Proponents of the free and open-source software development model are using the recently announced delays in the shipping of Microsoft’s Windows Vista and Office 2007 products as an example of how the company’s software engineering process simply does not work well. They are also pointing to how it stands in direct contrast with the way software gets developed in the free and open-source community, and using the delays to explain why theirs is the better choice.
That article was mostly about something different. The tendency to point to product delays as indications of process flaws that imply the superiority of a competitor’s approach at such a detached level from the development of the software in question is just propaganda. It simply encourages more propaganda in the form of equivalent attacks from others. Right off the top of my head I can foresee a criticism based upon the development time of the HURD as the kernel for GNU.
My first thought was “Debian”. Do I count this as a failure of Debian? No. Anyways, “Open-Source Advocates: Microsoft’s Development Model Is Failing” sounds very much like “Microsoft says Linux is less secure”.
My first thought was “Debian”. Do I count this as a failure of Debian?
Have they ever promised and not delivered?
In (F)OSS world it’s common to develop first and tell *after* release what’s new and or improved.
Microsoft’s Development Model Is Failing” sounds very much like “Microsoft says Linux is less secure”.
Not quite.No OS is absolute secure while humans operate them.Vista should have been delivered in 2003 but that’s hardly a development fault more a marketing/management blooper.
Not quite.No OS is absolute secure while humans operate them.Vista should have been delivered in 2003 but that’s hardly a development fault more a marketing/management blooper.
Windows Vista/Longhorn/some other name; was NEVER meant to be delivered then so please netpython, stop bringing up the same lies over and over again – just because you repeat a lie several times, doesn’t spontaneously turn it into the truth, no matter how much you’d like it to be the case.
The original road map was:
Windows XP -> Windows Longhorn -> Blackcomb
2001 2003 2005/2006
THEN, there was a move by management to rather than pushing out a new version with minimal enhancements, I don’t know who made the decision, but lets just say ‘Microsoft management’ which covers all those at the top, decided to scrap that road map in favour of pushing Windows Longhorn out to 2005/2006.
With this push out to 2005/2006, the idea, in a nutshell, was to bring an operating system that was absolutely revolutionary, threw the shackles off of old paradigms, embrace new technologies, and bring new enhancements – the general consensus was that 5 years would be sufficient time frame to bring about such massive technology additions to Windows.
What Microsoft faced was a number of problems that they didn’t count on, firstly, the code audit and pushing out of Service Pack 2; the move from baseing Windows Vista on Windows XP SP2 to Windows 2003 SP1, and finally, the feature creep that occured, and the idea that ‘oh, its ok, we’ve got tonnes of programmers and time, we’ll do it’, then I think the final nail in the coffin was also the need to modularise the development process and the code itself.
All these things were not written in the script 5 years ago, let alone saw as being things to be concerned about 5 years ago; the assumption was, development would carry on as usual, we’ll pump the thing full of features, boot it out the door, and make a speedy service pack 1 to fix up any random issues we couldn’t fix in time for RTM.
The situation has changed for Microsoft; their big customers are demanding better security, there is a less of a willingness to upgrade to the next versions of Office and Windows – in the past Microsoft could simply wheel out the products, the press would purr with delight, CIO’s would cream their pants, geeks would paddle the ground like cats, and so the sell-a-thon would begin.
Fast forward to now; and people are fairly happy already with Windows XP with Service Pack 2, and Service Pack 3 will apparently deliver even more security enhancements, with brings to question, ‘if Windows XP can do everything I want, do I really need to upgrade’? Office 2007, for the enterprise, they’ll find some of the new features very compelling, and if they have a complete end to end Microsoft workflow, the new intergration will make it even better; new computers will be coupled with the new cut down versions of Office – so essentially Microsoft will be very reliant on OEM sales of their software and enterprises towing the Microsoft line.
So the bigges threat to Microsoft and the success or failure of Office 2007/Windows Vista won’t be the threat of Linux, because lets be honest, it isn’t a threat to Microsoft given its crap-tackular hardware and software availability – no, what will be a threat is their current software line up and the number of customers right now, who are actually quite happy with the status quo.
So the bigges threat to Microsoft and the success or failure of Office 2007/Windows Vista won’t be the threat of Linux, because lets be honest, it isn’t a threat to Microsoft given its crap-tackular hardware and software availability – no,
There isn’t a threat for MS in the server room?
>Crap-tackular hardware availability?
You mean the same hardware XP is running on now?Or perhaps did you mean the availabillty of preinstalled Linux PC’s.Well we all know why is that.
>Crap-tackular software availability?
Take for example the amount of mirrors eg: Debian,Gentoo,Ubuntu,etc have worldwide.I wouldn’t say it’s insufficient.On the contrary,updating either of these mentioned linux flavors is a breeze and most of the times multiple times faster than any MS update.
Or perhaps you referred to the amount of (third party) packages/software titles avaible to the end-user.Well as you mentioned XP and the status quo i think more than 80% of the software desires can be substituted by equivalent Linux apps.As a lot of XP users demands can be met by any wellthought linux flavor.Exept perhaps for games.But than again the consoles are far more superiour and costworthy opponent to the current MS PC’s.
From a security point of view the overall not MS situation is by design more secure.Most of the apps are under the goverment of the distribution instead of all the third party vendors that make up your computer configuration together with any MS OS.Most people use the PC as a tool for entertainment or their whatever profession and don’t have the time nor the desire to continously monitor the vulnerabillity lists.Needless to say that one central repository that covers more of the overall code (OS and Apps) is better than only MS update and that’s more or less it.
