Slashdot Mirror
A Continued Look at Linux vs Windows
Rogier van Vlissingen writes to tell us Paul Murphy has an interesting writeup on his blog about the continued Linux versus Windows debate with regards to some of the recent insights provided by various groups. From the article: "Disinformation comes in three major forms: innocent mistakes, intentional disinformation (aka FUD), and (self) delusion. Delusions are easily the most dangerous of these. In the IT context the most common delusion is simply that what we know is right in general or applicable to some specific issue when, in reality, it isn't. We know, and we act accordingly - with frequently catastrophic results."
Since when do blogs represent news? While blogs are interesting to read, they are by no means a good news source. Please stop allowing blogs as news sources. They are usually biased and the writers are normally amateurs spouting incompetent opinions. This person may be acclimated to the information pursuant to a linux vs. windows debate, but his blog should not allowed in this site as news. Additionally, I am getting tired of reading about this debate. If I want an opinion on windows or linux I will damn well use them both and figure out which one I prefer.
who is paul murphy and why the fuck do I care what some blogger has to say? /. has gone from a blog reporting on news to a blog reporting on blogs. Why bother?
This kind of stuff is always an oversimplification. We are going to see these things forever. For instance the very nature of this discussion is already ignoring FreeBSD, Mac OS X, etc. and thats before they even get into their arguments about why linux is better then windows.
How do you compare Linux to Windows when there are hundreds of different linux distros that do things differently as well. It seems that the authors of these comparisons don't truly understand that this question can't be answered. Yet we will continiously see articles pop up that says one is better then the other and of course it will sway one way or the other depending on which OS the person who did the study is partial towards.
Lets count the differences outlined in the article:
:). Linux wins this one.
1. apply security and recommended patches on a simulated monthly release basis;
Is there anything out there equivalent to windows update? Windows wins this one
2. upgrade the e-commerce application with new functionality at the end of each simulated quarter (i.e. change it to meet changing business requirements); and,
This shouldn't be discussed under 'linux vs windows', this is more the case of 'linux software vs windows software'
3. upgrade the core OS from SuSe 8.0 to 9.0 and from Windows 2000 server to Windows 2003/XP server at the end of the simulated year.
This would be the comparison of genkernel and the rest of the beasts in the pack with "Files and Settings transfer wiZZard"
Seems to me that the whole article boils down to 1:1
When I read the SI study, I was *horrified.* The paper was uninformative, the methodology was flawed, and the analysis was unsupported.
My favorite quote though from the article is this:
This is absolutely correct. Treat Linux as if it were Windows, or vice versa, and you are asking for real pain.
LedgerSMB: Open source Accounting/ERP
As long we're making arbitrary, over simplified judgements about which OS is supperior, why not base our decisions on their mascots? I think the SuSe iguana wins hands down. Linux 1, Windows 0
It makes real cupcakes, with a 40 watt bulb, and there's icing packets....but the secret ingredient is love.
yes. it's call up2date. duh.
Or YUM or YAST. Depending on your distro.
Let's also point out that most major Linux distros have faster patch cycles rather than a month (or two or three or more in Windows case)
Score another point for Linux. And at the buzzer it's Linux 3, Windows 0.
This is my sig. There are many like it but this one is mine.
...is more poorly constructed than the study's own synopsis (which was woefully lacking the clarity of the 10 answered questions earlier today on slashdot.)
:)].)
Read, most amusingly, the blogger's attempt to repudiate the study based upon patching. LOL.
The basic problem with any study like this is that Linux and Windows admins approach things differently. *nix setups tend to spread the workload an application stack across multiple machines and Windows admins tend to load the entire stack on one or two machines. A study tends to try and mimic one or the other (Windows focused ones pick the 'all on one' stack approach, the *nix ones [depending upon what the scenario is] tend be less monolithic on the hardware level [oooh, flexibility
Loading...
