jeremy

[ash73]

Finshed now, nice ending. I've gone back and edited my posts where I mentioned things I thought were bugs, but were clues. Don't want to spoil it for others.

One possible bug towards the end, the enqvb puzzle is a bit fussy about the order you do things.

[EdwardianDuck]

I'm glad you managed to solve it and enjoyed it. Yes, the enqvb is a bug in the sense that I hadn't thought of ghavat vg va orsber fjvgpuvat vg ba myself, but I'd spotted that one yesterday and it's fixed in the pending Beta 5.

Thank you for all the bug reports and suggestions, these will certainly benefit future games using the same skeleton code.

Hopefully one or two more people will have a got at it.

Jeremy

[melchett]

Not sure if this is expected behaviour or not in Beta 4 but....

<rot13>
Vs V chg gur jebat vgrz va gur zvpebjnir naq ghea vg ba (r.t. gur one bs fbnc), gura vg qvfnccrnef naq zntvpnyyl genafcbegf vgfrys gb gur Ynaqvat Onl ybpngvba. Vf gung rkcrpgrq?
<rot13>

[EdwardianDuck]

Yes...

<rot13>
Guvf vf rkcrpgrq naq lbh'ir zbfgyl svtherq vg bhg. Va orgn 5 juvpu V'yy eryrnfr bapr V'ir svanyvfrq gur grkg pbzcerffvba, gur rknzvar zrffntr vf punatrq gb fhttrfg gung fbzrbar unf gnxra vg ncneg ng fbzr gvzr, nf n fhogyr uvag gung vg'f orra gnzcrerq jvgu.
</rot13>

Jeremy

[melchett]

Oh great, glad to hear that - wanted confirmation before I spent time using that as a "real" feature.

[EdwardianDuck]

I've uploaded Beta 5 to the top post. It's been a while since Beta 4, but I've had a few issues arising from the text compression introduced in this release, as well as some other strange behaviour.

Changes.

• Two objects near the start have had their initial locations swapped over. This doesn't change any puzzles, but the locations are now slightly more sensible.

• Gur enqvb abj jbexf jurgure lbh ghar vg orsber be nsgre gheavat vg ba.

• Gur zvpebjnir'f rknzvar grkg abj vapyhqrf n tragyr uvag.

• Gheavat gur gbepu ba naq bss trarengrf zrffntrf juvpu ner nyfb fyvtug pyhrf.

• Gur tnzr abj fgbcf lbh sebz thrffvat naq hfvat gur pbqrf. Lbh unir gb qvfpbire gurz. Guvf vf gb nibvq fxvccvat fbzr bs gur chmmyrf.

• Ybbxvat va n pbhcyr bs pbagnvaref abj gevttref gur fnzr erirny ybtvp nf rknzvar qvq orsber.

• Fixed a bug arising from my lack of understanding how CA65 handled switching between segments. Guvf qvq abg vzcnpg gur tnzr, ohg bar bs gur enaqbz fprarf qvfcynlrq ba gur ZbaxrlFcnpr zbavgbe jbhyq arire fubj hc.

• Fixed a few typos.

• Backed out all the tailcall optimisations as this suddenly started causing a problem for no apparent reason.

• New loader allows you to start the game in 40 or 80 columns.

Please report any oddities you find.

Jeremy

[EdwardianDuck]

There will be a Beta 6, I just spotted at least one more typo...

Jeremy

[ash73]

Possible bug in the box room. The game crashed when I picked up the gloves and wore them, second time (going straight there) it said it's too dark to see when I put them on... oddly since then it's been ok.

(I'm using BeebEm, 80 cols)

ETA - crashed again when I returned there again later in the game, this time just typing INV.

[melchett]

I think I found a few bugs in Beta 4, I will try on Beta 5 but I haven't downloaded it yet and only just noticed it is out. It might have saved me some time looking at your fix list

<rot13>
1. Vs lbh srrq gur ubg ebpx gb gur fanxr - lbh pna fgvyy tb onpx gb gur fgernz naq rknzvar vg naq svaq nabgure vevqrfprag ebpx. Gur ebpx vf nyfb gbb ubg gb cvpx hc gur frpbaq gvzr.

2. Vs lbh cerff gur xrl onpx vagb n srqben ung, vg yrnirf gur xrl va lbhe vairagbel gbb.

3. Frrzf lbh unir svkrq gur enqvb oht va Orgn 5 (abg gevrq) - V nyjnlf frg gur qvny svefg, gura fjvgpurq ba - frrzf V unq gb qb vg gur bgure jnl nebhaq.