The situation has changed for Microsoft; their big customers are demanding better security, there is a less of a willingness to upgrade to the next versions of Office and Windows – in the past Microsoft could simply wheel out the products, the press would purr with delight, CIO’s would cream their pants, geeks would paddle the ground like cats, and so the sell-a-thon would begin.
I’m glad they are more security and or resources aware customers.Makes (real) innovation more the prime than outstanding marketing alone.While Linux isn’t allways the only (F)OSS key to more secure and costworthier success there’re increasingly examples of OSS solutions that save a lot on licences,lately.
With Linux it’s now not perse anymore a question if it works but what now.And maybe not every professional branch can convert completely to Linux neither it should.
no, what will be a threat is their current software line up and the number of customers right now, who are actually quite happy with the status quo.
To be honest what choice does the majority have other than using what they know is an OS.Like coca cola or pentium inside instead of <fill in your local,national ,favorite speciality>.A lot of buisinesses are trapped in the MS focussed tredmill of continuously upgrading their investments in harware,software,training,etc..
That’s indeed an factor.Also i think it plays a role that people think Vista doesn’t live up to more than 5 years of development time.After all wether you say it’s on time or too early or even obsolete it doesn’t matter.People only remember the promises and scrapped features.They want software to keep them secured productive and entertained *now* and not any joker with boring pseudo visions or some lunatic billionair jumping around making neanderthaler noise on whatever stage.
Fast jump to the beginning of your reply.You suddenly mean Vista is in maybe coming to early?
Heh, i’m looking for the day the Chinese start selling cheaper lenovo machines preinstalled with either linux or just bare en masse and dump them in the west.That will be the day.The choice to buy any laptop,desktop without any OS preinstalled.
> Have [Debian] ever promised and not delivered?
You sound rather defensive… did I hurt your feelings? Actually, I said *nothing* about promises and delivery. Debian had delays which they themselves called unacceptable. I repeat: Do I count this as a failure? No. Vista also has serious delays. Do I count this as a failure? Certainly not. There are other, more interesting points to look at and judge it by: Features, stability, security, performance, licensing…
> > Microsoft’s Development Model Is Failing” sounds very much like “Microsoft
> > says Linux is less secure”.
> Not quite.
You misunderstood me. When one side tells you how bad the other side is doing, the only thing you should do is ignore it. Such reviews are hardly objective.
You misunderstood me. When one side tells you how bad the other side is doing, the only thing you should do is ignore it. Such reviews are hardly objective.
Agreed,or take it with a grain of salt so to speak:-)
Unless however there is some constructive debating based on undeniable facts,how bad the outcome might be for either side.
My first thought was “Debian”.
I think the Debian philosophy during the Sarge development is to make sure everything is stable before actually releasing it as “stable”. Sarge was actually quite stable and used by most Debian users during much of its development.
!!!
— “My first thought was “Debian”. Do I count this as a failure of Debian?”
Debian is not announcing a ‘Magic’ OS during years to not deliver later!
Also, the Ms production model, financial model and ethics are completely different. That means that if Ms do not deliver what is promising, and anouncing as the ‘Miracle OS’, actually, Microsoft’s Development Model Is somehow Failing!
And this fail leads to financial loses, that at the same time lead to more development fails…
For Debian, a delay in a release of the OS would not represent the problem that represents for Ms, precisely because of its different production model, financial model and ethics.
— “Microsoft’s Development Model Is Failing” sounds very much like “Microsoft says Linux is less secure”.”
Not really. It is a kind of propaganda, yes. But, still, because of the differences that I have just exposed (completely different production model, financial model and ethics), it is somehow true!
!!!
Edited 2006-04-06 01:15
That article was mostly about something different. The tendency to point to product delays as indications of process flaws that imply the superiority of a competitor’s approach at such a detached level from the development of the software in question is just propaganda. It simply encourages more propaganda in the form of equivalent attacks from others. Right off the top of my head I can foresee a criticism based upon the development time of the HURD as the kernel for GNU.
Well, you’re on the right track, but I think that the article was more of a sad attempt to justify the supposed ‘superiority’ of the opensource bazaar model over the more traditional cathedral closed source model.
The flaw isn’t in the business model, where the flaw is, and being correct in Windows, was how Windows was being treated internally, and how teams are structured; as interviewed in the MSDN/Channel 9 interview with some of the lead programmers in the Windows kernel coders, Windows used to be treated as ‘one big binary’, however, they have found that the development model was prone to scalability issues.
You could say its the equivilant of having a very large country with a large population, and trying to govern it with a strong central government – you’re going to face difficulties.
Microsoft has acknowledged this scalability issue, and are starting to break down Windows to defined and isolated parts – over the last year they have mapped out Windows, and gradually breaking down the parts into smaller, more isolate modules so that each module can consist of its own team, that team can then take away the code, make changes – as radical as they wish, and know that when they remerge their code back into the maintree, the whole trunk isn’t going to fall over because of issues.
So in a nutshell; Microsoft have acknowledge the issue with the way Windows is developed, and has actually taken steps to address those issues – just because they don’t run down mainstream with streamers out their behind declaring their changes, doesn’t mean that they’re not occuring as we speak.
As for the ‘changing of the guard’ at Microsoft – with the Office manager being pushed over to Windows; firstly, the delay of Office 2007 is probably more of a marketing decision rather than actually anything of technical merit – the Office division is actually one of Microsofts strong points (hence, when breaking up was put forward by the DOJ, as I said to a mate at SUN – you’ll see the ‘office company’ take off, and the operating system side of the equation, flounder), so the likely hood of delays in Office 2007 are non-existant; like I said, the release date for Office 2007 was probably pushed back due to marketing.