You're reading a Slashdot article...about a blog...which is criticizing a report...which is pretty obviously another paid-for-study.
Gah.
That's so many orders of removal away from meaningful content that it's amazing.
Plus, the argument is about the technical merits of Linux versus Windows. You know, I like Linux. I think that it's a pretty nifty system. But, I have to be honest. I think that the technical merits of Linux comprise a pretty small chunk of the real-world benefits it has over Windows.
I think that the biggest reason that I'd rather have a Linux box running something is just that the cluefulness factor of Linux folk tends to be significantly higher. Thus, the chance that the guy writing the software and adminning the machine actually knows what he's doing is significantly better. I know a couple of Windows hackers that I'd call competent, and one that's really good, but of all the Linux hackers I know, I can't think of even one that really doesn't know what he's doing, and most of them are extremely good. It's not that the Penguin is the end-all be-all, it's that his adherents are damn knowledgeable.
Any program relying on (nontrivial) preemptive multithreading will be buggy.
that MS is first and formost a marketing company where its second place position is heald by the legal department which also partakes in chess (the idea of sacrificing your own to obtain an advantage worth more then teh sacrifice). Third place at MS is not even innovation but rather imitation or buyout ...
When you understanding this, you understand MS. To understand MS you know that what was once something ignored by MS, then laughed at by MS and then lied about by MS.... there is something of history in teh direction of open source software.
To compare Windows to Linux is like comparing carrots to meat and potatos....
I was gunna mod this something bad, but then I thought, if slashdot is gunna spam me with these crappy articles, people should make the most of it and spam with crappy offers.
If the vast majority of (low wage) administrators are trained and have experience in, and solely in, the "Windows way," I'm not sure that allowing the Linux admins to use the "Unix way" would have been realistic. Yes, they could do it, and do a better job using the "Unix way," but that might make the study less useful and less accurately predictive given the shortage of people adequately trained in the "Unix way."
Also (and this is an honest question, I have no idea what the answer is), is the truly the "Unix way" to "duplicate your production environment on the sysadmin's workstation and debug any processes to be applied to production there before proceeding?" Is that even possible?
Only Women Bleed (Sex, Sharia remix)
I am a paid Linux consultand/admin. If I would have read what they wanted me to do... I would have said no. Methodology in supporting a linux server is all wrong. Still one admin mangaged to pull it off. He probably didnt fully follow there rules.
I've mangaged to live update a server with Fedora core 1 all the way through each core release till 4 and kept it live and running.
security updates? 'yum check-update' 'yum upgrade $X'
If you run Linux like Windows, expect Linux to have the problems of Windows too.
...Goofus and Gallant!
Goofus would rather turn on his computer and be a corporate tool for Microsoft without giving a second thought to how much richer the world would be, intellectually speaking, if everyone spent a little more time actually learning how computers worked instead of learning MS specific pointy clickety stuff.
Gallant spends time learning about how to utilize the resources in his PC as efficiently as possible, sharing his knowledge with anyone who will listen and helping people to help themselved by using Linux as the primary operating system and open source applications for true productivity.
Goofus doesn't care how much bandwidth he uses while downloading internet pr0n with his insecure P2P client that has trojaned his system and turned his system into a spam bot while at the same time complaining about how slow his system is because it's over six months old.
Gallant is a polite internet citizen. "Wow. This ISO download of Fedora Core 5 is going to take me good long time to download. I've got 25 meg down available right now, but my neighbors on the cable system might need to download some things too. So I'll lower my downstream during daytime hours to half a meg and only go up to 2 megs between 2:00AM and 4:00AM".
Goofus thinks that pirating software is cool because it saves him money that he can use to fill the tank on his gas hog SUV. "Haw haw!! Adobe thinks that we're all suckers who will pay them what they ask for their crap program! I'll show them! I'm gonna fire up Kazaa and get it for free! I'm a revolutionary who's stickin' it to the man"!
