I have been rather overwhelmed by the wonderful response to my Open Letter. There are many excellent comments among the 100 or so. Most comments are simple words of support for the idea, and are very much appreciated for that, but a small number had other messages that perhaps deserve to be a bit more accessible. I’m quoting a few of them here. Where more than one person made similar remarks, I’ve rather arbitrarily selected among them. Thankyou to everyone for your support, encouragement and ideas.
I should also say right up front that Tony Hey has responded briefly that he is checking, so I remain optimistic that something might just come from this.
Howard Besser’s early comment was typical of many comments in support of the idea. He also paid homage to Lee Dirks. I had mentioned Lee in an early draft of my letter, and I truly believe he would have supported this idea internally as strongly as he could. In the end I took the mention out as diluting my main point, but I’m delighted to see it brought out here by Howard:
“Chris, we all know that this would be an important step that would facilitate digital preservationists in doing their jobs. And it is particularly timely, given that our community has recently lost our major advocate within Microsoft– Lee Dirks.”
Libor Coufal of the National Library of Australia makes it clear that my particular problem with old PowerPoints is potentially widespread:
“Like many other memory institutions, National Library of Australia has a load of files in legacy MS formats. Just a quick search in our small testing sample of files returned several PowerPoint 4 files which can’t be open with the current PowerPoint version. Any initiative which would help to solve this problem is very welcome.”
Other advantages… and patents
Gary McGath (who writes a blog related to File Formats) pointed out other advantages for Microsoft:
“I hadn’t realized Microsoft has documented as much as it has till I started looking around its Open Specifications pages. Expanding their scope as you suggest would certainly be beneficial, and I’d like to mention another benefit to Microsoft: It would let other people do their work for them. Microsoft has little interest in spending money to support formats from the nineties, but if other people take up the slack in open source they will add to the long-term value of the formats, and thus give people more confidence in the long-term viability of their current formats.”
Gary also raises another concern, on patent implications. I believe this might well be covered by the terms of Microsoft’s Open Specification Promise, but it’s worth mentioning here:
“Microsoft alludes to patent licensing without getting into specifics. It would give even more confidence if we could be confident that open-source implementations wouldn’t have the threat of patent lawsuits hanging over their heads.”
Euan Cochrane takes me to task for the tone of a remark about emulation:
“This is a great idea and I would be very impressed if Microsoft were to release their standards documentation.
“I am, however, a little concerned about your statement that:
‘At present there appear to be only two routes for migration: one relies on technology preservation (or emulation) in the form of systems that can (and are licensed to) run a sufficiently early version of MS PowerPoint, and the second is via this small company, Zamzar. Neither of these solutions can be relied on for the long term.’
(emphasis added [by Euan])
“Emulation can be viable over the long term and therefore migration by emulation can be viable over the long term (e.g. this. Furthermore I fail to see why you believe that migration is likely to be any more viable. I can give countless examples of the use of emulation right now (such as for mobile phone software development), and examples of the use of emulation going back decades.
“Nevertheless, to be able to have a viable long-term emulation solution we will need access to the software of yesteryear. As such I would love to see your open letter extended to include a request for access to old Microsoft software. It would not have to be without cost and perhaps could include a custom license for use only by memory institutions and/or with other restrictions. “
My Open Letter didn’t make this point clearly, but licensing issues were at the back of my mind when I made that point about emulation. Others were concerned too. Mark Jordan wrote:
“If emulation is the only practical option for accessing these files, then Microsoft should relax the licenses on its older products to allow installation and use specifically for digital preservation purposes. These older products would not compete with current ones, and preservationists would have relief from the single biggest non-technical problem with emulation — software licenses.”
Meanwhile gwhatjeff (I don’t have a more complete name) had a practical suggestion that cuts across the emulation and hardware preservation approaches:
“Hi Mark – I agree that Microsoft could do a lot to relax licensing standards, but there are plenty of old Office licenses and media available for ~$10 per. I’m going to purchase some older versions (Off95/Win95 or Off2000,Win2k) to support getting this set up in emulation. Even without some sort of hosted emulation, I believe most digital archive organizations could set up a legacy OS and file conversion PC for < $200. The harder part would be to get the various hardware components set up properly, particularly networking equipment or floppy drives.”
MetalSamurai (again I don’t have a more complete name) was one of those concerned (as am I) that there are actually no specifications to release:
“You can bet that in this case the code *is* the documentation*. So you’re asking MS to release the source for old versions of Office. I can’t even keep a straight face thinking that.