As for the new manager, I would say that what needs to be done in the Windows division is this; continuing along with modularising the development process as well as clearing defining exactly what features are aimed at what delivery date; the problem with vista as I’ve seen (since the first leak), they never really put into stone what was going to be delivered.
Managers came out and spoke about this and that, promising technologies, hyping things, but never really said, ‘this is what will appear in the RTM, this will be seen in the first update….’ – have a rigid time line of when technologies were to reach stability, merging and pushed out to the testers, then ultimately RTM’ed.
Mozilla suffered the same thing for a while, but the success of Firefox shows what happens when you bolt down a strong target list of required features vs. desired features vs. possibly in the now, but some time in the future features.
So it would be simplistic to announce, as the individual did in the linked article, to claim that ‘Microsofts model is failing’ when clearly what he is comparing is oranges and apples.
As for the ‘opensource development process’, yes, there are some very good software titles out there, developed in the opensource way; Netbeans, Apache, GCC, Xorg and the likes, but at the same time, there are horrific examples of how, when you have software controlled by programmers, the outcomes are not consumer focused, but rather focused on the cooolness factor – case in point would be GIMP and the use of MDI rather than using the commonly used mainstream approach of MDI like how Photoshop is designed; or when things start to get boring, the project is simply abandoned – Inkscape, passepartout (Quark Clone), Pan, gtk on MacOS X natively (on Aqua).
The grave yard of abadoned code stretches for miles, and the number of crappy, poorly though out opensource programmes seem to go on for even longer; I think what the chap from FSF (Eben Moglen) needs to get out a little more and see the carnage in his own backyard before coming out to the front to declare that the FSF is superior to anything else out there.
So maybe they should just GPL Windows and Office. The value of the products would immediatley drop to $0, but at least they could get them out on time.
I doubt that. Apparently the code is very big and complex, and different licensing terms won’t change that.
A mess is a mess no matter what kind of mess it is.
I agree with you that a mess is a mess, but the post isn’t about “licensing”. It is about “development models”. (some licenses are more compatible with certain models, but that’s not what this is about, either)
I believe they are commenting on the modular approach of “do one thing, and do it well” rather than trying to make one large entity that performs all those desired functions.
One group makes bricks, one cuts lumber, one makes mortar, and together they can build a great house (sometimes they can make a crappy one, too!). But, if you have a large group trying to form a house as one, without those little autonomous specialists of small items, you can run into difficulties as things get more complex.
I think that is what the point was about.
In that case I’ll agree.
The bazaar is great, but if there were as few of them as their are MS devs I think they’d be way behind.
I’d blame Microsoft’s rush to be ahead actually. Their desktop system is a nightmare, they’re fighting issues with a system where there are so many interwoven dependencies that even Microsoft can’t streamline the system and know it’ll work.
They’re playing cleanup. They had a weak foundation and it couldn’t have anything else piled on top of it. It’d probably less their development method and more their goals which need to be looked at.
Microsoft should be striving to be building better software, not more advanced software. Robustness should be job #1, not just half of it (Security).
Job #2 should be quality of design, extensibility, that sort of thing.
And job #3 might be worrying about features, and being ahead of the game.
And of course, cutting the ambilical cord and writing quality software that helps their end users instead of cozying up to the developers with endless backwards compatibility might help a lot too!
What MSFT should of done with Vista. What they said they were doing but now aren’t is the key issue. Vista should of been released 2 years ago. It would of been the equivalant of OS X 10.0 A completely redesigned product that still needed fine tuning. Backwords compatiblity should of been managed through either an ehanced VPC implentation, or like wine/Classic mode A loading of the libraries.
oh well I can dream can’t I?
OS X was actually a redesign though. Vista isn’t nearly that big a change.
OS X had a rewritten interface, a completely different OS under it, and the general UI had even changed.
!!!
“The bazaar is great, but if there were as few of them as their are MS devs I think they’d be way behind.”
Nice SOFISM… LOL!
Well, if there were as few of them as their are MS devs, then it would not be a bazaar, isn’t it???
Being a lot of them and dispersed is in the essence and nature of the Bazaar. In a real Bazaar, and in the Free-OpenSource Bazaar model of Software production.
Therefore your sentence, although very nice, has not sense at all!
You can not compare a Bazaar pattern, that would not be a Bazaar with the Ms pattern, then…
!!!
Uh, no, you’ve completely missed the point of the essay and I suggest you re-read it.
http://www.catb.org/~esr/writings/cathedral-bazaar/cathedral-bazaar…
Many developers is only one point. And there are plenty of people at Microsoft to qualify as many, plenty. However, things like early and frequent releases are missing, customer testing, and small patch contributions are all missing.
There’s no spot where everyone discusses things, users yelling at developers, developers yelling back, etc. It’s a group of professionals getting paid to develop code in the manner which the company tells them, and it’s in no way bazaar; in fact, a lot of software houses come to Redmond to take notes on how Microsoft manages its projects.
!!!
— “Uh, no, you’ve completely missed the point of the essay and I suggest you re-read it.”
Well, no I have not missed the point! I have read the essay several times, and I know it.
