The Sinclair QL Forum

stevepoole wrote: I just hope the 'Forked Lightning' recursive program on QPC2 is not he culprit, but who knows ?

I know. It was not. I could be a millionaire if QPC could do such a thing.

Hi Marcel,

Thanks for your clear response.

I don't think I will ever find out why the PC went haywire, but all is now OK.... Oof !

So back to forked lightning....

Thanks,
Steve.

Well, this thread certainly took a strange turn. However, it's good to know there might possibly be a fourth QL entry on the way... though it looks like I should probably install QPC2 on my 14-year-old laptop and hope that it can deal with it. (It's Windows 7, running from a SSD so it should be all right... right?) That way, if Steve does send me this "Forked Lightning" recursion routine and it requires SMSQ/E, then at least I can run it on a machine that's somewhat expendable, i.e. I can re-install Windows 7 if I need to. I look forward to... something. Meanwhile:

END OF WEEK 21:

21 weekly reports for 2021, and barring one mild error on Sinclair ZX World I've managed to keep to the regular schedule. There was one entry this week, Salvador Camacho converting his own Zhunder Vlade for the ZX81 from its Spectrum original that was first entered in the 2011 CSSCGC.

"That is a promise, it is also a threat" said Richard O'Brien one day. To that end there has been one of these from John Connolly. I'm sure I also had promises from Peter Jones way back in the past, and "+3code" 9who has no other name) is threatening to make the most minimal SAM Coupé game ever. Also, I've had a reply via email from Colin Williams who's finally read the review of MotoRace81, and I'm trying to convince him to turn his ability with Z80 assembler to tackle the Blue Challenge so I finally get a ZX80 entry.

These promises are on top of the potential good news above. Is this competition about to pick up just as this miserable rainy weather clears?

EDIT: I've gone up to "bent pin expansion port"! I'm assuming this is better than a chuggy microdrive. Should I bend the pin back to where it's supposed to be?

Hi TMD,

Yes, QPC2 was not to blame for destabilising my PC, as It now runs the lightning progam seamlessly with no PC problems.... thank you Marcel for a great emulator !

But I have abandoned the game version for the time being, as the file got corrupted because of the former instability.

So here is the rewritten screen saver : Just RUN or EXEC it and 'ESC'ape when required. Tested on SGC - OK too. Compiled with Turbo...

Fingers crossed : it should be ok on any QL system. Works by recursion with very little code.

Apart from the game version, I am tweaking the code to do other things involving arborescences : trees, web diagrams etc.

All the best, Steve.

stevepoole wrote: So here is the rewritten screen saver : Just RUN or EXEC it and 'ESC'ape when required. Tested on SGC - OK too. Compiled with Turbo...
Fingers crossed : it should be ok on any QL system. Works by recursion with very little code.
Apart from the game version, I am tweaking the code to do other things involving arborescences : trees, web diagrams etc.

Nice work, Steve. Producing that in such short code was impressive.

Minor point to query: did you really need RANDOMISE DATE in line 120? The Jan Jones SuperBASIC book says:

If no parameter is given then a 'random' number seed is generated, based on the current clock value.

Looks like RANDOMISE with no parameter is the same as RANDOMISE DATE, although I admit I don't know the internal workings of the command.

Hi Dilwyn,

I did not know about RANDOMISE as mentioned by Jan Jones.

What I did notice on my JM QL, was that when RESET, you would ALWAYS get the same random sequence generated when BOOTing up.
This was why I got into the habit of using DATE.
Somebody even devised a 'True_Randomise' function, maybe in QL World magazine, getting a better seed value than DATE.

Perhaps SMSQ/E or emulators have their own ways of doing seeding ?

Best Wishes,
Steve.

dilwyn wrote:<>
Minor point to query: did you really need RANDOMISE DATE in line 120? The Jan Jones SuperBASIC book says:

If no parameter is given then a 'random' number seed is generated, based on the current clock value.

Looks like RANDOMISE with no parameter is the same as RANDOMISE DATE, although I admit I don't know the internal workings of the command.

For reasons lost in the mists of time, Ive always used RANDOMISE DATE. However, every now and again it is good to revisit old assumptions: It appears that in SMSQ/E if you omit the parameter RANDOMISE simply jiggles the existing random number in the systems variables, ie it is not the same as RANDOMISE DATE, but its all much of a muchness for most practical purposes. It saves a few nano seconds to omit the parameter. Under Qdos things may be different. I didnt look.
There (Qdos) it may also be worth considering that without an RTC one could in theory end up with the same sequence of "random" number after every boot; say you boot straight into the same game each time.
One thing I find strange though is if you dont use RANDOMISE at the start of a program that uses random numbers, you get the same sequence every time: RUN that once in an SBASIC and any number of daughter jobs and you get the same sequence of numbers.. RUN it again in each and you get a new sequence - but the same one across the board.

pjw wrote:<>One thing I find strange though is if you dont use RANDOMISE at the start of a program that uses random numbers, you get the same sequence every time:
<>

Ah, not so strange after all: SBASIC has its own random number store (sb_rand). This gets updated on use of RANDOMISE either with the given parameter or from SMSQ/E's system random number (at sys_rand, which is not affected by RANDOMISE). Theres more, but I wouldnt want to spoil your fun..

Hi,

Here is another variant of the previous recursive screen saver, which draws 'neurones' : RUN or ESCape.

The counter is to keep account of recursion, too much and the interpreter would fatally crash !

I hope a 128ko QL will have enough memory : Ok on SGC !

Regards,
Steve.

Hi,
Sorry for hogging this thread with screen savers. I will try finishing the 'games' versions soon.

Here is another recursive version, a tree designer : Lrun or ESCape.

The first counter is for recursions, the second for the total of branches.

As 'pruning' is random, the trees are more accurately drawn as scrub !

The code runs slower than the prevous versions, as there are a lot of details drawn.

But the core code is all much the same : it could even be adapted for sorting...

Steve.