For the half-year to 30 June 2014, the IPKat's regular team is supplemented by contributions from guest bloggers Alberto Bellan, Darren Meale and Nadia Zegze.

Two of our regular Kats are currently on blogging sabbaticals. They are David Brophy and Catherine Lee.

Thursday, 2 April 2009

Microsoft, TomTom and a dogfood post

The contents of the IPKat's postbag are rich and varied, and he suspected he'd be getting more than just the usual comments after he posted this item ("Microsoft, TomTom best friends ...") on Monday. The following item, sent in by reader Keith Braithwaite, is entitled "Eating your dogfood and having it?"  It reads as follows:

"Like the Norse gods, public companies are creatures of law and order. Unlike individuals, companies cannot protest against laws they find irksome by breaking them without inviting ruin, as the gods did. This is a good thing. It is only through respect for the laws regarding copyright that the Gnu Public Licence (GPL) can do its work, for example. Anyone who supports the enforcement of the rights and restrictions of distributors of GPL’d software and, at the same time, considers themselves entitled to infringe the protection of other content (copyright of music recording, say, or patents on technology) is in a curious position. One might not like the broad extent, in time or space, of IP law, but it is what it is.

The GPL is an interesting document. Perhaps it could only have been written by a physicist-turned-programmer. The GPL says what it means and means what it says, an unusual property in something labelled as the “Public” anything. The GPL doesn’t mean what many seem to think it ‘must’. In particular, it doesn’t mean that changes to a GPL’d program must be pushed “upstream”, nor that a GPL’d program may not contain patented technology.

It does mean that certain kinds of licensing arrangement have to be in place before software containing patented technology may be distributed under the GPL. The nature of those arrangements became much stronger in GPL v3, but the Linux kernel is still distributed under GPL v2.

In that limited sense there is nothing to see in the recent Microsoft/TomTom deal. Whether one is happy with the state of intellectual property law or not, it seems to have worked as advertised in this case. Arrangements have been made, GPLv2 has (we are assured) been respected, the legal rights and responsibilities of the various parties have been upheld.

A lot of web commentators really wanted this to be a Good v Evil showdown, with the sizeable backlog of uncertainties around Linux, Microsoft and patents settled in public. That hasn’t happened. There isn’t a consistent interpretation, either. Did Microsoft back down to avoid a courtroom test of the FAT long filename (LFN) patents? Did TomTom back down to avoid … being crushed like a bug? Did Microsoft win because they got the money (and dodged the countersuit)? Did TomTom win because they get to carry on using Linux (and recognition of their navigation patents)? Who can say?

What we do know is that TomTom now has two years in which to remove the FAT LFN technology from their Linux kernel. Some have expected this to be a re-run of the saga that ensued when Unisys decided to enforce its patents on Lempel-Ziv-Welsch compression (LZW), perhaps partly in the hope that Microsoft be embarrassed in the way that Unisys were.

In that case, the well focused Web technology community devised, agreed and rolled out the Portable Network Graphics (PNG) format to replace the Graphics Interchange Format (GIF), which depends on LZW, in fairly short order. This freed web browsers from the encumbrance of the Unisys patent. PNG was also an improvement over GIF, providing better progressive rendering over the narrow pipes common in the mid 1990’s. Similar conditions do not apply here.

Simply abandoning long filenames in Linux devices provides a clear reduction in the quality of the user experience. The FAT LFN mechanism is actually quite robust, so my SD card with long filenames will work in your non-LFN device, but I’ll have to explain to you that the file you want is the one with the ~ and the numbers … not good. Removing the LFN from TomTom product (while keeping FAT) will clearly degrade the use experience somewhat. And that's just for a very focused sat-nav product. To remove LFN from general purpose devices, netbooks, smart phones and such, would be very nasty for users indeed. Alternatively, FAT itself could be superseded, as GIF has been. On general purposes devices this would likely be a nightmare for users.

