13 March 2009

Shining Light on Why Microsoft Loves LAMP to Death

Here's an interesting little tale:

I was fortunate enough to spend last Thursday with a group of LAMP engineers who have some experience with Windows Server and IIS, and who are based in Japan.

The three - Kimio Tanaka, the president of Museum IN Cloud; Junpei Hosoda, the president of Yokohama System Development; and Hajime Taira, with Hewlett-Packard Japan - won a competition organized by impress IT and designed to get competitive LAMP engineers to increase the volume of technical information around PHP/IIS and application compatibility. The competition was titled "Install Maniax 2008".

A total of 100 engineers were chosen to compete and seeded with Dell server hardware and the Windows Web Server 2008 operating system. They were then required to deploy Windows Server/IIS and make the Web Server accessible from the Internet. They also had to run popular PHP/Perl applications on IIS and publish technical documentation on how to configure those applications to run on IIS.

The three winners were chosen based on the number of ported applications on IIS, with the prize being a trip to Redmond. A total of 71 applications out of the targeted 75 were ported onto IIS, of which 47 were newly ported to IIS, and related new "how to" documents were published to the Internet. Some 24 applications were also ported onto IIS based on existing "how to" documents.

So let's just deconstruct that, shall we?

A competition was held in Japan "to get competitive LAMP engineers to increase the volume of technical information around PHP/IIS and application compatibility"; they were given the challenge of getting "popular PHP/Perl applications on IIS", complete with documentation. They "succeeded" to such an extent, that "71 applications out of the targeted 75 were ported onto IIS, of which 47 were newly ported to IIS".

But that wasn't the real achievement: the real result was that a further 47 PHP/Perl apps were ported *from* GNU/Linux (LAMP) *to* Windows - in effect, extracting the open source solutions from the bottom of the stack, and substituting Microsoft's own software.

This has been going on for a while, and is part of a larger move by Microsoft to weaken the foundations of open source - especially GNU/Linux - on the pretext that they are simply porting some of the top layers to its own stack. But the net result is that it diminishes the support for GNU/Linux, and makes those upper-level apps more dependent on Microsoft's good graces. The plan is clearly to sort out GNU/Linux first, before moving on up the stack.

It's clever, and exactly the sort of thing I would expect from the cunning people at Microsoft. That I understand; what I don't get is why these LAMP hackers are happy to cut off the branch they sit on by aiding and abetting Microsoft in its plans? Can't they see what's being done to their LAMP?


Anonymous said...

Not everyone cares about the future. What they are doing may not be that bad, because IIS will still not be as stable as Apache. These developers need to feed their families too you know :)

Glyn Moody said...

Well, it was a rhetorical question in part. I think that they're simply not thinking through their actions - they're just having fun hacking. But occasionally, we *do* need to think about the future...

Anonymous said...

When did cross-platform support ever make an open-source project weaker? Firefox gets more core work and extensions because it has a lot of non-Linux users. And some companies ruled out git, even for Linux work, because of problems with Windows support -- now that it's there, it's getting in at more places.

And on a practical level, ease of migration is your best quality check. If there are a bunch of integration hassles keeping users in place, you don't find out about a threat until it's too late.

Glyn Moody said...

@Don: LAMP has been one of the strongest arguments to use free software for the last few years, at least in certain markets - just look at all the startups based on it. Once you port stuff to the Windows stack, you are inevitably going to get some people sticking with Windows rather than making the jump to open source.

It's different from Firefox, because Internet Explorer was dominant; LAMP has been dominant in its sector, but Microsoft's chipping away at this with the approach described in the post.

Anonymous said...

Not everyone is a Linux zealot. PHP already runs on a number of proprietary non-Windows platforms (e.g. Solaris). It isn't weaker because of it.

Chances are most of these folks write their PHP code on Windows workstations anyway. Now they can deploy on whatever server OS they want without clouding their heads with the politics of it all.

You asked the question, that's just my $0.02.

Glyn Moody said...

@Steve: well, Windows is rather different from Solaris, I think.

And I'm not saying you're analysis about coding and deployment is wrong; I'm just asking whether it's a good thing for open source as a whole.

Sean Lynch said...

"... These developers need to feed their families too you know :)"

Owen made a very good point. Now that these developers are creating software that has licensing costs attached to it, and are forcing their customers to take on the additional burden of software licensing audits will the developers still get paid as much. If the customer budgets X dollars for the job, and Microsoft increases its percentage of those dollars, will developers have to eat the difference?

With the economy the way it is right now I think the smart solution is to stick with the tried and true LAMP stack and leave the experimentation with IIS for University research projects.

Anonymous said...

It could have the opposite effect as well.

Think about it this way: As a IT Consultant in a Windows/IIS house I now have access to 70+ open source apps that will run in my environment. This makes it easier to sell open source software to the execs, and bring it into our environment.

I don't have to replace entire infrastructure to roll out a web app anymore.

This makes it easier to bring in Opensource in a slow way, and not having to retrain everybody on Linux as well.

This could in effect bring the message an philosophy of Opensource to more people - suddenly they can modify the apps to do exactly what they want to do.

Just look at how the French Police are handling the migration to Linux.

First they introduced Firefox, then Thunderbird, then OpenOffice all running on Windows. And now they are changing the desktop to Linux with minimal disruption.

With that being said: Anything that Microsoft is doing should be watched like a hawk. With their Embrace, extend and extinguish philosophy one never knows. At this moment all they might be busy with embracing opensource...

