Microsoft’s ecosystem: cracks in the reliability

microsoft
The microsoft ecosystem is often mentioned to be long term reliable and therefore be a good option for companies with long term plans or needs. However, recently some parts of Microsoft’s ecosystem fall apart and became incompatible, which is not what you expect from a reliable system.

The Microsoft ecosystem consists of thousands of services and applications provided by 3rd parties. However, the foundation of this system is provided by Microsoft’s services, applications and it’s Operating System. When these are suddenly not reliable anymore the ecosystem is suddenly not stable at all.
And there have been three large cracks in the near past which should make everyone worry who bases his or her IT on Microsoft’s products: PlayForSure, Outlook Express/Hotmail, and to some degree even OOXML.

PlayForSure

PlayForSure (DRM codename: Janus) was introduced by Microsoft in 2004/2005 to mark players which have been certified to follow Microsoft’s dream list: Windows Media Player compatibility, DRM support, MTP-only support, and so on. Originally this was also an attempt to push Ogg Vorbis out of the field.
The basic idea was that every player could be used on any Windows System which had Windows Media Player 10 installed. And that every player would play any music sold in any PlayForSure store.

The shortcoming for the user would be that the players and music wouldn’t work outside of this system (Apple, Linux, Windows 2000, Windows XP without the WMP 10, etc., other MP3 players). And if the user wanted to switch to a new computer, he/she had to re-download a licence for the new computer. That’s a usual problem with DRM, but is important in this regard.

So the ecosystem was built up, and almost every new MP3 player on the market was certified against PlayForSure. Millions of songs were presumably sold to probably millions of users.

But then Microsoft decided that PlayForeSure is not as cool as the iPod, and published an iPod enemy, the Zune. But since the Zune had to be Microsoft-Only it wasn’t Play Fore sure. The ecosystem got it’s first crack. And then Microsoft thought that having two systems was a bit confusing and closed down PlayForSure at the end of 2006. In the future the label “Certified for Microsoft Windows Vista” is supposed to replace PlayForSure. But while that certificate has some elements of the former PlayForSure it - for example - requires Windows Vista and has other, additional requirements. It is a new logo.

What we have now is a broken ecosystem. And that although most of Microsoft’s partners relied on the availability of PlayForSure - many of them still advertise their products as PlayForSure certified.
So now we have hardware vendors who have to push out a new line of players which can be certified against the new logo. All their old devices now depreciate because they are not certifed against the new logo.
And we have the webshops which sold PlayForSure music. Many of their customers for sure still use Windows XP - which is of course not Vista ready. They are left alone a bit.
Last but not least we have the customers: music which was bought with PlayForSure needs licence upgrades to work on a new computer. Since these licence upgrades will not be provided after PlayForeSure’s shut down the music is essentially worthless. Think of millions of songs here - think of millions of dollars/euros here!

Outlook Express and Hotmail

Outlook Express is an e-mail client shipped with all Microsoft versions since Windows 98 and before Windows Vista. Although it was responsible for many large scaled virus attacks and spreads due to a horrible security concept and had many more, other technical shortcomings and problems it was widely used. And it was of course often used with Microsoft’s own E-Mail service, Hotmail.

However, recently Microsoft decided that the old protocol used for the communication between Outlook Express and Hotmail wasn’t what they would prefer. They wanted to switch every user to the new DeltaSync - and so they decided to drop Outlook Express support in Hotmail. As a resultyou will not be able to use Outlook Express together with Hotmail in the future any more. Instead, you have to upgrade to Windows Live - if you can, because currently it does not support Windows XP 64 or Win2k3. Also, if you used to use another client, that one will also not work anymore.

To summarizes: as a user who relied on Microsoft’s in house e-mail technology you are screwed.

OOXML

The last problem is still in development, and the outcome is not really clear at the moment. But the root of the problem is ISO’s adoption of OOXML.
Originally published by Microsoft to have a pseudo-standard which is very hard to support for other software vendors, the ISO adoption process became necessary because no one was interested in not officially standardized formats by Microsoft.

