It’s always interesting when readers contact me. Most days, it’s just spam and companies who manufacture things in China wanting my business. But this was different. I received this genuine e-mail from Mark on 12th June which said the following:
I read your article about the SyQuest SQ5510 88Mb Removable Hard Disks. I have a Syquest 88mb 5.25″ SQ800 Cartridge. Do you have any ideas how I can get the files off of it ?
Thanks for your time…Mark
Being a bit of a digital historian/archaeologist, I was lucky to have the SQ5110 drive that I earned from doing previous recoveries for UNSW. It wasn’t completely healthy (given the age of the thing, it’s a bit of a miracle it works at all!) but I was definitely not letting its last few hours go to waste. I had pledged to use whatever life it had remaining in a responsible manner – to help the “occasional” forgotten SyQuest be “reanimated”.
Seeing as it was a genuine request, I was very much interested and replied the same day (e-mail trimmed to save reading):
Do you have a SyQuest SQ5110 or SQ5110C drive and an associated working SCSI controller? […]
If you don’t have the required equipment – your options are relatively limited. You could try contacting specialist data recovery firms […]. Or you could purchase the said equipment off eBay (and pay a fair premium for it, without many guarantees on condition).
Or, if you are not sensitive about the data, I am willing to donate my time and equipment to recover your disk(s) provided that you are comfortable with sending them to me (and covering postage fees) and understand that I cannot guarantee that your data will be readable (given the age) and I cannot guarantee that the files will be of use (due to file format changes). I can’t guarantee my SQ5110 drive doesn’t blow up trying either … as you can appreciate with things of this age, it’s not going to necessarily be a stroll in the park. Nor can I guarantee against loss in postage either.
I will render the services at no cost to you (other than postage, however, I will not return the disks to you), and I make no guarantees about time frames etc. If you opt to do so – I will return your files via a ZIP archive on my Dropbox (as this is most convenient).
Do have a think about it and if you choose to send your cartridge(s) to me, please give me a few more details and I will give you my postal address.
Just being my usual helpful self, or so I think. Anyway, as I am merely a hobbyist, it has always been my nature to refer people to “experts” – i.e. professionals. But this was a special case – SyQuest cartridges are rarely encountered and most data recovery firms now specialize on hard disks and won’t touch floppy drives, ZIP disks etc.
I received a quick reply from Mark the next day …
Thankyou for taking the time to reply at length to my email, it’s very much appreciated.
Unfortunately I don’t have the required equipment. The Syquest 88mb 5.25″ SQ800 cartridge I have contains Macromedia Director files […] since 1995 but the Syquest cartridge has been stored well.
I have emailed a couple of data recovery businesses, no luck so far but I might try a few more and I’ll let you know if I have any success. Otherwise I would like to take up your offer.
We only used Macs at the Uni so it would be Mac formatted. If the files are successfully recovered I have to then work out a way to render the animations out to a video file. Do you have any experience in that area?
Once again thanks for your time…..Mark
It’s not surprising that there wasn’t much luck from recovery businesses – however, that being said, the challenge was more than just getting the bits off the disk. Most recovery companies aren’t too interested in that. Anyway – I decided to reassure Mark of the “plan” –
Thanks for the info. With Mac formatted Syquest disks, I can always read them into a raw (~88mb image) file which can be formatted as a .dmg in which case double clicking it on a modern MacOSX machine will mount it *just like it was a real Syquest*. […]
I was lucky that the Mac carts I had worked with also contained a backup of the programs used to create the files themselves – so I managed to load that up into an old System 7.55 or Mac OS 9.0.4 emulated machine (as the application backups were for Motorola 68k or PowerPC architecture – long since abandoned with the Intel Macs of yore).
Unfortunately, I don’t think I have access to old versions of Macromedia Director for System 7/8/9 (which sounds like the likely candidate for creation of the files) – but you may (hence the image may be the best technique). If the software was backed up to cartridge, exporting to video could be tricky from within Mac OS […].
I hope this information gives you some ideas – getting the bits off of the disk is only half the job :).
It was at this point, our communications stopped. I had naturally assumed that he found success or found someone to do it for him – but that wasn’t the case. A month later, he replied:
Sorry to take so long to get back to you. I received some quotes to attempt to retrieve the data from my SyQuest Drive. To cut a very long story short there are no guarantees from any of them. I would like to offer you […] cash as a contribution for your time if you can attempt to retrieve whatever you can from my drive and get it to a video or graphics format that I can use.