SouL ReBeL said...

Basically with MS you can be sure that anything they do, they do it to damage the competition.
That's just the way it is.
They have been trying to reap the open source fruit for a while now.
It's a shame we don't have a license that says "may be run only on open source platforms", for some cases.

Anyway, and I firmly believe in this, the best web developers prefer Linux, because of its low requirements and footprint.

Glyn Moody said...

@Gerard: I think LAMP is different from Firefox and OpenOffice. The latter are all underdogs in their sectors: getting as wide as base as possible is clearly sensible.

LAMP, by contrast, dominates the sector of Web 2.0 - practically every famous name (except MySpace) runs on it. This means it has more to lose by this kind of porting.

Glyn Moody said...

@SouL ReBeL: we can't ban it, and I don't think we should. But that doesn't mean we need to help Microsoft siphon off users from the LAMP stack.

kozmcrae said...

@ Gerard Korsten

It only takes one application to keep a company locked in to Microsoft's proprietary environment. Companies are free to use Open Office but they continue with Microsoft's Office because "that's what everyone is used to" or "the legal department needs it". I think it's safe to say that Microsoft Office will never be Open Source.

Anonymous said...

To answer Glyn's 5:38 question "I'm just asking whether it's a good thing for open source as a whole.". Surely this true simply by the definition of open and free software. The licensing explicitely states that you can do whatever you like with the code, so from a practical point of view why not run it on windows?

Those projects which have been ported to IIS can now claim to run on all common platforms. If theres a company behind it they've just gained a huge potential market. As for those running open code on windows, they're now one step closer to making the complete switch.

Good news all around if you ask me, there's nothing to worry about

Alex Wauck said...

If the platform is good, you don't need to discourage people from leaving. If we needed to withhold Windows/IIS compatibility in order for LAMP to stay afloat, that would suggest that LAMP is flawed. That's why MS Office is not available for Linux. Microsoft knows they would lose Windows users left and right otherwise, especially businesses.

Realistically, though, which current users of LAMP want to switch to Windows/IIS? I seriously doubt there are many. This initiative can only hope to keep current Windows/IIS users on Windows/IIS. It poses no threat to LAMP.

Anonymous said...

I don't think it will siphon off anyone already running Linux. Who's going to drop their Red Hat license and sign up for Windows Server 200x? More likely it will expose Windows only shops (and there are still a lot of them) to WAMP and they will find the jump to LAMP a much smaller step.

Unknown said...

I get your point though many won't. The threat here is that Linux could be abandoned to the point of crumbling if everything ran well on MS and they promoted it. But once Linux is gone or in a terribly weakened state all MS has to do is go back to crippling these technologies on their OS. Why do you think they would want LAMP to run as well as ASP.Net? That would be a lost sales of Visual Studio.

Now I don't think it will happen that way because of the nature of FOSS. I don't think Linux would grow weak or die. But its definitely worth thinking about standardizing on a OS platform from a company that has an interest in seeing your application platform of choice fail. MS rips the rug from under their own community (ASP -> ASP.Net, ASP.Net 1.1 to ASP.Net 2.0 etc.). So why wouldn't they do it to a technology thats not their own.

Glyn Moody said...

@ABCC: I'm not suggesting we should actively try to stop someone from porting stuff, just wondering whether it's wise to accept Microsoft's offer to do so now in this accelerated way - 45 apps at a time.

Glyn Moody said...

@Alex, @moopst: I should have made clear that I'm not worried about losing people to Windows so much as losing the next batch of companies that want this kind of functionality, and that might have gone to LAMP in the past.

If it's readily available on Windows, they may decide to stick with Windows. Similarly, startups might choose to go with Windows (well, there are strange people out there...)

So it's about losing potential converts, rather than about losing existing users.

Glyn Moody said...

@Phillip: yes, that's right. Another fear is that at some future point, once PHP and Perl get used to running on Windows/IIS, then Microsoft will be in a much stronger position to suggest "improvements" that fork the code and generally weaken the open source alternatives.

And your point about Microsoft's competitive tendencies against its own software partners is spot on.

David Gerard said...

Speaking as someone who's done this a lot.

1. It's good for the apps, because cross-platform means a more robust, more cleanly layered and deeply better program.

2. Users want computers for applications. If they can get the same app on two operating systems, the one without enterprise volume licensing has obvious advantages.

3. (part of 2) This allows people in idiot companies that mandate Windows to run open source apps; (a) this proves open source apps are good, and (b) they can triple performance on the same hardware just by swapping the OS out from underneath and look like rock stars. I have done this. It really works.

All these three are real life phenomena that actually happen.

I can only presume that Microsoft are hoping against hope that the theoretical fears you put forward will come true, because real-life experience says they're shooting themselves in the foot in a freely reproducible and modifiable fashion. And here's to them continuing.

Anonymous said...

Glyn, you are being too generous here. Anyone doubt that "killer apps" help sell a platform? FOSS doesn't have lock-in. We don't need to help the Monopolist that leverages their position to snuff the better software. The more software that runs on their platform, the less likely people will work to break their lock-in and deal with broken interop issues that will arise. If MSOffice was the only thing holding people back, well, they would not be held back, they'd start to migrate permanently. "Everyone" uses MSO for many reasons.

In a tug of war, do you want to give people over to the other side? I mean software as well as human developers of course. Remember who the monopolist is here. Who is the one that leverages to beat out superior competition.

