Standard Bearers

There was a time when I briefly considered that Microsoft was being treated unfairly. It was around the point that the European Commission threatened to fine them two million euros every day that they failed to publish documentation for Windows.

I mean, give them a break, right? They make money from selling software and if they publish how they do it, other people will build on their efforts to steal their market share. It might be good for the consumer, but it's a bit hard on the poor saps who spent all that money on R&D so that somebody else could profit from it.

Fortunately this sense of pity recently passed when Microsoft demonstrated yet again the kind of corporate megalomania that brought the wrath of the EC down upon itself in the first place. Here's a few of the ways that Microsoft is completely screwing everyone who is foolish enough to buy into their products...

1) PlaysForSure

When Apple introduced the iTunes/iPod system, Microsoft scoffed. Apple's idea was simple - produce a great piece of hardware that is compatible with the current most popular free music standard (MP3). Once this catches on, create a music download service that supports this hardware. Now you have the only non-free service that is supported by the most popular player, hence you have the most popular service.

Microsoft however didn't get the point of this at all. The way they saw it, their proprietary WMA was the most popular format, and was supported by the most players and most existing commercial music download services. The fact that all the players that supported it were rubbish compared to the iPod didn't matter because as usual Microsoft assumed that quantity would trump quality.

It came as rather a surprise to them when the iPod not only stomped all over the existing music player market, but then iTunes proceeded to stomp all over the music download industry as well as iPod owners were no longer interested in buying music in iPod-incompatible WMA format.

Microsoft, entirely missing the point of why all this was happening (because the iPod was so much better than the other players on the market) assumed that the reason why WMA was falling on its arse was that the consumer was confused by the range of incompatible players and technologies, so they created PlaysForSure.

PlaysForSure was the one music standard to rule them all. The smart consumer would buy music from sites that carried the PlaysForSure logo, they would then buy a music player that sported the same logo, and that way they could be certain that their music would work and Microsoft would get a nice commission by charging all the music and hardware sellers for PlaysForSure certification, without which they couldn't display the logo, and would be avoided by consumers.

Except of course Apple didn't support PlaysForSure, which meant most users didn't either (it didn't work on their iPod). Which is probably just as well because when Microsoft woke up to why iTunes was succeeding (because the iPod didn't suck) and decided to the release the Zune (which does suck, immensely), they neglected to support PlaysForSure.

So let's just get this straight, Microsoft not only has the audacity to produce a music standard that isn't compatible with the most popular player on the market and call it "PlaysForSure", but they then fail to make it compatible with their own player. So assuming that (despite being a piece of crap) the Zune makes it to number 2 in the market by riding on the back of the Microsoft brand (and no doubt some FUD put out to confuse PC users into thinking that iPods don't work on Windows or something), then at that point it will be the case that PlaysForSure is in fact unsupported on the top 2 most popular players, which together will probably occupy something like 90% of the music market.

I bet all those PlaysForSure buyers are glad they backed MicroSoft now huh?

2) EOOXML

The format with many names and acronyms, including but not limited to OOXML, EOOXML, OpenXML and Microsoft Office Open XML is gearing up to be one of the biggest turkeys yet seen from Microsoft.

Ostensibly OOXML (or whatever) is intended to be an industry standard open format for exchanging Microsoft Office documents such as Word files or Excel spreadsheets. It supposedly marks a new era at Microsoft and stands as a testament to their claims that they are embracing interoperability with their competitors and bringing an end to dominance through obscurity.

Why would they do this? It's not so that people can write competitors to MS Word obviously, that would hardly be in their interests after all as the existence of a decent competitor would mean they would have to compete on quality in a market where they are currently the only option.

Most Office users will agree that Microsoft Office has only become more bloated over the last few years, with no significant new features. Many businesses wouldn't have upgraded from Office 97 had it not been for the fact that newer versions dropped support for the older file formats - a free competitor that could read these files would be a very bad thing for Microsoft.

