News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu

The Software Preservation Society

Started by dcdrac, 21:38, 22 May 13

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

dcdrac


Cholo

#1
Noble mission indeed. But i think it will be apropriate to mention that you will probably also need noble patience as well.

Short-ish story: I fell upon the SPS site quite a few years back. Got my hands on a Kryoflux board as well. Kryoflux being able to dump several systems was the real "draw" to get it. That there was a easy to use "DoDumpCPC.bat" batch file in the Kryoflux-forum was probably what convinced me to buy it in the end too. The promise of possible both read & especially writing back of amiga discs was really tempting. Anyways got the board & hooked it all up with a FD1 and it just worked great, i was getting green lights everywhere.

Then i send in my first email to SPS about "i got 3" disc+kryoflux and i want to help". First answer back didnt show untill 5 month later or so(i had pretty much forgotten all about it then). Communication was pretty ok from then on but it felt like i had to drag any information out of them really (should i even use that batch file, where , how, additional pictures/scans of dokumentation/box etc). Anyways i send i my first dump and they wrote it was ok etc.

During the next 2-ish years i must uploaded another 50+ dumps but didnt hear anything at all (no feedback). At some point they released their v2 of the software that added amiga write back .. sadly at the same time my kryoflux suddenly couldnt recognise the FD1 anymore? (tried 2x FD1 and even a "DDI" version as well). Still have the old v1 software on my pc and it still works ok and recognises the FD1 np. Could just be my version of the kryoflux that has gone "incompatible" for some odd reason (butl the v2 software work ok with my Kryoflux + 3,5" drives so i still think its in the software in the end).

Perhaps this is why we had a thread a while back about making DSK out of your own kryoflux dumps (aka people got impatient?). Also i do think SPS/Kryoflux hopes that people themselves does a bit of work to develope a way to make image files for their systems (aka released the source for dump/file formats etc). I do understand that as there is a LOT of systems out there. Initially i also think the goal of SPS was to just save the amiga floppies? Currently i do think they make priority to save the commodore 64 stuff as well. Do note that only amiga & 64 floppies can be written back.

Anyways, not trying to say anything bad about SPS. As i dont know really and this is all just impressions. I guess SPS is just a small lowfunded group trying to save everybody and i totally respect that. Just do take into consideration that you will most likely need a heroic noble patience as well  ;)

Note: The "wanted" and current games list on their main site hasnt change once in like 8 years.

Links:

Kryoflux board:
KryoFlux - USB Floppy Controller

Link to the kryoflux DoDumpCPC.bat file:
http://forum.kryoflux.com/viewtopic.php?f=10&t=53

Dumping your own non-protected data/system discs with kryoflux discussion/programs in here:
Image 3" floppy disks for emulator

pacomix

A thing that is funny to me is that you have to spend a lot of money on the device so you can help THEM preserve the software. Even having the device you have no chance of obtaining an .edsk image in example. If you ask for that in the forum you will be invited to code your own conversion program since the format is "open". And of course I assume you can not write back any of the existing .edsk's.
Also you have no access to the software used to create their native format. So you can only work with raw dumps making impossible generate your own ipf's. It seems the software is really expensive and privative developed by "I do not know the company since they ignored that question I did" and even having the software it seems that only a few mortals in this world are able to understand it and use it correctly. It seems the FM & MFM format and data provided by the device can only be used by "experts" on the matter.
Also they answered with doubts or in a not so clear way about the write back support for 3" units. Everything seems to be a mystery.
I was really close in buying the device but then I started reading the forums and I realized that for my purposes this device is useless. All I need can be done with SAMDISK and due to the amount of doubts and lack of usefulness of the device other than providing them with nice dumps for their PRIVATE collection (yes... you have no access to the database) or having your own raw dumps I decided to not buy it. It is simply not useful.
Why spending around 100€ for a device that can not create .edsk's, doesn't have write back support and whose the solely purpose seemed to be helping them in preserving software for their closed society?


It is a pity 'cos I am pretty sure they would receive a lot of support from the people if the were open since the idea and noble mission seems to be interesting.

robcfg

Well, to be fair, they also said that the process of converting raw files to other file formats is manual and it can take quite some time.


Some formats are already supported in the KryoFlux GUI, others not, but it improves over time.


I agree with you that, for our purposes, with SamDisk you have more than enough, but it doesn't work with very special protection systems.

pacomix

Hi paisa!