4. Vs lbh nyernql xabj gur xrlcnq cnffjbeq, lbh pna bcra gur fnsr dhvgr rneyl ba jvguva gur tnzr naq qba'g unir gb fbyir gur bgure chmmyrf. Fubhyq or n purpx gb rafher lbh unir pbzcyrgrq gur chmmyr gb trg gur cnffjbeq?
</rot13>

I liked this a great deal - well written, beautifully metaphysical and the craziest adventure I think I have every played. Even the duck portraits add welcome atmosphere to the game. And some good, strong puzzles although see a few comments further below in rot13. Some other feedback:

1. The shortcuts for LOOK and EXAMINE are inspired - I wish this had been a standard back in the day and I don't remember another adventure that did that (or maybe I just never tried it).
2. Marking which directions you have been by using the room name is helpful, especially when you are starting out.
3. Having the options to choose between 40 columns (nostalgia) versus 80 columns (better for my modern eyes these days) was a benefit
4. RAMSAVE and RAMLOAD - kept me sane when going through this, quick feature to go back to a few moves ago when I last saved
5. VERBS helped me solve some of the puzzles without having to wonder "what verb did the author want me to use to do x" - that was painful on adventures of the day
6. It was good to see more complex parsing beyond <verb> <object> - made it a little fun
7. So many red herrings, appreciated and also add to the game, at least I think they are red herrings or I have missed a great deal of the game! Pretty sure they are though...

<rot13>
n. V qba'g haqrefgnaq gur fanxr sbbq chmmyr - V znl unir zvffrq n pyhr nqzvggrqyl ubjrire V bayl qvfpbirerq guvf ol gevny naq reebe. Gur bhgpbzr vf "urnil qvaare" naq gung cebonoyl gura znxrf zber frafr ohg orsber lbh ernq gung, V qba'g erzrzore n pyhr bgure guna vg favssvat ng jung V fubjrq vg.

o. V nccerpvngr V'z n qhpx, jnxvat hc va n jbeyq bs qhpxf naq bgure penmvarff - ohg jul qb zvpebjnir vgrzf nccrne ryfrjurer naq nyfb gur bgure znpuvar'f vgrz? Gbbx zr n juvyr gb svther bhg gur ynggre naq vg jnf znavsrfgvat vgfrys ryfrjurer. Jnf gurer n pyhr fbzrjurer gb fhttrfg gung (bx bx bx V'z n qhpx...)

p. V jnfa'g 100% pregnva V unq svavfurq pbeerpgyl ng gur raq (nsgre frggvat bss gur sverjbex) be vs V unq qvrq va n fvavfgre jnl. V gubhtug vg zhfg or gur raq tvira gur yratgu bs qrfpevcgvba.
<rot13>

And thank you - kept my interest, was sufficiently difficult and great concept!

Andy

[ash73]

o. V nccerpvngr V'z n qhpx, jnxvat hc va n jbeyq bs qhpxf naq bgure penmvarff - ohg jul qb zvpebjnir vgrzf nccrne ryfrjurer naq nyfb gur bgure znpuvar'f vgrz? Gbbx zr n juvyr gb svther bhg gur ynggre naq vg jnf znavsrfgvat vgfrys ryfrjurer. Jnf gurer n pyhr fbzrjurer gb fhttrfg gung (bx bx bx V'z n qhpx...)

V guvax gur zvpebjnir vf vagraqrq nf n pyhr gb bcra lbhe zvaq gb gur cbffvovyvgl nabgure qrivpr zvtug qb fbzrguvat fvzvyne, engure guna fgnlvat fgngvp gelvat gb svther vg bhg. Whfg rabhtu bs n pyhr, naq vg'f irel pyrire vaqrrq.

[EdwardianDuck]

Thank you both for continuing to test and put up with / report bugs!

@ash73

The crash bug has been fixed (wings crossed) in Beta 6 which has been uploaded to the top post. This has taken all day to fix, but has had the benefit of causing me to learn how to use the debugger in B2. If anyone is interested, my sentence aliasing routine which hasn't been changed since the initial preview months ago at an ABUG meeting and is marked in the source code as "tested" was running off the end of its data table and causing havoc. I'm amazed the problem hadn't surfaced before. This manifested itself as INVENTORY being classified as a direction.

@melchett

1 & 2 are fixed in Beta 6, well spotted. 3 & 4 were fixed in Beta 5, if you could get that far without it crashing!

• a. Gur fanxr favssvat gur ebpx vf gur bayl pyhr.

• b. Sebz Orgn 5 gurer vf n uvag va gur zvpebjnir'f qrfpevcgvba gung fbzrbar unf gnxra vg ncneg va gur cnfg. Jung n penpx grnz bs nivna ratvarref zvtug trg hc gb jvgu n zvpebjnir, fbzr Zrppnab naq byq enqvb inyirf vf yrsg gb gur cynlre'f vzntvangvba. Jung @ash73 fnvq, cyhf gur znpuvar vf qrfpevorq nf rkcrevzragny, gurer orvat ab pyrne cerprqrapr nf gb jung n pryrel zngrevnyvfngvba znpuvar "abeznyyl" qbrf.

• c. V jnfa'g fher nobhg gung, fb V'ir nqqrq n pbatenghyngbel zrffntr naq ghearq gur "fpber" fjvgpu ba.

Personally I don't want a game to be too easy.

Some are red herrings, some are present just to amuse a friend from work and some are just things you can play with for fun...

Unir lbh frra gur fzhes, jngpurq gur ivqrbf naq sbhaq gur pryrel neg?

Jeremy

[EdwardianDuck]

Also, it's just possible I'm foreshadowing...