No, they've done it because their lucrative government contracts are in danger now that governments around the world are coming round to the idea that they don't want to have to pay Microsoft for regular software upgrades to let them read their own documents, and have begun mandating that all goverment information must be in formats that are based on vendor-neutral open standards:

• UK Gov's open source "mandate" policy attacked
• Massachusetts moves ahead sans Microsoft
• Open source's new weapon: The law?
• Brazil Gives Nod to Open Source

What was no doubt intended by the instigators of this mandate was that the goverment would switch to using the Open Document Format, a mature and well documented open document standard for office applications that already exists and is used by all the major contenders to Microsoft Office, including Open Office - a popular Linux-based office suite that has been ported to all major platforms.

So the theory goes that if Microsoft was to continue selling Office software to the government, it would need to update its applications to read and write documents in this format.

But that's hardly the Microsoft way is it? Why support an existing well-supported standard when they can create their own that's half as good for twice the cost? So they have created a completely new format that does the exact same job and submitted it for fast-track acceptance as a new ISO standard.

The motive is obvious, by using an open standard they have met the minimum requirement of the Government to let them keep providing office software, and by using their own complicated, quirky open standard they have ensured that no existing software producers will have had a chance to make their programs compatible.

So, on the launch day for their next version of Office, Microsoft will still have no compatible competitors, and by the time they finally come along, Microsoft will have managed to sell another generation of its software.

As an added bonus, those competitors who have managed to reverse engineer the existing Microsoft closed formats will now find that the new ones are incompatible and will have to start again.

This is a petty move, but seen from an objective capitalist viewpoint (and by objective I really mean a viewpoint where we discount the effect that a company's reputation has on profits) it makes sense. That is all assuming of course that the standard they have produced actually is open.

It isn't.

It has yet to be approved of course, but there are already good grounds for thinking that it won't be.

Setting aside the fact that it completely duplicates an existing standard for apparently no good reason other than to extend Microsoft's monopoly status, it has a number of significant flaws ranging from small technical niggles to giant man-eating gaps in the documentation, and there are laugh-out-loud failures to grasp the point of an open standard.

The OOXML standard is 6000 pages long. Clearly the hope was that if they made it big and complicated enough then a) nobody else would ever bother to implement it, and b) it would get through the fast-track approval process before anyone had actually finished reading it. It hasn't - here are some flaws that have already been discovered:

• From Groklaw:
  
  

  OOXML does not conform to ISO 8601:2004 "Representation of Dates and Times." Instead, OOXML section 3.17.4.1, "Date Representation," on page 3305, requires that implementations replicate a Microsoft bug that dictates that 1900 is a leap year, which in fact it isn't

  Instead of using the ISO standard date format for Gregorian dates, OOXML uses a proprietary date format. Why? Because a longstanding bug in Office has meant that the year 1900 was erroneously marked as being a leap year, so if Microsoft were to use a standard format then all their dates would be off by one day. So instead of fixing this bug themselves they expect everyone else who implements their standard to add functions to convert to and from "Microsoft time™".

• The OOXML standard includes support for other document types to be embedded within an OOXML document. Seems sensible, so what document standards do they support? PDF? HTML 4.0? Open Document Format?

  Guess again, they don't support other document standards at all, they support existing closed Microsoft Office formats, and vague references to HTML (which version? There are half a dozen variants, each of which are semi-incompatible, besides which Microsoft's HTML rendering doesn't currently support any W3C standard anyway, and it seems unlikely that the version in Office will) and plain text (with no reference to character encoding, so is this Ascii, or UTF8, UTF16, Japanese, or just whatever they feel like) and RTF (AKA Rich Text Format, which despite the name has no fixed format and was once defined as "Whatever MS Word outputs when you export to RTF").

  To support the OOXML standard fully, an application must be able to read any of these document types and convert them to standard OOXML markup. In other words, to support the documented OOXML standard an application must be able to read a file that can contain an arbitrary number of other formats that are not only undocumented, but in some cases patented by Microsoft. Conversely to support this standard all Microsoft Word has to output is the exact same closed formats it does currently, but wrapped in a file with the OOXML extension.

  Ah of course, so with this new format anyone can open and parse an Office file, it's only the actual content (irrelevant little things like the text in a Word file, or the tabulated data in a spreadsheet) that will be inaccessible.