“[* Really. Even now the MS Office document format is "what the software decides it is". That's why the Mac version never quite displayed documents the same way as the Windows version. There was no proper documentation. And MS will have no interest in writing it now. MS have however promised that the next version of Office will finally conform to the ISO standard they bought a few years ago. I'll believe it when I see it.]”
This was backed up by Jeff Meyer, who wrote:
“Hi – as a former Microsoftie who worked in Office marketing in the late 90′s, I fear MetalSamurai is correct & that this is neither simple nor straightforward, as Jerome has described. The letter’s assertion that the number of people who might be familiar with these formats at Microsoft is declining is an understatement. It might be declining from 2 to 1 or 1 to none at this point. […] I highly doubt there’s an old spec doc sitting around in their files. If there is, it’s full of errors.”
I’ve extracted a couple of sentences here that are off this particular point, but include them later. You can see Jeff’s comment un-edited in the comment stream.
There were several suggestions for other solutions or approaches. Henk Koning wrote:
“I would like to suggest that you consider less far going alternatives. It might very well be impossible for Microsoft to do what you ask, for instance because there is no authoritative file spec, or specs have been changed in an inorderly fashion, or there are anomalies in the specs that would not be understood by the world nowadays (and ridiculed). How to make this a safe journey for Microsoft?
“What I can think of:
- Microsoft starts giving support to specific migration problems
- specs are given piecemeal wise, related to specific migration problems, on request, and after signing a confidentiality agreement
- Microsoft opens a migration service with limited responsibility (but an estimate of the success of the migration)
I’m not in favour of confidentiality agreements in the long-term preservation arena; openness seems to me the key. Nevertheless Henk’s suggestion might allow competitive commercial alternatives to become available.
Jeff Meyer also wrote:
“Euan’s answer suggests the only reliable – and already available – method for this preservation, which is emulation. In fact, it would probably be easier for Microsoft to set up a hosted instance of an older version of Windows & Office just for converting old files than it would be for them to reverse-engineer their own standard. […] Not trying to be a spoilsport, just trying to suggest a solution that will get what you want – access to your files – reliably and quickly.”
(The sentence removed here is already quoted above.)
Jeff also added, in a later comment (remember, he’s an ex-Microsoftian!):
“Chris – I’m interested in figuring out a way to help, but am confused by many of the comments on this thread (maybe people aren’t reading the other comments?). Is the goal of having the specification for the sake of having the specification itself, or is it to reveal the content of potentially unreadable files? If the former, that may be searching for the nonexistent. If the latter, then you do not need an explicit specification document in order to do that. You just need working software that implements the specification.”
Libor Coufal of the National Library of Australia responded to this:
“Jeff: Yes, the ultimate goal is (not only) to reveal the content, but more importantly to save it in a newer, working version. If you can have access to software which can do it, then you’re saved. But what if you don’t? How much trouble (and expense) you’ve got to go into to get you there? And is this a long-term viable solution? I guess, having the specifications available would give everyone a greater confidence that such a solution can be developed, not only now but also anytime in the future. Having said that, I perfectly understand that it may not be viable neither, but if it is, it would definitely be very appreciated (as you can see from the comments).”
Other suppliers too
A number of comments highlighted that we should take this initiative to other suppliers as well as Microsoft. Apple was mentioned a few times, as in this comment by Ben Fino-Radin:
“Full endorsement. Now, who is going to write the open letter to convince Apple to update their ‘old software’ page?”
Kara Van Massen supported this, and points out another powerful argument:
“Very much in support of this. I hope Microsoft sets an example that other companies (ahem, Apple) might follow. Not only is this important for cultural heritage and memory institutions, but perhaps even more so for corporate assets in legacy formats that have business or legal reasons for preservation.”
William Anderson has a good point on the need for rescue missions. To some extent, “File format November” might form part of that, but in reality many more tightly focused efforts will be needed:
“Microsoft can help set a standard of practice for others to emulate. However, as has been pointed out […], specifications may be missing, and knowledge already lost. If this is the case, then perhaps these formats need to be nominated for a rescue mission. It’s clear that the content they encode is at risk of permanent loss.”
Thanks to everyone who has commented, and as noted all comments are accessible (unless they’ve inadvertently been killed by the spam filter, in which case, my apologies).