Gallant respects software licensing: "No Jim Bob. You see, even though I no longer use Windows, I am well aware of Microsoft's licensing requirements and you can't just take that copy of Windows and install it again on your cousin's PC because it's a license violation. If your cousin wants Windows XP Pro, he's going to have to buy the legitimate upgrade copy from a valid retailer".
Goofus doesn't care about other people's property or privacy: "Hey... looks like that hot neighbor Jolene's PC is accessible in Network Neighborhood. Well, well, well... Let's have a looksee at what's ono her hard drive. Oooohhh... C:\Private\JPEGs\XXX\Me, Branden and Rand Partying. That looks like a keeper"!
Gallant warns his neighbors that their machines might be insecure: "Sorry to bother you Jenna, but I noticed that your computer is readily accesible to anyone else in the apartment complex. If you want I can show you how to make it secure". Jenna: "Why thanks Gallant! I'd like that. By the way, if you'd like I could make us something for dinner when you come over. It's the least I could do". Gallant scores.
-"...bad old ideas look confusingly fresh when they are packaged as technology" - Jaron Lanier (Digital Maoism on Edge.o
Yes it is possible under many circumstances.... Your desktop has to fit close to the hard drive, memory, and processor requirements. If your database takes up a gig of ram, dont expect it to work good on a desktop with 256MB.
I copy the linux install from the server computer to the desktop computer and enable it to boot and setup the directories correctly. After that its as simple as compiling/installing the new software and running it. Load testing and border conditions are the hardest to test. After that I copy the install packages (normally rpms in my case) to the server and install them under low usage times.
Is there something like 'strace' for windows?, I have fixed problems countless times with that in Linux.
Well in our case, we have a full fledged QA environment that mirrors our production environment except for the number of app servers. It's even hosted in our datacenter to mimic connectivity.
We even restore a copy of our production database before each major release to the QA box.
Interestingly enough, we do the same thing for our few Windows servers (Navision for instance. Just did an upgrade over the weekend).
I can't understand who would apply patches to a live system without a qa run first. The other thing that bugs me is that they didn't use the same application stack across the board. A better test would have been something like WebSphere or tomcat talking to a DB2 or Oracle database. Those products would have been better tests.
The other thing that bugs me is that they did a major OS upgrade for some vendor binary. Would the same vendor binary have required a 200 to 2003 upgrade?
"Fighting the underpants gnomes since 1998!" "Bruce Schneier knows the state of schroedinger's cat"
The best performing enterprises do NOT compare themselves with others; they just keep asking the question "how can we be better?". The process of comparison is a waste of energy and an exercise in futility. Put that energy to use building the most amazing system ever created and the customers will be bashing down the door. "Build it and they will come."
I don't make predictions, and I never will.
Seen this one so many times I'm not even going to read it. Here is what it says:
Most people use windows, and are ignorant.
Linux used to be rough, but is growing fast.
Linux is better than Windows in 4 of 5 ways (take your pick).
People should use Linux.
It is now Linux's time to shine, in fact, 112% of computer users will switch to Linuxin the next 4 hour.
Saved you 15 minutes.
Go ahead, mark me a flamebait, but even I (I use OSS software and OS all the time) get tired of these repetitive and incredibly biased compairisons.
Typical slashdot-prejudice. "So it's a blog. Well that automatically means it's full of crap about the writers mood and sexual activities and his/hers dogs daily life. Oh and I didn't bother to RTFA, because I have this 5-Insightful-O-Matic which helps me to write witty and cynical remarks and get respect."
Google Search
Doesnt look like a very non-bias opinion if you ask me ..
Sure there are billions of blogs that are basically worthless. But there are also blogs that are doing real reporting, that are good sources of info.