FOSS running over Windows is not FOSS. You can't control or know what Apache and the other apps do or how robust they are etc because they rely on the OS for so many services including timesharing, hw interrupt servicing, memory allocation and copying, process creation AND software running on the high privilege ring on Intel chips determines what the other software does. The OS can butt into the memory maps and registers and everything and do what it wants. It can intercept calls not even to the OS. FOSS running on Windows is like HTML running on Internet explorer. It isn't FOSS. These people were working on proprietary work.

A "contest" also gives the impression the people were working hard for little money.

Microsoft is desperate that their platform will become stale with so much action going on with FOSS. They can't compete with variety, quantity, and quality, long-term, despite their lock-in. But they can compete if Linux is only a little better and too much trouble to interop.

I suspect at least one or two of the people posting here know very well what is going on and favor the Monopolist.

Anonymous said...

Remember it's about numbers. It's not one app or one level of lock-in. For Monopolysoft, it's about having as many intertwined layers as possible to thwart interop and escape. How many people have made an effort to migrate but perhaps found, in their case, things were just a little too difficult? The amount of lock-in matters.

Other proprietary vendors exist, but they are not Monopolysoft by a long shot (in terms of aggression, thwarting of interop, insistence on total control, etc). Someone mentioned Solaris. Besides Sun not being a monopolist or like Monopolysoft in many ways (including levers), there is also an OpenSolaris .. umm, OpenWindows?


MS service providers will be able to service Windows for some time to come as legacy. Linux is getting so easy, that I don't even see most wanting to do so at some point in time. It's better to work on fresh projects or have time to yourself than to have to be running around like chicken without head attending to Windows. I'm sure some MS supporters (eg, within companies) are a little fearful, today, however. That is normal, but many of us started on Windows. There is life after Windows, and that is more true today than it was years ago. It will be more true tomorrow.


Appealing to people's sense of fairness aside, lock-in is no joke. It is expensive (time, dollar, control, etc) and frequently means some amount of loss of documents and material. You are a slave to the whims and business needs of Monopolysoft. You ride their treadmill. Starting from scratch on new platforms and paying more according to their calendar. They clearly don't respect you very much when they go out of their way to snuff superior alternatives. There is no second source.

Many people are surprised once they get accustomed to Linux. All change takes a little bit of time, if you are heavily accustomed to a different routine, and tends to be scary.. that's just a fact of human nature. Change takes time and sometimes people will kick and scream to avoid it.. but that difficult moment passes.

Anyway, tomorrow, there will be a huge service industry for Linux because third-parties (everyone in the "community".. all users) can do so much with Linux that they can't do with closed opaque Monopolyware. It's good to see so many countries now giving FOSS the priority it deserves.

It's noticeable how Linux improvements speed up as the community grows. I must say that Monopolysoft, the incumbent, doesn't shy about pulling out all the stops. How long can they keep increasing their expenses fighting Linux? We'll see.

Glyn Moody said...

@Jose_X: thanks for those comments.

Glyn Moody said...

@David: all good points - I just hope you are right.

Anonymous said...

What a lot of comments. p

David Gerard said...

@Glyn - Jose_X and I have disagreed on this point before, but I will still put my - admittedly anecdotal - experience against theoretical worries.

Another good example: the Eee with Linux. Could users tell Firefox on this from Firefox on XP? No, because there's no difference. So Microsoft had to resort to blatantly dumping XP at $0-5 a copy, to such an extent it trashed their bottom line numbers for January. Amazing what actual competition can do.

For the fears of Microsoft dominance from this to come true, a ridiculous number of other bad things would need to happen first - patent lockin, free (libre) software no longer being even possible to make gratis, CentOS/Debian/Ubuntu becoming legally questionable to give away ... all of which are much bigger worries than this one.

So no, I really don't know what Microsoft are thinking they can actually do to leverage this one.

Glyn Moody said...

@David: I'm a big fan of the Eee PC, too, but as I've written above, I think this is the other way round: open source trying to gain a foothole where MS is strong.

The LAMP stuff is where open source is *already* strong, so the danger is that it may lose out from this move.

Just a thought.

David Gerard said...

I do see the point. OTOH, I can't think of any way that trying to stop people from porting the stuff - legally or by social approbation - wouldn't be morally questionable.

The definition of free software includes that appallingly neutral phrase "for any purpose". Note the lack of "but not for evil people."

OpenBSD puts it even more bluntly: "But software which OpenBSD uses and redistributes must be free to all (be they people or companies), for any purpose they wish to use it, including modification, use, peeing on, or even integration into baby mulching machines or atomic bombs to be dropped on Australia."

Basically, free software is fundamentally better ("open source"; science rather than alchemy) as well as more ethically justifiable. So we can't actually lose. Trying to pressure people not to port things to Evil OS may or may not be a tactical win (I don't think it will) but it will be a strategic failing. Either our stuff is better or it isn't.

Glyn Moody said...

@David: of course you're right, and I'm not suggesting stopping people porting stuff, which we can't. But the situation here isn't some hackers who have a burning desire to port this to Windows/IIs, but a bunch of hackers who are taking part in a competition organised by Microsoft.

That's slightly different, in that Microsoft obviously sees an advantage in doing so...which makes me worry that there might be a disadvantage from an open source viewpoint.

David Gerard said...