Maybe you do NOT want to recognise that you have written a “Sofism”, but is a fact, as I demonstrated in my previous post (I said: “Being a lot of them and dispersed is in the essence and nature of the Bazaar. In a real Bazaar, and in the Free-OpenSource Bazaar model of Software production”).
Now you answer me with this:
— “Many developers is only one point. And there are plenty of people at Microsoft to qualify as many, plenty. However, things like early and frequent releases are missing, customer testing, and small patch contributions are all missing.
There’s no spot where everyone discusses things, users yelling at developers, developers yelling back, etc. It’s a group of professionals getting paid to develop code in the manner which the company tells them, and it’s in no way bazaar; in fact, a lot of software houses come to Redmond to take notes on how Microsoft manages its projects.”
??? Well, I agree with all this, but this is not exactly what you said in your 1st post.
You said exactly: “The bazaar is great, but if there were as few of them as their are MS devs I think they’d be way behind.”
And I answered you: “Nice SOFISM… LOL! Well, if there were as few of them as their are MS devs, then it would not be a bazaar, isn’t it??? etc.”
Besides, I don’t mind if “a lot of software houses come to Redmond to take notes on how Microsoft manages its projects”.
Sure they go. But it does not mean that the OpenSource model of production is worst tha the Ms one.
By the way is noticeable that you say “COME” to Redmond, instead of “GO” to Redmond. Are you in Redmond? Do you work for Ms? …Hmm strange!!!
The Open Source model has already proved it’s validity and sucess, and that in many ways is superior to the model that closed source companies use. Even if a lot of software houses go to Redmond to take notes on how Microsoft manages its projects.
A a lot of Closed source commercial software houses go to Free-OpenSource projects to take notes on how they manage their projects too; and to understand the mechanism of their sucess to try to apply them themselves in their closed source projects.
What they do not understand is the real nature of the openSource and the Bazaar. So, many of these mechanisms of the Bazaar and OpenSource, could never be applied to the Closed source companies, because, not only the production pattern is diferent, but also the essence of the sotware itself, what impedes the same level of integration, production and interoperability between programers and programers and testers, and users…
Also because of the different legal nature and status and mentality of the bits produced among the two models.
For instance Ms is trying to reproduce the BitTorrent schema, or even modify the program to use it to apply the software corrections and security patches that the Windows users download, in order to save time, money and server use.
But I do not see thousands of people and companies torrenting the patches, and therefore keeping open their torrent security patches, and at the same time paying money, and wasting time and server resources to make the work that Ms should do, and the work that they have already paid to Ms to make (when buying the licences). So what MS will do is making it possible and make the future Windows torrent the patches, without the user knowing that!
In the Bazaar pattern of production and collaboration, I do not think it would happen in the same way. So if Ms would be an OpenSource company (
), it will be probably easier for them to use a kind of BitTorrent model to update Windows and to patch the security holes of windows!
It would be in the interest of the software and the users. But what Ms pretends is in the interest of the users, the software, and mainly in the interest of Ms …
!!!
Hahahaha you just can’t resist accussing anyone who you disagree with of working for MS, can you?
Pay attention, ma_d criticizes Microsoft a LOT, but is very smart about his criticism.
It’s not that the closed model doesn’t work. One could easily point to Apple, VMware, XGL, Photoshop or many other examples that show it does work.
It’s specifically microsoft who can’t make it work. They have become so big they are as bloated as the software they put out. There’s the huge MSN department, the Xbox department, the computer/hardware department, and all of the various software departments, microsoft mobile(pda dev), as well as others that I’m not mentioning.
Microsoft needs to put the “micro” back into their soft. They’re just too big to do anything meaningful.
Edited 2006-04-05 20:01
On the one hand, you have an excellent point. There are countless examples of excellent software that is totally proprietary. Microsoft got rich selling software that people were willing to pay for, after all.
But the “Stallmanists” do NOT have it wrong. The open source model cranks out an array of excellent software that boggles the mind.
For a few years after I heard about GNU software and then Linux, I thought it was a nice idea but it would never work. I was thoroughly enjoying it while it lasted, but I was sure it couldn’t support its own weight.
But, man alive! Every time I go looking for a solution, I find there is even more and better stuff than there was the last time I looked. And despite my initial assumptions, profit is being made.
^^^^^^^^^^But the “Stallmanists” do NOT have it wrong. The open source model cranks out an array of excellent software that boggles the mind. ^^^^^^^^^
That wasn’t my point at all. The OSS model does crank out excellent software. But if you listen to a stallmanist, that’s the only way to get excellent software. Thus, they have it wrong. They are trying to paint MS as the sole proprietary dev. It’s classic propaganda.
Suse 10 user here, looking forward to 10.1. Windows free is the way to be.
The OSS model does crank out excellent software. But if you listen to a stallmanist, that’s the only way to get excellent software
That’s a straw man, as RMS himself does not say that. He says that the freedom of software surpass every other aspects, as long as the software is functional of course.
The software is better in the sense that it is free. I tend to agree with this view, given my experience with software.
That wasn’t my point at all. … But if you listen to a stallmanist, that’s the only way to get excellent software. Thus, they have it wrong. They are trying to paint MS as the sole proprietary dev. It’s classic propaganda.
Oh. You’re actually making a point very similar to mine. And the author, in this particular article, is guilty of the propaganda you describe. But I think most of us eager proponents of free software will acknowledge that this is an exaggeration, probably founded on an otherwise healthy exuberance.
Actually, XGL can probably count as bazaar. Developing a starting chunk of code (even demonstratable start) and then releasing it for public contribution is not only considered bazaar, it’s considered to be one of the best ways to start a bazaar project!