[melchett]

Well that's quite some service and turnaround, thank you. Will try Beta 6 sometime on Sunday - benefit of lockdown is I have a little more idle time than usual. Agree it shouldn't be too easy and it's the right balance right now. Wanted to be sure I followed the logic and hadn't missed anything else. Will check the end message too.

[ash73]

Yes I love the zbaxrlfcnpr npgvivgl zbavgbe very amusing.

No more crashes, one small bug - vs lbh qebc n pregnva vgrz lbh pna urne vg cvat sebz nal ybpngvba

If you fancy a challenge... the already good parser could be enhanced with lists, e.g. take x, y & drop x, y.

But it already gets 10/10 from me.

[EdwardianDuck]

I've uploaded Beta 7 to the top post. This fixes:

• Gur "cvat" oht. Bs pbhefr V pbhyq unir purngrq naq fvzcyl nffregrq gung gbl qhpxf bevtvangvat sebz QhpxFcnpr ner erabjarq sbe gurve cnegvphyneyl nphgr urnevat, ohg gung jbhyq or purngvat (naq jbhyq orpbzr pnaba).

• A more general problem with timer resets which arose due to over zealous moving of data into sideways RAM coupled with a cavalier attitude to bypassing my own API to fiddle about with them. In terms of game play this has no real impact, you'd have just missed the rgurerny ibvpr anttvat lbh gb fbyir gur chmmyr vs lbh erznva va bar ybpngvba gbb ybat message.

Regarding lists, that would involve throwing most of the parser and command processor away and starting again, so, sorry, I'm not adding that. The way the parser is written, even GET ALL would be an ugly hack, unless the game included an allosaurus.

Jeremy

[leenew]

When I added GET ALL to SIGMA EXPERIMENT, I ended up taking it out again.
Each object must be checked for consequences as you take it.
e.g GET ACID may kill you, but GET GLOVES then GET ACID might not.
I found it too fiddly, so disabled it.
The code is still there I think.

Lee.

[EdwardianDuck]

I concur. Given the resource limitations of 8-bit machines, if we have to drop "fancy" in order for the parser and game logic to be robust, that's what I'd recommend every time.

Jeremy

[EdwardianDuck]

No new beta today, just a note to report that I played it through on a real Master 128 for the first time yesterday evening.

As my Master 128 has a DataCentre fitted, saving game position to USB works, which is useful (just specify :5.NAME) when prompted.

Jeremy

[EdwardianDuck]

Roughly how many locations, objects and puzzles by the way?

I guess it depends on how complex something needs to be to be classed as a puzzle, but my checklist of actions you need to complete in order to finish the game comes to 15, one of which is optional, if that helps.

So I guess that makes it quite a small game.

Jeremy

[EdwardianDuck]

I've uploaded Beta 8 to the top post.

It's been a while. I've been working on an application in C# to manage the game data and text compression.

This is mostly a maintenance release as I've had to fix a few minor errors which I spotted while writing the C# application. The text compression makes entering the text directly via CA65 macros too hard to attempt. Text is now about 53-54% of its original size, so the game is reduced to needing only SWRAM banks 4 & 5. Other minor changes...

• A few spelling / grammar issues.

• Changes to the memory usage (in anticipation of larger games in the future). Some code is now running from SWRAM banks.

• A test for NMOS 6502. While the loader might work with this, the game won't, so to avoid confusion, this is now checked.

• The FLAREGUN is now FLARE GUN (adjective FLARE, noun GUN) because according to the spell checker in MS Word, FLAREGUN isn't a word.

• A new verb. It doesn't change the game, but to give a reasonable response if anyone tries it.

Jeremy

[EdwardianDuck]

I've uploaded Beta 9 to the top post. Yes, I know, two betas in one day!

This fixes a problem with loading when a 2nd processor is enabled. I also noticed that loading into SWRAM via *SRLOAD seemed very slow on a real Master 128 with a DataCentre installed. I've changed the loader to handle this separately and it seems a lot faster. While the game doesn't use a 2nd processor, I had not allowed for one being present during the load process, due to inexperience.

Jeremy

[EdwardianDuck]

On the off-chance anyone is still interested in this, I've uploaded Beta 10 to the top post. Hopefully this will be the final beta. Mostly just internal fixes to the skeleton code & enhancements for the future. Other than a few text changes, there are no changes to actual game play.

Jeremy

[EdwardianDuck]

The latest beta may still have issues. The development workflow became very messy after I'd added text compression and I think there may be some object attributes missing. However the game should still be playable, just with some constraints on what you can do with certain objects missing.

I haven't looked at this for a while, the port to PunyInform has been my focus recently.

Jeremy

[EdwardianDuck]

Well that took a while (although I didn't even look at it for a couple of months)...

Attached is the latest beta, which is a rewrite from scratch in Inform6/PunyInform. It's got to the stage where it can be played through, but there are some things which need sorting out, mostly cosmetic / tightening up some responses.

I have extended / changed some of the puzzles to try and give it a bit more depth.

Compared to the original version, it is a lot slower, but should run on a wider range of machines.

The forum software won't allow me to add another attachment to the top post, so the latest beta is here.

Jeremy