"If Microsoft likes it, it's probably bad" is indeed a useful and generally correct heuristic.

But it's only a heuristic. As closely as I examine this particular example, I still can't see how it works for them without a large number of much more significant worries coming to pass first.

I still have to ask "what on Earth is their plan with this?" I can't think of one that actually joins up. None of the hypotheses above seem to join up without things like the much worse things I listed happening.

David Gerard said...

And there's obvious implicit pressure on people not to port software to Evil OS; I don't really see how you can say there isn't. There's already a (IMO fallacious) meme that porting open source to Windows hurts free software (rather than, as I think is widely seen in practice, acting as a gateway drug).

Glyn Moody said...

@David: well, it will be interesting to see how this develops. I predict we're going to hear about many more of these "competitions" and their like....

David Gerard said...

I assume noteworthy performance discrepancies will be taken as bug reports, much as the Mindcraft reports were.

Really, if open source wasn't a fundamentally better way to make software, free software would have been dead on arrival.

Glyn Moody said...

Good point about Mindcraft.

Anonymous said...

David, sorry, I took so long to get back here. Maybe I'll bring this thread up when I see you posting on boycottnovell.com.

I'll happily disagree with you again. Monopolysoft was being helped. The surest and fastest way to break Monopolysoft's levers is to shun their platform. Every time we code to Linux standards (open standards defined in part by open reference implementations) and integrate with Linux, Monopolysoft must spend extra dollars to port. This creates ongoing costs for them to maintain their own distinct set of standards. More importantly, to break their levers, you want to do what you can to communicate and convince users to keep Linux around as a hedge to closed monopoly platforms.

>> Really, if open source wasn't a fundamentally better way to make software, free software would have been dead on arrival.

We each agree with this, of course.

The difference is that I recognize the field is biased against open source, so I believe we should want to make it costly for Monopolysoft's platform to gain FOSS apps (or at least as costly as would be the case when we code exclusively for Linux/LSB/etc). It's ridiculous to give freebies to the player with the advantages.

You apparently also believe the field is biased against open source, but you don't seem bothered that developers contribute to keeping Monopolysoft's platforms useful.

Remember, Monopolysoft will likely leverage BSD. Linux+FOSS must not just be better, but significantly better in order for users to be motivated to change and to overcome various obstacles Monopolysoft puts in the way of Linux+FOSS adoption (eg, lock-in and market leverage).

I don't know if you mean well or not (wrt Linux+FOSS), David, but refusing to address "theory" while talking vaguely about anecdotal evidence is not very convincing. There is also anecdotal evidence and a whole lot of other types of evidence to suggest the "theory" is not too far off the mark.

The playing field will be a little bit better balanced when Monopolysoft's closed platforms are removed from active use. Linux will face a lower onslaught of obstacles when Monopolysoft's money supply takes a real hit.

David, I also noticed you didn't acknowledge above how easily it is for Microsoft to create interop problems with whomever threatens them. I think we may have butted head on this issue as well during the time I was on the OASIS ODF OIIC-formation-discussion list.

>> Another good example: the Eee with Linux. Could users tell Firefox on this from Firefox on XP? No, because there's no difference. So Microsoft had to resort to blatantly dumping XP at $0-5 a copy, to such an extent it trashed their bottom line numbers for January. Amazing what actual competition can do.

Are you implying that if Firefox didn't exist on XP, people would not buy an inexpensive useful product with a usable interface, web browser, and other apps?

In fact, if firefox did not exist on XP, the Linux netbooks would have an extra attraction. The more apps that worked much better (or at all) on Linux, the better that would be for Linux. People might even buy 2, but what we see when all interesting apps were to run on Windows is that many more people would just buy one.

To the extent Microsoft needs firefox on XP, they will not toy around too much with it, but it's rather easy for Microsoft to have firefox be degraded. Microsoft degrades threats on their platform.

Mozilla should put their efforts on firefox designs and debugging that optimize in all possible ways for the Linux architecture rather than the Windows arch. I know Firefox comes from Windows before Linux and that they want market share and many of the devs have experience with Windows, but focusing on Linux would help Linux and FOSS apps (like firefox) more down the line.

David Gerard said...

@Jose_X - I do think they wouldn't regard said web browser as being as usable if it weren't identical to something they were familiar with.

I do think Firefox on XP is a gateway drug to open source in general, as is GIMP - people who are sceptical about open source suddenly seeing that it's good are IME happier with the idea of Linux in general.

(It's amazing the effect it can have getting an actual piece of open source software in front of someone - it's concrete proof the engineering model makes good things. I was most pleased one day in 2001 when I showed my then-boss GNOME October or whatever it was then on Red Hat 7. He was literally speechless that something from a bunch of these open source weirdos could look and work that well.)

So no, I'm quite sure open source on Windows is a win for free software.

I took my comments on this post and turned them into a much longer coherent comment on BoycottNovell. Someone there suggested the aim could be embrace-extend-extinguish. Though most suitable prospects are under copyleft licenses - I'm surprised IIS these days isn't basically a hacked-up Apache.

Anonymous said...

>> So no, I'm quite sure open source on Windows is a win for free software.

Keep in mind that your main example in your comment was about GNOME and Red Hat not about winfoss.

>> I do think they wouldn't regard said web browser as being as usable if it weren't identical to something they were familiar with.