There is also the risk of yet another costly, time-wasting, confusing “format war”, just as everyone is getting comfortable with the idea that an SD card or USB drive or whatever will work well with any device into which it will fit. OEMs could try to absorb that problem by adding drivers for the many and various filesystems that could possibly fit on a stick. Even if they did that, there’d then always be the risk that I take my card with the video of baby’s first steps (or whatever) and pop into your camera to grab your graduation snaps and whoops! It’s been reformatted. With a filesystem that my device can’t read. Ouch.

The more optimistic Linux advocates even talk in terms of end-users installing new drivers. I don’t see that happening much. If vendors did move in that direction I wouldn’t want to be on the receiving end of the support calls from members of the public with freshly bricked wristwatches, phones, cameras, sat-navs, cars, etc. etc. etc.

There are very powerful positive network effects that come with a ubiquitous technology, as FAT LFN has become. It would be a shame to lose them over a point of principle about which the law appears to be silent, especially as the patents in question expire in a few years.

This settlement between TomTom and Microsoft can be seen as the bullying of a smaller patent-holder by a larger, as the long-awaited first shot that turns the cold “patent war” into a hot one, as the launch-pad for a period of instability and confusion in a thriving market. Or, it can be seen as a healthy incremental step towards normalised relations between the intensively open and the intensively proprietary parts of the software world".
The IPKat thinks this is really useful: having read a batch of material that has discussed (if that be the right word) the settlement at such a high level that, once any name-calling is disregarded, there's not much to illuminate the reader as to what the issues are, here's a piece that at least gives a clue as to what the technology and the legal issues are about.  Merpel says, I've read it twice now and I still couldn't find anything about dogfood. Or catfood, for that matter ...

Microsoft and dogfood: an American idiom demystified for non-US readers here
More dogfood here and here
Dog eat dog here and here
Cat eat dog here
Typo about cooperation between Widows and Linux here

3 comments:

Gerontius said...

Much as I hate to disparage an otherwise excellent post, Keith Braithwaite's understanding of how PNG replaced GIF does not appear to be correct.

Yes, in the face of the patents covering LZW, PNG was developed pretty quickly, but it never EVER managed to replace GIF. For proof, go to the front page of Google and have a look at what format the Google logo is in.

In reality, take up of PNG was very slow. Why? Because this idea that the patents on LZW caused a significant encumberance on the use of the GIF format is just plain wrong. People carried on using it, people carried on supporting it. The only people it caused a problem for were those who were unable (or, more likely, unwilling) to pay the licence fee or to use licenced software for creating GIFs. Even then, Unisys only went after people who used GIFs commercially, just as Microsoft has not gone after your average Linux user and only people like TomTom who are making money from their products.

The whole GIF farrago spread the lie that if something is patented, nobody is allowed to use it at all. That is completely untrue - nobody is allowed to use it "without permission" and both Unisys in the past and Microsoft now have shown willingness to give permission. Nor does it appear to me that the terms are in any way unreasonable unless you want to hold onto the untenable position that everything in life should be free.

Anonymous said...

Further regarding the GIF it should be pointed out that a GIF can be animated, something a PNG cannot. Rather the MNG was conceived for animations, a standard that has not caught on after all these years, not even in many web browsers.

PNG offered alpha channel control, however implementations were frequently of such a low standard that these were hardly every used.

All in all GIF is more an example of the power of inertia than than the power of innovation through pressure.

Anonymous said...

Internet Explorer's horrible implementation of PNG is what held it back from being a force on the web. IE6 didn't support the alpha channel-- it rendered it as grey. So, your spiffy graphics sucked on the dominant web browser. Nobody used them for anything except graphics that were rendered illegible by lossy compression (like screenshots). PNG (like GIF) uses non-lossy compression, so it doesn't get artifacts like JPEG and (unlike GIF) it supports 32 bit color.

Subscribe to the IPKat's posts by email here

Just pop your email address into the box and click 'Subscribe':