But yes, the other examples are great.
Like I said earlier, I don’t think it’s their development model as much as their rush and bad start.
You’re probably right. XGL is the best of both worlds.
This article has almost nothing to do with Microsoft’s engineering processes which do, admittedly, sound less than optimal. It is entirely about Eben Moglen and the progress of the GPL v3. I got as far as Moglen’s sniffy putdown of Linus Torvalds – in the form of a backhanded compliment – and then thought “Not again”. Good luck to the gpl v3, but as long as the FSF keep giving the impression they’d like to destroy someone who’s actually achieved something and asked for almost nothing in return, they aren’t getting my respect or support.
The delay of Ubuntu Dapper Drake is an example of how the free software engineering process simply does not work well.
Please, this kind of argumentation is useless. And yes, I’m an Ubuntu user.
From that, you can determine that Ubuntu is a poor product? Or Linux is? Or Open Source development? No. Focusing on one item and drawing a sweeping conclusion from it will often lead to inaccurate results.
I know that ESR things that Microsoft has gotten too big and is collapsing internally from their own development weight (from his ibiblio blog). But I don’t subscribe to his views, either (he is a very intelligent person, but is sometimes too fanatical about his beliefs).
From that, you can determine that Ubuntu is a poor product? Or Linux is? Or Open Source development? No. Focusing on one item and drawing a sweeping conclusion from it will often lead to inaccurate results.
That’s what I was trying to say, but in a sarcastic way. 🙂
The delay of Ubuntu Dapper Drake is an example of how the free software engineering process simply does not work well.
I believe they wanted to release Dapper in a more stable state since this is supposed to be their first “stable” release.
Even if this is not the real reason behind the dapper delay, I think you still can’t say that Dapper promised but didn’t deliver (ie. WinFS)
Fedora Core 5 was also not released on time…
MS is in the pickle that it’s in because of several important reasons:
1. It was trying to juggle too many internal dependencies. For example, putting all of its eggs in the managed code basket. WinFS, etc.
2. Some of the dependencies weren’t coming together. Managed code memory footprint was too big. Perf wasn’t satisfactory. Startup costs were prohibitive. Misaligned cross-divisional ship dates.
3. Management wasn’t willing to make cuts at a point where they SHOULD have made cuts to ship. When Allchin finally did so, it was already too late. Dueling VPs were counting on other divisions to slip to cover their asses. Bad management, plain and simple.
But there’s an upside for MS. They’re changing management in a radical way. They’re realigning their dev groups to bring overlapping technologies together. They are smart people. They WILL fix this problem — particularly with Sinofsky in charge now.
Sinofsky isn’t in charge of Vista, only Windows releases after Vista.
There are a ton of smart people there though, you are right. And there are a ton of cool things being worked on. It’s just the whole matter of bringing it all together that worries a lot of people, even me.
mhm, XGL is not closed software, Apple is only succesfull because they have all the BSD-crowd to fall back on, And VMware was opening their specifications? photoshop well, I don’t know much about that one.
The ubuntu 6-week release-delay comparing to an OS that was scheduled for 2003 and has a four year delay is not completely correct. Debian doesn’t have a real roadmap, they release when they feel it is ready, which can take some time.
Apple is only succesfull because they have all the BSD-crowd to fall back on
I don’t think Final Cut Pro studio and Aperture are based on any open source products.
As for OS X, note that while Darwin makes use of portions of BSD, it is based on a different kernel and has numerous tweaks builtin to support feature built on top of it (ie. Spotlight and Quartz)
Apple/NeXT also wrote all the libraries above Darwin.
!!!
— “Stallmanists have it wrong”
Not so quick, man… Maybe Stalinists have it wrong, maybe Bushists have it wrong too…!
But up to the moment, and for the last 15 years it seems that, little by little, Stallmanists are having it right!
Anyway, BSD people are not precisely Stallmanists.
— “I don’t think Final Cut Pro studio and Aperture are based on any open source products.”
Yeah, probably you are right, And I can agree with you. But, we will never know, since the source is closed… Isn’t it ???
— “As for OS X, note that while Darwin makes use of portions of BSD, it is based on a different kernel and has numerous tweaks builtin to support feature built on top of it (ie. Spotlight and Quartz)”
Yes, You are right in all that. But you do not pretend that they have to get all the cake for free, isn’t it?
C’mon, plenty of heavy work, and millions or code lines that the excellent MacApple engineers have saved to make…
They would have never get MacOS X as it is now, if not for all this work done for them… Oor at least it would have taken them 10 years more… and it would not hve been the same, probably!
—“Apple/NeXT also wrote all the libraries above Darwin.”
Good. NeXT wasn’t Apple at the time. But, that proves that they have good and talented people, and that many of them, earn their excellent wages.
They could start to think in giving back some more money to those they have been sucking from (even if legally, that is absolutly clear), now that they have their pockets plenty!
!!!
Edited 2006-04-06 01:45
It took apple 10-15 years to finally get a version of MacOS out the door with preemptive multitasking and proper memory protection out the door, critics, pundits, columnists and consultants had been declaring the company dead since the early 90s, but they got MacOS out the door, released the iMac, the iPod and all the other Mac goodness, and have really turned the company around. We have heard similar things about MS since the delays in Vista, and I think the same thing will happen here, all those people will get a real wakeup call. People will buy Vista, mostly with new PCs, people will install it, and MS will go on. Hopefully, in the interim, GNU/Linux, the BSDs and Apple will gain market share, and lead to a more balanced marketplace, but to declare that thier “Development Model Is Failing” is just retarded.