Basically, if this format gets through the ISO standardisation process it would be a complete travesty. If it doesn't it will be a huge embarrassment to Microsoft. I'm genuinely curious to know which it will be (we'll find out in February).

3) Immortal Computing

There is a fundamental problem with digital data that is only now coming to light. How do we preserve it long-term? I don't just mean the relative fragility of magnetic or optical media versus, say, clay tablets. After all if this were really the problem then we could just archive all our data as 1's and 0's scratched into clay.

No, the real issue is that unlike conventional writing, which is designed to be directly decoded by the unaided human brain, most computer data needs to be converted into a visual or auditory form by a computer program.

The most longstanding formats in which data has survived have been in plain text (ascii) and bitmap form. The reason why these document have survived is basically because they are so simple. The first thing you would think to try when decoding them will work. But even this is only because of a few conventions in computer science that happen to have stuck around for the last few decades, notably: the use of 8 bits to represent a single block of data, the use of the ascii standard for text , and the convention (due to the requirements of legacy hardware such as CRTs) of storing images in row-major order. But advances in technology will make it likely that eventually even these standards cannot be relied upon.

And for anything more complex, it becomes hopeless. If you wish to store fonts and formatting with your text, or to store vector graphics instead of bitmaps, there is no particularly "obvious" way to do this, and the data structures used would invariably appear as complete gobbledegook unless fed back through the exact same decoding algorithm.

To further exacerbate the problem, when we decipher ancient texts found by archaeologists, the process involves pattern recognition. But most modern data storage formats use compression and/or encryption to save disk space, and these processes systematically eliminate all the common patterns in the data by design.

Microsoft to the rescue! Yes, after their previous success with building reliable, transparent data formats that can stand the test of time, Microsoft has decided to throw their hat into the arena and solve this grievous problem for the good of mankind. No doubt it was next on Bill's to-do list after defeating spam.

They've filed a patent for a new way to preserve data long-term. Dubbed the "Immortal Computing" project, Microsoft's aim is to preserve not only personal data but also "civilization memories", by which they presumably mean cultural history. The idea is to create self-describing, interactive "artefacts", that would potentially feature novel power supplies.

Perhaps Microsoft has some vision of a post apocalyptic future where child-like Eloi will all gather in awe around a newly uncovered artifact as it explains to them in a booming voice how Microsoft foresaw the apocalypse and built this device to educate the survivors and rebuild society, starting with fire and the wheel and then swiftly, via MSDOS, back to the panacea of 21st century civilisation - but this time a better, more peaceful society where Apple never existed to bring about dissent.

But let us suppose that in fact this is not a cynical attempt by Microsoft to gain authority over what parts of history are preserved, and lets assume for the time being that these artefacts could actually work, and would be better in some way than, say, burying a printout of Wikipedia in a Teflon time capsule.

Should we not praise Microsoft for this effort to preserve precious information for our descendants to enjoy? I did after all just take an inordinately long break from criticising them in order to explain just what a huge problem this is we're facing...

No. It's actually worse than you could possibly imagine. Not only are Microsoft patenting a new technique for preserving data (which in itself would allow them to charge your great-grandchildren to read your journal, or view your photo albums), the terms of the patent are actually so vague that they are essentially trying to patent the whole idea of preserving data long term.

The following is included in the patent application:

What has been described above includes examples of the subject matter. It is, of course, not possible to describe every conceivable combination of components or methodologies for purposes of describing the subject matter, but one of ordinary skill in the art may recognize that many further combinations and permutations of the subject matter are possible. Accordingly, the subject matter is intended to embrace all such alterations, modifications and variations that fall within the spirit and scope of the appended claims.

So there we have it - in the space of a few months, Microsoft have...

1. Invented a digital music standard that is not even compatible with their own music player.
2. Submitted an open standard that requires developers to support several closed standards in order to implement it and requires anyone who uses it to regard 1900 as a leap year when it wasn't.
3. Attempted to patent not only one (already derivative) method for preserving data, but the entire concept of data storage and preservation.

It seems like the new, improved Microsoft still has a way to go...