IPF and Openness

[mr.vince]

Bring it on, Jeff. Would love to try that!

[DrCoolZic]

I dont see why anyone wants to create their own ipf files though?

context: I am interested in IPF format for preservation of Atari FD
The one thing that I do not like about IPF is the fact that it can only be created by SPS people.
Why? because user do not have control over the process: When will I get back my IPF? What is done with my image? ...

Nor am I sure why anyone thinks they are qualified to do it?

Really depends on goal. Preservation means different things to different people.
Your goal (supported by IPF) is to Preserve (with capital P) existing FD with all original information.

But in most cases people have less ambitious goal: what they really want is to continue to use their games/programs. This could be done by running the sw on an emulator (relatively easy) or by making a backup of the FD (relatively difficult).
In the Atari world there are two solutions:

• one is to use a format that allow to run protected sw on emulator (.stx / pasti format)
• one that use a specific hw to reproduce any protected FD (DC cartridge).

The nice thing about these solutions is that you do not rely on 3rd party to get the work done.

But whatever license you want to use is fine with me as long as I can use the IPF format in a free product like the Steem or Hatari emulator
So many thanks for opening the door

[mr.vince]

The one thing that I do not like about IPF is the fact that it can only be created by SPS people.
Why? because user do not have control over the process: When will I get back my IPF? What is done with my image? ...

Oh, we've heard that a million times before. It is actually good the typical user does not have control because otherwise we'd already have the situation that is status quo for all others, e.g. ADF, G64, FDI... You want a good one, and you have to test ten to find one good. I am not saying that not exchanging information is good, but setting a barrier in terms of quality isn't wrong.

I understand that right now waiting for IPFs is a pain, but that's the price you pay for having it done for free, at a professional level. I understand that you are one of the few that would be able to get involved in the process of making IPFs. But the casual user could only use some automated process, e.g. Keir's code in the dev subforums.

As for what's done with your IPF at SPS... easy thing. You won't submit your private data, but computer art for preservation. So your image gets archived and is given to other institutions, e.g. national archives and museums upon request. I would say this is something good, not bad, and contributors get their image back, too.

Nor am I sure why anyone thinks they are qualified to do it?

Really depends on goal. Preservation means different things to different people.
Your goal (supported by IPF) is to Preserve (with capital P) existing FD with all original information.

But in most cases people have less ambitious goal: what they really want is to continue to use their games/programs. This could be done by running the sw on an emulator (relatively easy) or by making a backup of the FD (relatively difficult).

You are mixing up preservation with archival. The one is keeping something in its original pristine state (as good as possible), the other thing means just keeping it and using it, but without the intent to create anything that's of use to the community. While the latter is ok and understandable, it's dangerous to call it preservation, as people could see such IPF and think they have the real deal when in fact it would only be a user made thing, that can have errors and might be modified as well.

[Darkstar]

I still think that digital signatures are the way to go here. If SPS would sign all their preserved images, everyone could make sure if he got the "real deal" or not... The only downside is that it would make IPF dependent on an external crypto library, but since signing (or rather signature checking) could be disabled in the source that wouldn't be a problem in most cases.

-Darkstar

[DrCoolZic]

...
Preservation means different things to different people.
...

You are mixing up preservation with archival. The one is keeping something in its original pristine state (as good as possible), the other thing means just keeping it and using it, but without the intent to create anything that's of use to the community. While the latter is ok and understandable, it's dangerous to call it preservation, as people could see such IPF and think they have the real deal when in fact it would only be a user made thing, that can have errors and might be modified as well.

Yes I agree with you The term preservation is used with different meanings...

However I think that if there is definitively a need for "real preservation" with digital signature as proposed by Darkstar, there is also a need for "something lighter" that allow the user to continue to use a game/program. Of course the "real preservation" (that is signed IPF) is a super-set of "something".

I have been thinking at the usage of preserved information and here are some thoughts.
The IPF files contains all the original FD information but what will you do (not what can you do) with it?

• use it to recreate a new FD to use on a real machine
• use it to "play" the original FD. This can be either a software emulator or a FD hardware emulator (like HxC)
• any other use?

All the above is still possible today. But what in 10 or 20 years from now?
I am not a chemical expert but from what I understand the magnetic particles that coats FD support will probably "fall apart" after so many years (sorry for my English). I still own few hundreds of FD that I can use today to reproduce a damage original but what in 20 years? I do not think industrial will create new batch of DD FD ? So having the capability of making perfect backup of FD is probably not going to be important in the future? Plus even if you can still make a FD 20 years from now will the original hardware still work?
So this basically leads to the usage of emulators. But in that case you do not need all the information provided by IPF files. If you have fuzzy bits on FD do you really care if the underlying mechanism used is unformatted sector, or no flux area, or variable bit length?

Do not get me wrong: I think it is important to preserve as much information as possible for the future and IPF seems the right solution. But for an average user all the contained information is probably not necessary.