It’s not the development model, it’s the 20 years of app compatibilty, activex controls and IE integration that has caused most of the problems with Windows, and that is the cause of the newest delay. Security and code maturity
but they got MacOS out the door, released the iMac, the iPod and all the other Mac goodness, and have really turned the company around. We have heard similar things about MS since the delays in Vista, and I think the same thing will happen here, all those people will get a real wakeup call. People will buy Vista, mostly with new PCs, people will install it, and MS will go on.
The situation here is not quite the same. Mac OS classic has an ancient kernel and the system is very unstable. Mac users were desperate for something that works (they didn’t think of eyecandies back then). What was presented to them exceeded their expectations and they were willing to put up with Cheetah’s (OS X 10.0) sluggish performance and bugginess.
I don’t think the same could be said about MS.
They already revamped the base of Windows with the stable NT kernels and replaced the aging FAT with NTFS. Win2k and XP are not crashing every few hours and customers are very happy with the result. Basically, their OS works already.
Of course, one could argue that the Windows security problems is comparable to the crash-prone Mac OS classic. However, Apple is replacing the aging OS with something that has been proven to be stable and secure, while the effectiveness of Vista’s security enhancement is questionable.
Lastly, Apple was in a desperate state of affair during that period. While their financial situation improved with the release of the Classic iMac, they are still doomed without a revamped OS. MS, on the other hand, is flowing with cash and they haven’t got any ideas of how to make use all the money.
It’s not the development model, it’s the 20 years of app compatibilty, activex controls and IE integration that has caused most of the problems with Windows
MS should learn from Apple here and use a sandboxed emulator to support the old DOS, win16 and win32 apps. By doing this, they can move forward instead of bending backwards to support old undocumented behaviours and comprise the security of Vista.
Apple solved this problem with the Classic emulation environment. Their Carbon API, while based on Toolbox, is to facillitate porting of Classic applications and cross-platform development (Cocoa must use objective-c). In fact, certain features are only available in Carbon since they don’t fit the Cocoa model.
Edited 2006-04-06 01:29
When Windows XP was released in 2001 MS said do not expect the next big MS OS release until 5 years down the road. (2001 + 5 years = 2006)so technically the delay is not going to be that great the big thing that will be hurting MS with this delay is the fact that all the PC’s sold during the holidays will not have VIsta on them and in this world people get their new window’s OS by simply buying a new PC otherwise they typically leave whatever then got on there.
When Windows XP was released in 2001 MS said do not expect the next big MS OS release until 5 years down the road
I HIGHLY DOUBT this is the case, as IIRC, that’s the time at which they launched they new subscription licensing (what was the name ? License 7 ? Software Assurance ?), with a sales pitch that customers would get discount on the next MS OS. I’m pretty sure they made everyone believe there would be some new OS before the end of the license, which would have been 2003. And in 2003, businesses (which were screwed already) had to renew such a license (I think it’s 2 years). If they knew from the start that MS would not have put out some business OS before 2006 (like you say), no businesses would have taken these licenses. And no, Media Center and Windows CE do not count, that’s not what businesses had in mind. MS diminished the backfire by putting out new Office versions.
(2001 + 5 years = 2006)so technically the delay is not going to be that great
I’m amazed there, “not that great” ?!! We’re talking in years of delay from the “initial” mi-2005 date !! Not months, years !!
the big thing that will be hurting MS with this delay is the fact that all the PC’s sold during the holidays will not have VIsta on them and in this world people get their new window’s OS by simply buying a new PC otherwise they typically leave whatever then got on there
I don’t think it will hurt MS … Vista share perhaps. But as you say, people keep what they have on their PC.
Oct 2001 + 5 = Oct 2006, not mid 2005.
> Proponents of the free and open-source
> software development model are using the
> recently announced delays in the shipping of
> Microsoft’s Windows Vista and Office 2007 products
> as an example of how the company’s software
> engineering process simply does not work well.
Really now.
I guess these so called “proponents of free software” convieniently just forgot about the various Linux kernels that have shipped a year late, or the fact that it took 10 years for Wine to get to a beta release… Or the fact that Perl 6 has been in development for what… 6 years now and is still no where even close to a release?
They would do well to look at some of their own failures to deliver software in a timely manner before critisizing Microsoft’s development model too heavily. Basically, their development model really doesn’t work any better than Microsoft’s.
Edited 2006-04-06 13:07
That’s a good point.
But theres another message that’s equally important.
Delivery dates have nothing to do with software quality.
Give me something that works. I don’t need version .01. Some of the linux software makes me laugh because it sounds like they write the first 10 lines and then release it to great fanfare.
I’m certain microsoft wanted to release vista earlier, but lets be frank: Windows xp is a very decent os. I use it at home, our users are all on xp, they like it. Vista will run a little bit better and they’ll like it too.. and we’ll run it at home. If its late, so what.
Morglum
I guess these so called “proponents of free software” convieniently just forgot about the various Linux kernels that have shipped a year late
They have not.
There is no release date road map for Linux kernel, only Linus saying some dates because people (mostly journalists) pressed him to do so.
This is very different from MS announcing some date on its road map and making it slip everytime by months.
or the fact that it took 10 years for Wine to get to a beta release
I wonder if you even understand what you’re talking about. Wine never gave any date on when it would be ready, if ever.
Wine was working way before a “beta” release too.