Thanks for your time…Mark
From there, it was a simple matter of replying with my address and having the cartridge shipped to me.
The Cartridge Arrives
On Monday 22nd July, the cartridge arrived by registered post. Now, the work could really begin. First, a quick examination of the cartridge as the design differs slightly from the other cartridges I had:
Interesting the “five year warranty” on the front of the outer card. The other ones I had used a different colour scheme and didn’t have the guarantee. As this cartridge was “dated” as being written in 1995, this is 18 years old!
Unlike the other cartridges I had, the serial number of the cartridge was on the rear, and the front featured a full size moulded SyQuest logo.
Onto the Job at Hand
The first step was quite straightforward – prepare the recovery environment. Luckily, I have a machine here dedicated for recoveries. It has Linux on it (lubuntu 12.10), and virtually any modern and old interface available – IDE, SATA, SCSI, USB, Firewire, Parallel, Serial just to name the most important. The SyQuest drive hasn’t been touched since the last set of jobs completed – so I hooked it all up and started it up.
It’s quite normal that Linux takes almost 10 minutes to start with the SyQuest drive installed. During boot-up, it seems that Linux wants to auto-mount/find the geometry of the drive, and with no cartridge inserted, it keeps trying for a while. You might think you can speed it up by sticking a cartridge in, but alas, with my drive, the heads seem to be a bit dirty and the disks a touch difficult to read – so instead it takes almost just as long as it tries to read the sectors and either slowly reports them or gives up.
My tool of choice is ddrescue – the package name is gddrescue. It’s not the same as dd_rescue. ddrescue is special as it’s capable of automatically partitioning and retrying “bad” areas, reading forwards and backwards to complete the job. My previous experience with the SyQuests have shown that they are quite reliable provided you are patient and you are willing to let it retry several times. In my case, I set retries to 20, and had it spit the log file out (so we can resume in case of power failure).
It’s at this point, I should point out something peculiar to SyQuest cartridges – the hardware write protect switch. I ensured it was in the write protect position, to prevent inadvertent corruption and lack of “forensic” quality recovery. But what I noticed from the 44MB SyQuest manual on BitSavers was that the drive itself had inbuilt defect management. Quite an impressive feat for that period when PC’s still had MFM drives with a print-out of defective sectors.
This in itself shouldn’t be a bad thing, although it can cause problems. ZIP disks also have an internal defect management as well. If the drive used to read the disk is fine, then it should be okay, but if it isn’t, there is the potential that the drive will read the data, incorrectly correct it (say, buffer RAM failure or ECC algorithm damaged) and rewrite it back to the disk (and mangle it, say a bad write head) which may destroy any and all chances of recovery. If the hardware write protect doesn’t stop these reallocations/rewrites, it could be a disaster. It’s definitely interesting to think that the simple act of reading a disk could alter the disk and should be concerning to all “historians”.
But I know from experience that it isn’t a problem with this drive – and with the ddrescue provisions, as long as it read back correctly just once, that was all I needed.
In 26 hours, the disk was fully recovered after about 12,000 retries in total. In fact, it would have been faster to download 88Mb on a dial-up modem than it was to squeeze it out from the drive. Step 1 was complete. I offered Mark a ZIP archive of the .img file itself, which I renamed to .dmg to allow modern MacOSX to mount it, although with a warning:
Unfortunately, as it is a raw image file, it will not be accessible on the PC, however, you can double click on it on your Mac to mount the cartridge image and explore the files directly. The programs on the disk will not likely run on modern Macintoshes. Please make several archive copies of it before opening it. Any opening of the image under emulators or Macs will cause alteration to the disk image which will corrupt things like the driver partition and date/time stamps and will not preserve forensic integrity. As it stands, this image can be used to produce a perfect 1:1 copy of your disk provided another machine with a Syquest cartridge drive and cartridge is used. The copy produced from this image will be indistinguishable when placed into a drive on a classic Macintosh – barring the cartridge serial number which is not part of the image.
Phase Two Recovery: Data Interpretation
Unfortunately, this post might get a little boring to some people because there are many things which I just can’t show you all – after all, the data belongs to Mark, and it’s my duty to keep that between us. As a result, there will not be much to see – but just a few thoughts for all.