Without a doubt familiarity doesn't happen instantaneously. The issue is gains vs losses. The lack of familiarity of something that is similar to what you know or whose capabilities you have been sold on is a *small* hurdle. People face and conquer small hurdles every day/week/month/year. People do so of all ages, sizes, and sexes.. of all skill-levels. The key to being sold is: is it good.. does it do something worthwhile.. is the user motivated to see it through to the end?

The most important characteristic is motivation not 10000% familiarity.

Now, are there costs to Winfoss? I believe there are as I have been stating. If there were no costs (eg, the ports were automatic, didn't aid a monopolist, etc), then I wouldn't care.

>> I do think Firefox on XP is a gateway drug to open source in general, as is GIMP -

These two were your other examples that "foss" is liked by people; however, as the GNOME example showed, as I believe is the case, *LINUX* in action is a much better ambassador for open source than is any handful of FOSS apps running on Windows. In particular, the LiveCD/etc and license terms, demo'd and explained (with the help of youtube and any other medium), means there is no need to have (much if any) Winfoss in order to get people's attention. You yourself managed to demo GNOME/Red Hat to an influential person. We have VMs, not to mention that many people have more than one PC (the PCs from a few years ago almost require Linux to be useful today).

David, I believe you find cohabitation with Monopolysoft to be acceptable and perhaps even desirable. If I felt similarly, I would champion WinFOSS. I don't, however, because the market is clearly slanted (because of Monopolysoft's actions) against what is best for users and the economy.

>> I took my comments on this post and turned them into a much longer coherent comment on BoycottNovell


BTW, I don't mean to fight. I'm passionate about this topic. I find it sad that with all the headaches Microsoft brings to FOSS and all their work and need that a healthy Linux ecosystem be neutered, that people would go out of their way to help support their closed source monopolies and the company.

Anonymous said...

David, sorry if I almost can't keep myself from stating here and there that supporters of WinFOSS almost have to believe they are supporting Monopolysoft. I know that the most efficient way (theoretically) to reach Windows users is through Windows itself (VMs would come in handy here).

My views are principally as a developer who would not trust my time (unless I had no choice) to coding for an opaque platform whose behavior can and will be leveraged against the code I write proportionately to such code being a threat to the creators of that platform or those creators needing more money or them simply being imperfect and changing things in a way where all the key details are not revealed (no source) and my application fails to work (especially frustrating after a long time spent debugging to resolve nuances of undocumented semantics/ interface limitations from the prior set of bugs/behavior). It's just a very unhealthy "partnership" this relationship with Microsoft, much like http://www.youtube.com/watch?v=8To-6VIJZRE where Ballmer can't possibly really respect developers (a very one-sided relationship in terms of control) yet recognizes that he must talk the talk as his mansions, influence, and power depends on enough developers choosing Windows over the rest. I recognize the costs to my time and that applications do add to the value of his platform.

As a user, I would not care as much about this issue so long as Windows appeared to be cost-effective and allowed me to keep up with the (freebie) goodies of other platforms.. and I didn't have plans to leverage open source through a business (this last is applicable to business owners and service providers but not (yet) to most users).

When one looks at the work done with FOSS, why wouldn't users want these open platforms to dominate?

[Speaking as a hypothetical Windows user aware of FOSS...] Perhaps many do, but only as long as they can still have their toys today on Windows. The user finds time spent porting to be more fruitful (though they aren't doing the porting) than time spent adding another feature or redesigning. You're more likely to want "the dev" to port to Windows ("great job your are doing") than to want to re-adjust your workflow and environment to make room for Linux.

I am human. I understand. I behave this way as well. Laziness rules. Self-interest rules.

Of course, if you are sold on the apps and if they are basically only available on Linux, you may bxxxx and moan but eventually you'll overcome inertia. One reason I came to Linux years ago was because I did not imagine a comparable environment was possible on Windows [It wasn't, but if I had more knowledge and lower requirements, I may have found enough to be satisfied staying put].

I don't want to appear to be rude. Fact is that, longer term for most developers (and shorter term for those currently knowledgeable), it's more efficient and less risky to code for a totally open platform than for a closed one. There are many synergies when everyone can contribute to any part (the easier and cheaper, the better). Society will benefit more in the long run the more ways they find to support FOSS.

The best marketing move for Linux is to have it become easy to use and easy to be leveraged through source code modifications. At that point, users will have a personal stake and find Monopolyware to be unacceptable. [Eg, much organization of program behavior can be abstracted (eg, xml configuration). Additionally, projects should incorporate better docs, including greater use of rich meta-tagging and video tutorials, and better Hack the Source for Dummies tools.]

A more practical short term marketing move would be to spin unique distros for friends/family/customers/etc. Create the re-mixes (or an original distro) from info about tastes and needs of that person. Surprise them with an initial freebie and then maybe work together on something more advanced (especially easy to do if they fund some of this effort or carry out enough of the work themselves). Video a demo if necessary. Look for businesses as potential clients.

David Gerard said...

@Jose_X: A comment at http://boycottnovell.com/2009/03/14/ms-path-of-lamp-destruction/ .

As I said there, I'm still not sure what coherent plan Microsoft thinks it has that will work.

Anonymous said...

So as to create a duplicate copy, I'll just inline here my very long response to you on that thread. Many of the arguments you made there and which I address were also made here anyway.


>> 1. It’s good for the apps themselves, because cross-platform means a more robust, more cleanly layered and deeply better program. IBM’s work to port Apache to NT did wonders for its internals, for example.