So OOXML went through the ISO process, and got several comments where it has to be reworked - Microsoft could have avoided this by working together with others right from the beginning! In the end the new format was accepted as a standard. But the necessary revision of the format resulted in a non-compatible Microsoft Office:

Such a test is only indicative, of course, but a few tentative conclusions can be drawn:

  • Word documents generated by today’s version of MS Office 2007 do not conform to ISO/IEC 29500
  • Making them conform to the STRICT schema is going to require some surgery to the (de)serialisation code of the application
  • Making them conform to the TRANSITIONAL will require less of the same sort of surgery (since they’re quite close to conformant as-is)

Given Microsoft’s proven ability to tinker with the Office XML file format between service packs, I am hoping that MS Office will shortly be brought into line with the 29500 specification, and will stay that way.

I’ve included the last part (emphasis not by me) to make the point of view of the poster clear. He thinks/hopes that MS will update the Office suite soon (an ODF test suite can be found here, btw.). However, Microsoft haven’t published any time frame, schedule or plan yet about such an update.

So in the end Microsoft first created an ecosystem around its new file Office Suite and on full purpose introduced a file format which would never become an ISO standard in that form - to afterwards apply for the said standard and change the file format again. As a result Microsoft Office 2007 still is not compatible to the new OOXML version itself atm.

Every organization and every software vendor who already jumped on the OOXML train before the ISO changes therefore has now to reconsider the software strategy: it should prepare for the switch to the updated OOXML - but also wait with it because it is unclear when the switch in MS Office will come.

However, preparing such a big update and keeping it ready consumes quite some resources. And small software vendors might not have these resources. Additionally, even smaller vendors which only had the resources to accept the new file format once are screwed.

And of course we have the users: everyone who tried to base a long-time backup solution on top of the old OOXML is screwed as well and should re-build the entire update again with the new OOXML.

Final words

The main problem behind these cracks in Microsoft’s ecosystem is of course the lack of well proved, open standards. But since Microsoft is not willing to support standards, the cracks are there now, and they are huge.

Of course from Microsoft’s point of view all these things make totally sense: everyone who strictly follows Microsoft as close as possible (use Vista only, etc.) and is ready to lose quite some money once in a while (because the music must be bought again and again or similar) will survive all these changes.

So to summarize the problem with respect to Microsoft’s position: problems only occur where users don’t have money to waste or want to have choice. Microsoft does not want the user to choose anything. The user is there to consume, not to choose - or actually think. This is best expressed by a comment from Omar Shahine, one of Microsoft’s Lead Program Managers, who tried to explain the Outlook Express/Hotmail development:

What we didn’t want to do is offer POP access and then have [...] customers use POP over DeltaSync.

If customers decide to stay with a product despite you offer them a new one - that means the customers prefer the product. Usually a company has to stick with such things in the real world.

However, Microsoft thinks it knows better. Microsoft forces it’s users against any opposition. And against any possible casualties on the consumers side.

The new one in town

VirtualBox for OpenSolaris hosts available - FreeBSD soon?
Sometimes old ones have to be replaced - with something fresh and new. I’ve got a new laptop.

The old one

I’ve had my old laptop now for over three years. Back then it was state of the art: A Cebop T900 (the company doesn’t exist anymore, but the series is comparable to Siemens Amilo series of that time) with fresh 1,8 GHz Centrino (Pentium M 745, Dothan core) with - at that time quite large - 1 GB RAM and a 80 GB Hard disk. The graphics card was an ATI 9600 mobile with 256 MB RAM which actually was more of a desktop graphics card then a laptop card. The wireless card was an Intel ipw2200, and there was even a card reader installed, though I never got that one working. The screen was a 1280×800 15.4″ wide screen. Some benchmarks can be found here

The old one - called Perricum, btw. - served me well. However, due to the ridiculous oversized graphics card (and probably due to the entire construction at all) the fan never stopped working. And the fan was really, really noisy. Additionally, the entire chassis was not very stable. Still, the laptop accompanied me during the last years (and to several countries, btw.). For the distribution interested person: I started back then with a pre-release version of Mandrake/Mandriva (because it was the only one which started on that machine at that time due to kernel problems), later on Fedora and also for some time OpenSuse followed. The virtual machines started inside of the system run almost everything, from Ubuntu over the already mentioned distributions to even some BSD versions I gave a try. Also, for some time a virtual machine contained a Windows copy.
Speaking about Windows: besides the virtual machine that computer never saw a a Windows Kernel. The machine came along without a preinstalled operating system, and I was already in a state where I didn’t need any Windows. The need came later up for some months due to some specialized Windows-only program, but that was all.

