Development of Blake's 7 (was OASIS development)

Want to talks about games you like, would like to see developed on the Oric, it's here.
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Dbug »

As far as I can see, the current version of the OSDK has the 1.1 version of floppy builder:

http://osdk.org/index.php?page=document ... ppybuilder
Top right of the page:
Version 1-1
Added support for sector interleave: The 'DefineDisk' now has a additional parameter which indicates how far the next sector is (defaults to 1 for a linear search)
Which should maps to this revision in SVN:
http://miniserve.defence-force.org/view ... 61&r2=1393
User avatar
Chema
Game master
Posts: 3013
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Chema »

You may think I am mad, but look at this video. Created 4 disks with the same image and same tools: two are DD and two are HD.

https://youtu.be/CN44y-atxLM
User avatar
Symoon
Archivist
Posts: 2301
Joined: Sat Jan 14, 2006 12:44 am
Location: Paris, France

Re: Development of Blake's 7 (was OASIS development)

Post by Symoon »

Are you using a DD flopppy drive or a HD floppy drive?
Could a HD drive mix the "fake DD" floppy with a HD floppy and read data faster?
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Dbug »

Regarding the theory that the HD floppies may already be formatted, I can confidently say with 500% confidence that this cannot be the case: There are no HD floppies that are pre-formatted using 256 bytes sectors.

Maybe we should ask to Jean-Francois Del Nero (of HxC fame) if he has any idea about what could be the issue.
User avatar
Chema
Game master
Posts: 3013
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Chema »

@Symoon, not sure. The Microdisc controller chip 197x is the one which should decode the signal, and it does not support HD disks (I think). In any case, I put tape on the hole, so the disk should have been written as DD (the bit density is half in DD).

The transfer rate should be the same, and when using HD discs it is near the time when running the game in Oricutron (4-5 secs), so the difference should be somewhere when reading sectors... the interleave is a clue. It is as if it only worked in HD floppies...
User avatar
iss
Wing Commander
Posts: 1637
Joined: Sat Apr 03, 2010 5:43 pm
Location: Bulgaria
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by iss »

@Chema: May I ask you for the image file with interleave 5, which you used for test? I want to repeat your tests with and I can do some low level analyse on PC to see how different are the HD/DD floppies.
User avatar
Chema
Game master
Posts: 3013
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Chema »

Sure, here it is...
B7en.dsk
(500.25 KiB) Downloaded 368 times
User avatar
iss
Wing Commander
Posts: 1637
Joined: Sat Apr 03, 2010 5:43 pm
Location: Bulgaria
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by iss »

Thanks for the image, Chema! Things getting more mysterious for me :)...
But first sorry for my bad suggestion in my previous post, that FloppyBuilder doesn't implement interleave - I wrongly mixed the 'geometry' value with the 'interleave' - they are completely different things.
I compared your test image 'B7en.dsk' with the official one 'B7en-1.1.dsk'. I found some differences in boot sectors and loader - this is OK, but what puzzles me, there is obvious difference in the GAP2 for each sector - in the 'official' image it's filled with the correct value 22 x 0x4E, and the test image has 22 x 0x22. More strange is that this wrong value exists in all versions of FloppyBuilder, even more I found that the origin of the code is in 'old2mfm.c', where someone for unknown reason commented out the right value and changed it to 0x22:

Code: Select all

    /*                for (i=0;i<gap2-12;i++) trackbuf[offset++]=0x4E;
    */ for (i=0;i<gap2-12;i++) trackbuf[offset++]=0x22;
 
Screenshot_20180315_120522.jpeg
Although I think this is bug in FloppyBuilder, the images are working fine in emulators and writedsk2, but I'm curious how did you generate the 'offical' image with correct GAP2?
Last edited by iss on Thu Mar 15, 2018 11:10 am, edited 1 time in total.
User avatar
Chema
Game master
Posts: 3013
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Chema »

Beware you now have version 1.2, so it is different from the official one!!! In fact this one supports Jasmin and has some minor typos corrected (though has not been thoroughly tested yet).

But the official version also includes the same geometry and interleaving.
User avatar
iss
Wing Commander
Posts: 1637
Joined: Sat Apr 03, 2010 5:43 pm
Location: Bulgaria
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by iss »

Please, see my edited post (I pushed preliminary 'submit' instead 'preview').
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Dbug »

http://forum.defence-force.org/viewtopi ... x22#p16274

Looks like I never actually pushed the version 1.2 in nay official OSDK, and I did not commit the change.
I guess I could do that during the weekend, push the various changes I had around, and release a new version of the OSDK with these.
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Dbug »

Ok, I committed the changes I had into the SVN depot, and here the actual matching executable:
FloppyBuilder_1.2.zip
(104.41 KiB) Downloaded 352 times
What would be nice is if you people could validate that:
- the source code change make sense
- this version of FloppyBuilder is actually working as expected

would suck I release an OSDK version with a completely broken floppy builder :)
User avatar
iss
Wing Commander
Posts: 1637
Joined: Sat Apr 03, 2010 5:43 pm
Location: Bulgaria
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by iss »

This is great! Thanks, Dbug. I'll check the news and will do test with real HD/DD floppies asap.
User avatar
Chema
Game master
Posts: 3013
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Development of Blake's 7 (was OASIS development)

Post by Chema »

I think writedsk2 puts the correct values, iss. Don't bother too much, because both images work well with real disks.

And, for what we are discussing now, I made all the floppies with the new image, so the difference is not there.

Based on the symptoms I'd say that it is as if the interleaving is not working as expected on real DD floppies, because we have the timing which is expected if we have to wait for a full rotation to read consecutive sectors. We already made some calculations and numbers were quite in agreement with this. However, using HD gets to timings which are closer to the ideal (with the emulator it takes a bit more than 4 seconds to boot, vs 6 seconds with a real disc, which is not bad at all!). And I am using an interleaving factor of 5, not 3 which was the optimum when I made the calculations (a bit too just, I'd go for 4 at least).

This is something impossible, but also agrees with the fact that all the tests I did with different interleaving factors had no effect on the loading time.

Writedsk2 records the sectors in the order present in the dsk file. The hardware should not be altering this... or maybe is it?

I am also making the assumption that the Microdisc controller's chip 179x does not support HD drives, so there is no chance that the duct tape is not working correctly and disks are being recorded and read in HD. Besides, I don't see how this (if it happens) could alter the loading time at all.
User avatar
Symoon
Archivist
Posts: 2301
Joined: Sat Jan 14, 2006 12:44 am
Location: Paris, France

Re: Development of Blake's 7 (was OASIS development)

Post by Symoon »

People on Atari ST are using specific cards, like KryoFlux, to analyse and dump disks (especially damaged or protected ones).
I don't know exactly what those cards do and how they work, but maybe then can help understanding this case ?

Maybe worth contacting someone on an Atari forum, to ask for a hand - and at the same time see if such a card can dump Oric disks ;)
Post Reply