And Wine is not at all in the same league as an OS : it’s an attempt to provide a reverse engineered Windows framework.
But I’m not surprised all you have is red herring like Linux kernel and Wine.
Or the fact that Perl 6 has been in development for what… 6 years now and is still no where even close to a release?
And what’s your point ? Perl 5 is still developed actively and used in production since a VERY long time.
What was the date given for the release of Perl 6 ?
They would do well to look at some of their own failures to deliver software in a timely manner before critisizing Microsoft’s development model too heavily. Basically, their development model really doesn’t work any better than Microsoft’s
How is it so ? None of the projects you’ve cited had schedules, so you’re talking apples and oranges.
In case you did not notice, Wine, Linux, Perl each provided lots of stable releases in the years of Vista development, which has still delivered nothing significant.
What is even more glaring, is the fact that MS and its shills touts the ease and speed of development of Visual Studio and .NET, and how MS tools help productivity, and yet, the FOSS crowd speeds along in development with their inferior tools, while MS had to scrap its project to use mostly managed code, and is amazingly late on its next OS release.
What is even more glaring, is the fact that MS and its shills touts the ease and speed of development of Visual Studio and .NET, and how MS tools help productivity, and yet, the FOSS crowd speeds along in development with their inferior tools, while MS had to scrap its project to use mostly managed code, and is amazingly late on its next OS release.
Operating System development is a completely different ballgame. Visual Studio itself was not the problem at all. It had nothing to do with any problems Windows Development had or has. .Net just isn’t cut out to be used at such a low level for an OPERATING SYSTEM, but that does not make it inferior or any less of a framework.
Development in Visual Studio and .NET IS easy and IS fast and CAN make you a lot more productive. I’m sorry, but no FOSS IDE holds a candle to Visual Studio, yet. This coming from someone who used KDevelop for 6 months (in late 2004/early 2005).
Edited 2006-04-06 15:50
“They have not.
There is no release date road map for Linux kernel, only Linus saying some dates because people (mostly journalists) pressed him to do so.”
Actually, kernel 2.4 did have an expected release date. And they ended up missing it by about a year. And yes, you are right. There are no official release dates set for the kernel. You know why? Because it is a convinient way for them to be able to avoid the kind of criticism that Microsoft gets. Unfortunately, Microsoft just doesn’t have the luxery that open source does. Can you imagine the kind of heat that Microsoft would get if they just said “Vista will ship when it ships.” It simply doesn’t work that way when 96% of the computers in the world run your software.
“But I’m not surprised all you have is red herring like Linux kernel and Wine.”
They are not red herrings at all–especially some versions of the Linux kernel. It is an example of a product that was delivered nearly a year after it was expected to be ready.
And note that I never said that Perl or Wine had a scheduled release date. I said that they have failed to deliver releases in a timely fashion, which is perfectly true.
“the FOSS crowd speeds along in development with their inferior tools”
Sorry, I don’t consider 10 years to a beta release, or 6 years since development started on Perl 6 with no release date even on the radar yet to be speeding along.
The problem is not in the development models of either MS or open source. The problem is that software engineering is one of the most complex activities humankind has ever undertaken. And to this day, we still don’t have very good ways to control complexity. I don’t care how good your tools are. When you have a few million lines of code in a few thousand source files that have 50 or 60 dependant libraries, each of which also has hundreds of thousands of lines of code, this stuff just gets very hard to manage.
Actually, kernel 2.4 did have an expected release date
Which is not the same as a date on a roadmap as MS.
Which did not prevent any Linux kernel user/business from taking the development version and deliver a product based on it.
Which is not the same as MS, as customers have not paid a subscription license to have discount on Linux kernel when it’s out.
Which is different, because Linux “execs” did not make bold statements beforehand as to how they will be so far ahead of everyone.
And so on.
There are no official release dates set for the kernel. You know why? Because it is a convinient way for them to be able to avoid the kind of criticism that Microsoft gets
No, that’s just because it’s FOSS made by benevolent people, not by a company, but this has escaped you completely.
You apparently can’t understand FOSS, having to take it as a company to grok what it is about.
Can you imagine the kind of heat that Microsoft would get if they just said “Vista will ship when it ships.” It simply doesn’t work that way when 96% of the computers in the world run your software
96 % of the world’s computers don’t run MS software (due to the fact that MS software is limited to MS OS), far from it, get your facts straight.
I agree with the rest : deceit through delayed and delayed release dates is better, especially when businesses paid licenses to have discount on an OS that won’t be out during the time frame they paid for.
They are not red herrings at all–especially some versions of the Linux kernel. It is an example of a product that was delivered nearly a year after it was expected to be ready
You seem to have difficulty understanding the difference between “expected to be ready” and “announced date of release”.
One is your guess, the other is backed by the entity releasing the product.
And note that I never said that Perl or Wine had a scheduled release date. I said that they have failed to deliver releases in a timely fashion, which is perfectly true
Which is perfectly wrong. Sorry, even businesses have put out products based on the numerous releases of Wine. Perl have made numerous releases too.
If you really believe what you say, it only shows you really don’t understand a thing about development.
The beta status of Wine is just a facelift.
Sorry, I don’t consider 10 years to a beta release
That’s because you don’t realise what a “beta” release mean, especially in the case of Wine. Cedega, which is an old product not in beta, is using Wine, appliances are using Wine too, CrossOver Office too, …
6 years since development started on Perl 6 with no release date even on the radar yet to be speeding along
As was intended. And Perl 5 is still actively developed. You’re completely wrong anyway, you talk like lots of Perl resources were on Perl 6 (like MS with Vista), which is just not true.