The new one

In 10 years my mobile phone will most likely have more power than my new laptop - but today it is state of the art: A Dell Latitude D630 with a Intel T9300 with 2.5 GHz Dual Core (6 MB Cache, btw) - yes, that’s the new Penryn, Intel’s new 45nm processor series. Again, the RAM is quite high even for today: 4 GB. The hard disk is comparable small, just 160 GB, but for me power consumption was more important than size. And I do have an external hard disk.
But I also from the problems with my old machine: the graphics card is not a desktop card this time, but a Nvidia NVS 135M, a business card from Nvidia’s Quadro series. I could have also taken the laptop with Intel’s X3100, but since this machine is supposed to serve me for a longer time I wanted to have at least a bit more power. The wireless card is a 4965, and the screen is a 14.1″ with a resolution of 1440×900. Some benchmarks can be found here.

Additionally, this is a business notebook: so there are less fancy things (no card reader for example, no TV out, etc.), higher costs, but also more reliability: a much, much better and more stable chassis, high quality hardware (the speakers for example), and so on. And it has docking station support, which I plan to use quite often in the near future.

The operating system choice was a lit more difficult in this case, however. Dell doesn’t sell its flagships with Linux - at least not to usual customers. Additionally, since I am an author for technical magazines sometimes I am asked if I could do a story on some Windows related topics - like “Connect Windows to Linux” or “Compare xyz on Windows against abc on Linux”. Since I didn’t have a Windows copy in the past I had to pass on these tasks. Also, having a look at Vista once in a while makes me able to compare it against Linux distributions.

So I payed money for a Windows Vista copy. The disadvantages are now that that Microsoft earned some money and that my hard disk is littered with a lot of crap - almost two GB are taken by the problem that I only got a “Recovery DVD”. Sigh…

Sooo…

As a result I will publish more posts about the new computer - for example the task of installing Linux on it was much more difficult than I have thought! Also I will install a more recent KDE on this machine - one way or the other. But for now I’m just happy to have a new computer which is strong enough to do all the things I would like it to do but is not too noisy to work with or to heavy to carry around. :)

fsdaily.com - a digg-like portal for Free and Open Source Software

VirtualBox for OpenSolaris hosts available - FreeBSD soon?
Recently I stumpled on the project fsdaily.com. It is a digg.com-clone but is focused on Free and Open Source Software. While the community seems to be rather small yet it might be an interesting news source in the future.

When I came across fsdaily.com I noticed an interesting piece of news about future EU plans to buy more Open Source software. It caught my eye because I hadn’t read that anywhere else at that time. Therefore I had a deeper look at fsdaily.com, and must say that it looks rather nice.

The more important question however is if it will suffer the same problems as digg.com. While I loved digg.com when I first discovered it I dropped it some months later due to quite some disappointments: digg is, after all, just a place for the masses. So the relevance is also defined by the masses. And well, the masses love sport, glittering things and naked skin. The equivalent in the software world might be iPods, themes and desktop screenshots. That’s ok for the masses, but as a result digg.com is just not the place if you want to get decent information or want to follow technological development.

That however might be the case with fsdaily.com - although I must admit that even the FLOSS world might have it’s equivalent for sport, glittering and naked skin in the form of a dozen hot flame wars. But at the moment the situation looks rather calm, and due to the specialized focus I think the topics will stay more technique related and really interesting than just glittering.

Also since it is still a quite small project, you can still influence it quite a lot with just a couple of people. So where is the next Commit Digest? In the end it’s the community which decides which news are interesting.

Graphics and Free Software: a great 2007, but where is OpenGL?

cube-with-matrix
2007 was probably The Year Of Free Graphics: AMD/ATI’s specs, a new totally Mesa , output hotplugging via XRandR and the announcement of new shiny OpenGL specs. While this all was truly great, the OpenGL releases never happened, and there are no updates on the topic.