The most challenging part of the recovery, in the case of non-damaged media, is often making sense of the files. I say that because there are, and have always been, a plethora of proprietary formats. Simple text files will often be easy, but when we’re talking about multimedia, it’s not so simple. The applications that were used to create them are often missing – and the systems on which they run are no longer available or functional.
We knew we were dealing with old Macintosh stuff – so as a prerequisite, “Classic” MacOS was a necessity. Luckily I had already had a setup of Basilisk II and Sheepshaver, along with a set of OSes ranging from 7.5.5, 8.1 and 9.0.4 already set up and ready to go. The image file can be directly attached to these emulators, and explored on the “emulated” classic Macs just like having an actual SyQuest drive.
The disk itself had a copy of Macromind Director 3.1 on it, but at some point in time, the actual files had been opened up in Macromedia Director 4.0. Unfortunately, this upgraded the file format, meaning that getting it to open in 3.1 was impossible. Maybe it could have been possible with a bit of “flag tweaking” using Resedit, but the details aren’t clearly available.
My next thought was to “export” the files from the SyQuest cartridge and get my oldest archived PC version of Director to open it. This was Macromedia Director MX 2004. Surprise surprise …
Or maybe not so much a surprise. Even within the same franchise, it seems, the file formats are updated so quickly, and the support window for older file formats extends only to “the next version or two”. This means that even a new version of the required software is entirely useless in resurrecting the files.
From this, the next action was clear – obtain a copy of Macromedia Director 4 or 5. It took a lot of searching to uncover the one place I could find it – Macintosh Garden. Legally, this move is ethically and legally questionable as the status of the software is technically “abandonware”. What people perceive to be “okay” differs – but I think it’s rather obvious when time is running out, it’s better to get the job done before you can no longer do it. I suppose that if they didn’t want us to do this, they would have allowed their modern versions of the software to open the files in the first place!
It took a bit of juggling to unstuff the file – which revealed many disk images which could be mounted in the emulator to do the install, and then the upgrade. HFVExplorer comes in very handy to inject these downloaded files into the emulated systems, although, the resource fork is almost always missing and this causes problems like the dialog above which basically says that there’s nothing that can open them, and the files aren’t even listed in the application’s Open dialog at all. Luckily I had an older version of StuffIt! Expander which had an “all files” feature in the open dialog which allowed me to unstuff the files.
With the program installed, the files opened but they couldn’t play at full speed and export was broken, complaining of running out of RAM. You have to remember that when you play with emulated machines, they can obey the limitations in the original hardware and add some more. Emulation speed is not exactly cycle-accurate, and neither is the CPU emulation, meaning that Virtual Memory isn’t available. Giving the emulator a large amount of RAM didn’t help as the application just didn’t want to use it all before complaining.
Oh I love a challenge …
What I did? Exported the Director files into a Projector file – this was self contained 68k and PPC fat binary which I then executed under the faster PPC emulator. The screen was recorded using Camtasia Studio running on the host, which was resynced manually externally with the audio file (an AIFF) which was taken out of the SyQuest image. After all the editing, it was encoded with MediaCoder to some “modern” H.264 MP4 file.
I couldn’t resist the challenge, so it was all done by Wednesday 24th July!
As far as I can see, this is the best I could achieve given all the hurdles and limitations with what I had at hand. That being said, the result was okay, but I suppose, could be considered not entirely “authentic”. I e-mailed back with the files and the list of “issues” which were encountered, to receive a reply:
Thanks you very for your excellent work. I’m happy with what you were able to extract from the drive and the video files are usable.
Your skills are remarkable and if there is any way that I can sing your praises in the form of a reference or comments on sites etc. please let me know I’ll be very happy to do that.
I have downloaded the files from your drop box and made multiple copies. I haven’t tried to explore the cartridge image on my mac yet but will let you know the results when I do.
Well, what can I say? I can’t be happier with that. If I can do something for someone, and they’re satisfied with it, that’s exactly what I aim for.
Even better was the fact that Mark was generous enough to contribute some money for my time and efforts – this will go to help supporting the site and keeping everything running – so it’s a win-win situation. Thanks Mark!
As a footnote – if anyone else has a long-forgotten SyQuest 44/88Mb Cartridge they want the data from, and they’re willing to part with it, cover the cost of postage and perhaps even donate something for my time, this may just be your lucky day …