The problem is not in the development models of either MS or open source
I’m not qualified to discuss that. I disagree with people saying “Linux advocates” do this or think that, like in this stupid article, anyway.
The problem is that software engineering is one of the most complex activities humankind has ever undertaken
BS.
And to this day, we still don’t have very good ways to control complexity
BS again.
I don’t care how good your tools are. When you have a few million lines of code in a few thousand source files that have 50 or 60 dependant libraries, each of which also has hundreds of thousands of lines of code, this stuff just gets very hard to manage
Which is so if you didn’t follow basic good practices of engineering, which seems the case at MS, contrary to what I hear constantly from MS and its shills : they listen to customers, they do strict QA, they innovate, they have the best IDE, …
I guess these so called “proponents of free software” convieniently just forgot about the various Linux kernels that have shipped a year late, or the fact that it took 10 years for Wine to get to a beta release… Or the fact that Perl 6 has been in development for what… 6 years now and is still no where even close to a release?
Linux isn’t about wine neither is XP only about one of it’s (third party) apps.Perl alone doesn’t make a linux distro neither does the msn messenger makes what’s XP is all about allthough some disagree.
Or the fact that Perl 6 has been in development for what… 6 years now and is still no where even close to a release?
Isn’t it continuously been upgraded and updated?
Isn’t it functional *today*?
Why would you perse need a different version number?
They would do well to look at some of their own failures to deliver software in a timely manner before critisizing Microsoft’s development model too heavily. Basically, their development model really doesn’t work any better than Microsoft’s.
Who’s they?
Name an OS that doesn’t have perl included by default!
“Isn’t it continuously been upgraded and updated?
Isn’t it functional *today*?
Why would you perse need a different version number?”
Isn’t Windows XP continiously updated with Windows update?
Isn’t Windows XP functional today?
Why do you need a different version number of Windows XP?
My point is the the criticism by the FOSS crowd of Microsoft not delivering on a timely schedule is the pot calling the kettle black. And Perl 6 is a very good example.
And to answer your question about why we need a different version number of Perl? Simple. Because Perl 5 has been stretched way beyond its limits. Features have been grafted onto it that feel exactly like what they are: A gross kludge (Perl OOP being a very good example). That’s why Perl code is so unmaintainable. That’s why even experienced Perl hackers like ESR dread having to maintain even their own Perl code. Perl 5 is, quite simply, a bad design. And Perl 6 aims to fix all of those problems. Of course, if Perl 6 development continues at the snail’s pace that it is going at right now, it will be irrelavent anyway–having lost too much developer mindset to Python and Ruby to be a major player anymore.
That’s already happened to a large extent. Perl was once the dominant language for writing Web applications. There was a time where virtually all of the dynamic content on the Web was driven by Perl. But those days are gone. And most of that dynamic content is driven by Java or PHP these days, with Ruby On Rails starting to make itself a nice little presence on the Web too.
Edited 2006-04-06 16:44
Its completely non sense reason that why MS delayed its products.
“96 % of the world’s computers don’t run MS software (due to the fact that MS software is limited to MS OS), far from it, get your facts straight.”
Yes, they do. Roughly 96% of the computers in the world are running some version of Microsoft Windows. My facts are straight. Yours are not. You are simply one of the FOSS zealots who is defending open source software past all point of reason. To you it is a religion–a religion of perfection, that is right even when it is wrong.
“One is your guess, the other is backed by the entity releasing the product.”
Both are guesses. It’s been said of software engineering, take your worst case estimate of how long it will take you to get a product out the door, and then double it.
“That’s because you don’t realise what a “beta” release mean”
As a software engineer for 20 years, I assure you I know what a beta release means. Let me give you a few key phrases “testing”, “only for early adopters”, “not intended for production deployment”, etc.
“As was intended. And Perl 5 is still actively developed. You’re completely wrong anyway”
No, it was NOT intended. It was not intended for the Perl 6 development effort to stall out the way that it has. And part of the reason it has stalled is parrot. Parrot doesn’t have its act together. You are the one who doesn’t have a clue here.
“I’m not qualified to discuss that.”
Then don’t discuss it, and don’t accuse me of not understanding software engineering. As a software engineer who has worked on both closed source and open source projects for 20 years, I AM qualified to discuss it.
And no, it is not at all BS. Software engineering IS one of the most complex activities ever taken on by humankind.
And yes, to this day, we don’t have very good ways to control complexity. Good engineering practices only get you so far. After that, you are still limitted by the available design methodologies. OOP was a step in the right direction. But at the same time, OOP has failed to deliver on a lot of its promises. And some studies have even suggested that C++ is so much more complex than C, that the overall lifetime cost of maintaining a C++ program is higher than that of an equivalent C program. So it’s arguable whether OOP even met any of its intended goals, at least in the case of C++.
Again, you are simply defending FOSS past all point of logic. Because you are so brainwashed by their ideals that it has become a religion to you. Everything out of the mouth of the FOSS leaders is god given gospel for you, even if it is totally wrong. Start thinking for yourself man.
Bottom line, FOSS has shipped more than its fair share of products very late. Linux kernel 2.4 was a year late. FreeBSD 5 was a year late. Perl 6 isn’t even on the radar yet. Fedora Core 5 was late. And I am not going to play semantics with you about “But that was a guess. Not an officially backed statement”. That my friend, is the red herring. Both are, in fact, just best guesses.
Edited 2006-04-07 13:17