   Is stated somewhere which formats can not still be read/written back with Samdisk?

ralferoo

Quote from: pacomix on 12:33, 31 May 13
A thing that is funny to me is that you have to spend a lot of money on the device so you can help THEM preserve the software. Even having the device you have no chance of obtaining an .edsk image in example. If you ask for that in the forum you will be invited to code your own conversion program since the format is "open".
I know very little about the project, but it is fairly true to say that the FM and MFM formats are fairly well understood, and presumably these images are just raw bistreams and a particular rate or timings of when the bit polarity changes (more useful for Apple formats).

There is plenty of information on such matters if you search, so if their dump format is open, it should be possible to write this data back to a disk or convert it to an emulator file for a particular emulator. The real problem is for copy protected disks knowing exactly where to stop writing data. The problem with disks is obviously after a full rotation, you're overwriting data you've just written. Normal disk formats typically have sectors spaced through the disk with some padding gaps. The Amiga is different where it has all the sectors lumped together with a padding at the end. Most protection systems deliberately write data so that the part that is important goes through the padding area which would normally get slightly corrupted when writing. So you appear to have copied all the sector data correctly, but the gaps have slight differences. The real problem is trying to figure out exactly where this data is and where you really should start writing from...

That said, if they have tools that convert from their format to others, I'm surprised they don't share them. That seems crazy. But definitely, you should keep hold of the original dumps as they're likely to have more data than a converted form.

Finally, €100 for this device seems overly expensive. I realise it's a limited market for a dedicated device like this, but essentially it's nothing that couldn't be done using a parallel port and an op-amp on a fast enough PC. You could probably cobble together something to dump raw disk images for €10-€20 if you tried.

Bryce

Quote from: ralferoo on 13:34, 31 May 13
I know very little about the project, but it is fairly true to say that the FM and MFM formats are fairly well understood, and presumably these images are just raw bistreams and a particular rate or timings of when the bit polarity changes (more useful for Apple formats).

As far as I know, KryoFlux is a bit more exact than just a bit stream. It reads the data as an analogue stream, so weak bits and other tricks are copied identical to the original.

Bryce.

ralferoo

#7
Quote from: Bryce on 14:40, 31 May 13
As far as I know, KryoFlux is a bit more exact than just a bit stream. It reads the data as an analogue stream, so weak bits and other tricks are copied identical to the original.
This got me intrigued, so I did a bit more digging. Amongst the gems I found was an interview with Rob Northern himself (who did a lot of the Amiga/ST protections): An interview with Rob Northen » Codetapper's Amiga Retro Site! but also a whole load of useful info here: http://www.atari-forum.com/viewtopic.php?t=9012

The take home point is that the actual voltage of the bit transition didn't matter so much as it was always TTL by the time it got to the drive's connector, so the controller only saw a stream of transitions between 0 and 1. So, actually I was wrong in a way, as you don't even need an opamp to accurately dump a floppy drive!

But actually, how these protections worked was by changing the bit density. One case was to create "random" sectors (so e.g. writing 0101010101 at double the normal bit-rate would result in a bitstream that was essentially random when sampled at the correct bit-rate) which were used to create sectors that could be detected as they'd return different results every read, and by limiting the random portions to specific areas you could identify a genuine disk or not. The Rob Northern technique relied on disk controllers always operating at almost exactly the correct bit-rate when writing, but being tolerant of bit-rates up to about 10% fast or slow when reading. So, this system relied on measuring the read time of the "slow" sector compared to the read speed of a "normal" sector on the same track. If the sectors read at the same speed, they had been copied...

Either way, dumping a disk at a bit level should be a case of asserting the MOTOR signal, using STEP and DIR to seek to the appropriate track and then sampling the DATA at a high-enough bit-rate (4MHz or a sample every 250ns should be about double the normal disk rate). If you were to sample it significantly faster still, you should be able to dump any disk. Obviously, determining the blank areas and so how to write that back to a disk is another matter...

ralferoo

There is also an interesting document here which has a whole load of information about lots of protection systems on the ST (but you'll need to subscribe to their forum to read it): http://www.atari-forum.com/viewtopic.php?f=95&t=21952

It's probably relevant in this thread because it talks about KyroFlux a fair bit...

SyX

If i remember fine, the most part of that information is already in the page of Jean Louis-Guerin :)

Powered by SMFPacks Menu Editor Mod