Graphics in Free Software were always a difficult topic in former years: specs for many hardware bits were missing, proprietary drivers often didn’t support new hardware for several months, the graphics API (OpenGL) was seriously falling behind Microsoft’s competing API (DirectX) in terms of features and hardware support and last but not least the main free implementation of the API, Mesa, was hopelessly outdated.

But this all changed, and many important changes were in 2007:

While this list is indeed impressive, there is one big gap: the announcement of the new OpenGL specs. These never happened. Not even a cleaned-up version of the current OpenGL API ever came up anywhere. And, although it was said that the communication with the community would be improved there aren’t any information available about the current state. In fact, all possible channels are deserted. The only bit of information available is a post by Barthold Lichtenbelt (OpenGL ARB Working Group chair) in the OpenGL forum dated end of October 2007 stating that the development of the next spec faced difficulties and that the release will be delayed:

The OpenGL ARB found, after careful review, some unresolved issues that we want addressed before we feel comfortable releasing a specification. [...] None of these issues we found are of earth-shocking nature, but we did want them discussed and resolved to make absolutely sure we are on the right path, and we are. [...] The ARB meets 5 times a week, and has done so for the last two months, to get this out to you as soon as possible.[...] We don’t want to spend time fixing mistakes made in haste.
[...]
More details will follow soon in an upcoming OpenGL Pipeline newsletter.

Of course unresolved issues can turn up. But in such cases a vital part of maintaining a community - and also a stable customer base - is to deliver more information from time to time. And it is important to state schedules as soon as it is possible. If something is difficult, than that must be stated. Communication also means to say something from time to time just to show: we are alive, we care.
But the OpenGL team has failed in that regard: the mentioned OpenGL Pipeline newsletter is totally outdated and therefore a joke, and the forums don’t contain any other official post explaining the situation.

I’m pretty sure that there is work going on behind the scenes - there are many different parties with a vital interest to ship a next generation OpenGL. But the total lack of communication, the inability to handle basic communication principles in the right way is depressing. OpenGL is not only a technology, but also a product, and if the product is managed that bad it can’t be good for the technology.
Also it leaves the sad feeling that, in respect of the API, the free graphics ecosystem is still back in the old days.

VirtualBox for OpenSolaris hosts available - FreeBSD soon? [Update]

Tux
Innotek today announced the first Beta version of their well known virtual machine VirtualBox for OpenSolaris. It is also possible that a FreeBSD version will be available soon.

The announcement can be found in the news section, however there are no further details available than the simple announcement itself:

Feb 11, 2008. innotek today released the first beta of VirtualBox for OpenSolaris.

Neither the Press section nor the Forum have any additional information, and there is also no screenshot provided this time. Heck, there isn’t even anything mentioned at the e-mail lists or at the changelog. But of course a prebuild package (tagged beta 1) is available, and there are also build instructions for the more experienced users.

Sigh, as much as I love the product, as much do I wonder about their news politics. They should definitely put a little more effort into their public relations - or at least provide some press material for the interested users and bloggers to create some PR for them. And they shoudl rework the wiki in a way that the interesting information are presented to the reader, and don’t have to be searched.
Anyway, searching around around the project web page for more information also brought me to a chart comparing VirtualBox to both VMware and Parallels Workstation/Desktop in regards to the features. The list is pretty interesting and shows very well that VirtualBox supports more host systems than the other solutions. Several other features are also unique to VirtualBox or at least are only shared by one of the other solutions (also it is questionable if these features are needed by others). Other features which are not supported by VirtualBox like “3D-acceleration” are also listed.

Last but not least the list mentions a development version of VirtualBox for FreeBSD hosts. Since VirtualBox is now supported on Mac OS X, Solaris and Linux the porting effort should be manageable. It is likely that there will be soon a new VirtualBox release (1.6 or something) for all major platforms (Linux, Mac OS X and Windows) while the Solaris version and a then released FreeBSD version are available as Beta until their reach the feature level of the main versions.
Now the only thing missing is that the VirtualBox kernel module gets into the Linux mainline kernel, that would give VirtualBox another boost, I think.

Update:
Sun just bought Innotek. The announcement says Sun will continue to publish VirtualBox as an Open Source product. That might even mean that future versions of VirtualBox are fully Open Source, including the atm still closed parts as some USB enhancements and RDP features. Anyway, looks like all big companies are on a shopping spree these days.