You can't design for everything and expect to gain. Some architectures should be dumped. Maintenance costs can go up and performance can suffer. Some designs are not very compatible with each other.

I do understand your point even though the details of specific cases are always debatable (I'm not familiar with the Apache case).

There are many other platforms that can be kept in mind when designing. Is this being done? What makes Monopolyware special over other platforms wrt design? Should we write code in Esperanto as well?

Performance is an important consideration. It's not crucial in all cases, but it is a an important feature to many users. Just watch how frequently it is sited as a positive (both in favor of FOSS and against (eg, OO.o vs MSO)). Overly general design will have a negative impact on performance (and maintenance). There is a reason the LSB exists. It's so that app writers don't have to design to every possible way of doing all things. Focus can be very good.

I do like the multi-platform perspective, but that doesn't mean I think we should follow Microsoft's lead and changing interfaces. It's better to look towards better design instead of trying to fit in with anything in particular. We should ask, not is MSware different, but is it better?

Let's not forget that changing designs midway is the sort of reason why Monopolyware has the security issues it has.

You can't build a palace and army barrack at the same time. Each is acceptable when built in focused fashion. Don't try to build a building that serves all purposes or you will be good in neither. It's OK to build for lots of things, but not necessarily within the same app. It's OK to build for many things, but not simply for the sake of doing so.

I think we generally do a good job, but Microsoft will only push harder for us to become like them, regardless of there being any other reason (or will fabricate reasons that have little to do with technical merits). Trying to be ultra-flexible is time not spent building utility, simplifying design, etc. We should be flexible to where we have to gain.

In the end, we have to ask ourselves, is there a positive reason to want to adapt to Microsoft's way while necessarily taking a hit in terms of costs? The design will be more complex. This is costlier. If it becomes simpler, then maybe our initial design was simply that bad, in which case, the particular design decision was a good one and was done for the right reason.

Finally, Roy had a quote on this thread: "Do encourage fragmentation of the Java classlib space." Well, Microsoft is doing it again. They are encouraging Linux devs to take into account all their dotnet garbage despite the huge overlaps with what exists. They are encouraging ports, which is nothing but an internal fragmentation of the classes and libs originally created optimized for Linux. Microsoft can use internally anything they want. Meanwhile they can expose dotnet, OOXML, and anything else with aim to achieve the greatest negative impact on those that try to follow or incorporate those design decisions. Microsoft has been known for arbitrarily making changes in things to throw off everyone else (undocumented bugs do this automatically, btw, when you control the market and people must interface with you).

>> 2. Users want computers for applications and couldn’t give a hoot about the OS. If they can get the same app on two operating systems, the one without enterprise volume licensing has obvious advantages.

So then why don't we focus on improving our applications? What does WinFOSS have to do with this? Isn't Linux free? Doesn't it run on most hardware? What is the problem with Linux+FOSS?

Cloning actually slows down as it's a look towards the past and overlooks that there may be better design decisions that can be made.

We should work on having the best of any given functionality. We only stand a chance of being in that position if we control all layers of the stack.

Here are a few reasons to avoid Windows.

Microsoft has many levers and control at their disposal. They can adopt open source much as IBM and others have adopted it and use it to lead the game their way. Their specific optimized (not general) software with lock-in features will give them an advantage *on Windows* over all FOSS apps and open source software layers lying on top. The closer the apps are in design to Monopolysoft's investments, the easier it is for them to study code and build their optimized version with all the lock-in and exclusive proprietary trimmings.

Our hope is when focused and accessing the source of all layers.

>> 3. (part of 2) This allows people in idiot companies that mandate Windows to run open source apps; (a) this proves open source apps are good,

That is done when Linux runs circles around Windows. This can be shown in many ways. We can publish numbers, videos, other marketing materials, and let the market know that the ones leading are using Linux and FOSS (eg, Google). We can do live demos. Linux is free (the cost of a download).

WinFOSS is a horrible compromise that actually puts FOSS in the inferior position.

Show our best with all the trimmings, not a compromised effect. Firefox and Apache getting hit by worms, virus, etc, shows bad for FOSS. It makes it appear that the fault is totally or mostly of the app or is an unavoidable issue.

>> and (b) they can triple performance on the same hardware just by swapping the OS out from underneath and look like rock stars. IT staff like to look like rock stars to the CxOs. I have done this. It really works.

Ah! So performance is important! Glad you also agree.

Anyway, 3x performance increase, besides not usually achievable and almost certainly not sustainable, has its best chance to be achieved with a full FOSS solution.. not with WinFOSS. Ditto for 2x, btw.

The more we accommodate Microsoft's designs, the less likely we are to find a way to stand out in terms of performance ..and hence to become Rock Stars.

Time spent on Winports and redesigns to accommodate MSware is time Microsoft works on their features and less time we do.

There is also a human tendency to lean on the platform. Linux will suffer as Windows specific hooks are leveraged by the Windows devs.

Redesigning to Microsoft's next greatest and latest (without design input and without being able to track the development) leads to more wasted time. Dealing with unavoidable bug/feature interop issues is a big time waster. Having apps break later on without warning is a time waster and speaks bad for professionalism. Getting second hand treatment as suits Microsoft strategically over time speaks bad for out ability to execute. It's also a big time/resource waster.

I'll repeat, running over Windows is not a way to try and beat Microsoft at almost anything.

Here is a link with more examples: http://boycottnovell.com/2008/11/25/jose-on-mono/

>> Could users tell Firefox on the Eee 701 with Xandros from Firefox on XP? No, because there’s no difference.

I replied directly to this wrt some other issues, here: http://opendotdotdot.blogspot.com/2009/03/shining-light-on-why-microsoft-loves.html?showComment=1237376400000#c1250602402856177969

People don't really care if something looks a little different. If you have a different look and design (that is usable of course) and people see it in action and see it is, eg, faster, this will elevate the entire platform, methodology, philosophy, etc. Conversely, if your differences perform worse, it will not help your cause (particularly true if Windows stability and malware problems and other quality negatives get inherited.. as would be the case with WinFOSS).

If you want a perfect clone, copy IE on all accounts for Linux.. or run MS IE with Linux+wine (the latest IE may not run, but then, the assumption was you were after identicalness, in which case, you are running an older IE anyway, one that may very well run with wine).

If people can change to new IE and Vista, etc, then clearly there are more important issues here than total sameness.

We can't win the lock-in game, and through levers and deals (and proprietary access), Microsoft can win the cost game as well (pre-loads deals). Look at the netbook market share evolution if you don't believe. The existing MS partners (the OEM industry) are not giving Linux a fair competing presentation, neither in hardware power, nor in pricing.

Anyway, Firefox is not about sameness or we would have cloned IE. Actually differentiating, as I stated, is a plus. Plug-ins are what sell Firefox. Allowing the user to do new things.

Variety is interesting. Don't help Microsoft along. If Linux can run a bit older IE, or if people can keep their Windows boxes around, we achieve success in the department of "similarities". But let's not help them with the more important features war.

To belabor the point further:
Features trumps performance, which itself trumps familiarity. And Microsoft can use their levers with partners (as well as the feature game through closed source exclusives) to win the full battle if we help them leverage the features of FOSS and compete on their platform.

To win the feature game, make Microsoft have to reproduce all our features -- don't port for them or make ports easy.

To win the performance game, control and have access to all software layers.. and integrate -- don't port or design overly general.

To win the similarity game, spend time cloning IE or getting wine to work for existing apps, but keep perspective. The feature and performance are more important and add more to brand for a greater number of use cases than does similarity. So if you have to pick, work on features and performance ahead of cloning. Similar enough, need not be identical.

>> I admit this is anecdotal, but all these are real life phenomena that actually happen, and I’d put that against theoretical worries.

Unfortunately, the "theoretical" side also has many anecdotal examples. The "theories" I'm talking about are based on actual things that have happened.

>> I really don’t understand what Microsoft think this will do for them.

Well, that much seems clear. I suppose it is a good thing you don't work for Microsoft because Microsoft has, for many years, been using many of these techniques and goals I have been talking about in order to successfully thwart competition and preserve their monopolies and extend them into new areas.

Not all theory is wrong. Believe it or not, you actually covered many theoretical reasons to explain your pov. In particular, it is a theory that WinFOSS does anything except waste FOSS dev resources on an imperfect uncontrollable solution.

I also think the theories you are championing in this thread are inconsistent in a number of ways.

>> A ridiculous number of other bad things would need to happen first for them to win - patent lockin, free (libre) software no longer being even possible to make gratis, CentOS/Debian/Ubuntu becoming legally questionable to give away ... all of which are much bigger worries than this one.

I don't follow your argument (not to mention I don't agree with the earlier part of your posting), but let me say a few words about these points you mention here.

Patent lock-in is something they are working hard on. Currently it is the law and has been for a while. Patent would be an easy play.

Let's keep in mind, they did not use patents in the past to achieve their successful and defend it. One very important reason was trade secret: they don't reveal the source of everything that ships (if even of anything). Another reason is because of deception in the market place. Just look at this: http://boycottnovell.com/2009/02/08/microsoft-evilness-galore/ . For example, they get people to champion Microsoft as if they were independent observers and regardless of what is the truth or accuracy (frequently in spite of it). Microsoft wins many battles because they are willing to stoop to lower levels than their adversaries. Let's not take Microsoft deception for granted.

Wrt cost: The netbook story and many other scenarios show that Microsoft can overcome costs to some degree by pulling various levers. They have extensive partnerships and current monopolies that carry much weight in determining which partners succeed or fail. The antitrust authorities are always many years late and many dollars short in trying to stop these abuses.

So long as patents affect the Linux commercial scene, as obviously it has to a nonnegligible degree, Microsoft automatically overcomes $0 to a growing degree.

Besides $0, there are patent conditions [RAND] that add a lot of gum to the gears of FOSS.

As an aside, Red Hat supporting patents to try and defend against something like CentOS would be them gaining a small sword while enabling their opponents with heavy artillery. The greatest gain to Red Hat is to continue to work against patents. That's one way to deal a massive blow to their strongest adversaries. It's OK to hedge (I expect this from a corporation) to a limited degree, but you want to continue to work against patents. Red Hat should not spurn those working against patents. I hope they don't because they probably couldn't afford that, and may ultimately short-change their stockholders. [Shareholders may not have liked Red Hat working with the FFII.. I don't know. Red Hat should try to argue their case better. They need the support of the community to survive against Microsoft's levers.]

>> LAMP is where free software is ridiculously strong and it could lose some share to Windows. However, see point 3 above for how this plays out in practice.

This much was mentioned by Glyn in that other article. I don't understand your pt "3" rebuttal though.

>> Free software is fundamentally *better* (”open source”; science rather than alchemy) - if it wasn’t, it would have been stillborn.

Let's say that FOSS will be around forever. This doesn't mean it can't be embraced and extended. It doesn't mean Microsoft will fail to leverage it to help preserve their controls and high profits. They may take a hit, but ... well.. call me greedy, but I want more than simply for them to take a hit. I want to break their stranglehold 100%.

Obstacles that will be around forever, don't mean they will play a significant role .. ever, even.

>> So we can’t actually lose.

You just argued we might be like scrapes and cuts -- around forever. Microsoft likely agrees to some extent as they have decided to "embrace" having gotten past the denial stage.

You haven't argued FOSS will flourish.

Will Linux flourish? Will full FOSS platforms flourish?

In case you also mean WinFOSS when you use the term FOSS, WinFOSS is a shell of openness just like HTML code is a shell of openness when managed by Internet Explorer. In both cases you "make calls" that get implemented/extended/etc by the underlying platform.

Tomorrow Microsoft can easily choose to implement the html tags badly (even within IE) if IE+MSwebtools+MSserver have control. This would foil all parties working on html tools, websites, browsers, etc. Of course, that day doesn't seem to be near, but surely Microsoft is working on it. If they lock in an important server market (like the web) at some point in time (Novell can really help here), they can really be deadly (if they continue to hold on to the desktop). On the other hand, on Windows, they have a much stronger position today than they do on the web.

>> Trying to pressure people not to port things to Evil OS may or may not be a tactical win (I don’t think it will) but it will be a strategic failing. Either our stuff is better or it isn’t.

You haven't convinced me.. except that, since "our stuff" does not include Windows, that, yes, either our stuff, Linux+FOSS, is better than theirs or it isn't. Linux+FOSS being better won't be enough to break Microsoft's hold, but it's a start.

And surely I don't want to give Microsoft any help.

WinFOSS is a dead end in many ways.

>> What plans could they make that aren’t self-sabotaging by the nature of open source and free software

Grabbing our features for cheap and overall having us add value to their platforms, getting us to ignore Linux more while running their rat race, making it easier for them to embrace and extend us (eg, mono)....

David Gerard said...

@Jose_X - I must admit that my current project involves porting Wine to Windows. Probably using Interix.


So that's using a proprietary Unix interface that uses a lot of GNU code and runs on Windows NT to compile and run a Linux program that lets you run Windows programs so you can run Windows programs on Windows. It's a sort of Klein bottle of thought.


Anonymous said...

David, I hope Microsoft realizes the value of what you are helping to bring about and pays you your million+ USD. By extending your work some to fill in gaps, they will have a tool to allow them to move into a cleaner design (eg, even leverage BSD as did Apple) yet remain relevant by being able to run past Windows applications.

On a positive note (from my pov), perhaps Windows devs will choose to target Wine as a cross-platform option if they really don't want to go to a different interface. Then again, many Windows devs will likely target the Win flavor based upon which target is supported by the tools they like best.

Everyone's got to pay the bills, I know. [We probably understand each other's position.] That's life. I'll hope for the best. Maybe Microsoft will implode. In a less imperfect world, Monopolysoft would not be an issue and we could all get along much better doing useful stuff.

vardomescro said...


I'm green to GNU/Linux & to programming, so my perspective is quite different from the knowledgeable and versed opinions of others.

After reading all comments, however, the one that stuck with me the most is "The threat here is that Linux could be abandoned to the point of crumbling if everything ran well on MS and they promoted it."

When I installed Debian, it didn't recognize my monitor. Once I could see, I had to fix my wireless (my sole connection to the net). Once I could get on the net, I had to fix my audio. It was a big pita.

I did it, and would again, because I knew of Debian's reputation. And it hasn't let me down. I've had sessions of installing various and sundry apps, many of them huge... HUGE, only to remove them half an hour later. I did it without rebooting once. Not once. My Debian, she purrs like a milk-fed, sun-beam napping kitten.

I also did it for another reason. I enjoyed it. An analogy I use with my wife is this: I don't know, much less care, how my car works. If it doesn't I take it to the mechanic. I could Chilton it up and get to the bottom of it but I detest the work. I just want my car to run. As I am to cars, many are to their computer.

And so the lie here is this: Microsoft just works. Slide the install disk in, push a button and your computer is usable. You don't have to peak under the hood. Those of us who know, however, realize that this is only smoke and mirrors and your computer will suffer as it lugs about that beast of an OS.

$soft could port the Holy Grail to IIS, it wouldn't matter. Their OS behaves like a house of cards and their business model appears to be based on planned obsolescence. As a community, I feel, we shouldn't care if Redwood comes, stays, lays or prays. Whatever happens, our toes are still tappin'.

As to programming, my goal is to create web-based apps. And with more and more services being offered on the web, how long can a demonstrably abusive proprietary OS stay viable? About as long as it takes for open source to remove the need to look under the hood.

Then when the hundredth monkey starts tapping away on their shiny GNU/Linux 'puter, the rest will be history.

Glyn Moody said...

@Jeremy: thanks for that interesting perspective. You're right, the strengths of Debian et al. are profound; but I think the issue is that Microsoft's tactics just make converting the world to free software take *longer*, and that's tiresome.

Yes, we'll get there, but wouldn't it be nice if it were tomorrow, not in twenty years?