Yes blogs are more biased. But they wear their bias openly on their sleeve. I greatly prefer that to a writer that pretends (even to him/herself) that they have no bias and writes what they think is "Objective" but always has a slant. I can read a right-wing blog and know where they are comign from. I can read a left-wing blog and knw where thety are coming from. If you range widley you can get a pretty good picture of what is going on, and a lot of interesting stories that the real media just pass right by or else make light notice of.
Furthermore blogs are often more accurate because they are (if the blog has a decent reader-base) correctly quickly. I've been involved with a few stories that have gone in the paper over the years and EVERY one of them had major facts wrong. Those are the ones I know about, how am I supposed to think that newspapers or other media get the other facts right as well?
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Fedora has a new yum based tool called pup (the joke being that it's a python program->pup.py). It does not have a notification icon (yet). It shows far less info then up2date and does not allow you to select repositories, supposedly to make it more suitable for endusers. Experienced users may prefer up2date (or running yum from the cli) but considering its purpose its a nice tool for people who just want to get regular updates installed without a hassle.
I wonder, is this an example of delusion that the author referred to? Do you really know enough about low-level kernel resource management and subsystems to judge that the unix architecture is superior? Or is this just something you "know" to be true?
Maybe I'm way off base and you have very specific reasons for believing that the unix architecture is fundamentally better. But almost everyone I talk to who complains about the "architecture" is really complaining about tools, like shells and programs (e.g., PHP, Perl, etc.), and is relatively clueless about the OS architecture (disclosure, I think the modern NT kernel is pretty well-designed, even if the default tools are sub-par.)
If you don't know where you are going, you will wind up somewhere else.
I used to have your mindset. I stopped using windows as my primary OS around 2001, and worked exclusively with Linux and Solaris.
I now do some Windows development again, and have an XP Laptop and I have to admit i'm very impressed. It's stable, fast, easy to use and with a few GPL tools installed I'm pretty happy. Visual Studio.NET is a pretty decent tool and is catching up to eclipse and netbeans.
OTOH I can't stand windows servers. SQL Server is a nightmare, they aren't easy to administer remotely and scriptability is pretty lacking. They have a place in small companies without a full time IT guy, but that's about it.
Windows has it's place, and for now that place is bigger than the place that linux has. I'm certain in time Linux will take over, but it wont happen this year or next.
I see the same problem when dealing with students who come from a Solaris or Linux background -- usually they get tripped up in IP address configuration, which is very different on Mac OS X than it is on a standard Unix system. The Mac OS X way is much more dynamic and self-configuring, but this means that essentially ifconfig(8) is only useful in a read-only mode and cannot be used to write changes.
My respect for Paul Murphy is only increasing.
--Paul
http://arstechnica.com/news.ars/post/20051117-5590 .html
excerpt:
"As one might expect, the Linux system did not even come close to stacking up to Windows Server. The "granularity and high modularity of Linux" led each administrator down a different path when issues occurred due to the ambiguity of the problem. The Linux administrators were also portrayed as being confused when updates needed to be found, and at one point, a system was rendered useless by a GLIBC upgrade that went awry. On a positive note, once the SUSE server was upgraded to version 9, everything went back to a state of normal operation.
Overall, the study displays Microsoft as king of the server hill. The 49-page study (which I managed to read in its entirety), although claiming to be unbiased, reads like a huge piece of Microsoft propaganda. The Linux administrators were portrayed as lab monkeys at certain points, whereas the Microsoft administrators came off as drones that just went out to Windows Update for all their system needs. It's very difficult to read this study without believing that an obvious bias was in place."
LWN.net has more comments on their link to the original report (http://lwn.net/Articles/160247/#Comments).
Linux is now going through what Mac OS has been going through for years:
People --including well-educated techies-- have misconceptions based upon things that happened in the past and keep those in mind for ever. For instance: about the Mac people still say "it's got weird connectors and you cannot exchange files with Windows". About Linux the same thing, people still say "you have to compile and tweak everything yourself before it works" and "no software available" for both.
Now... articles such as TFA are NEVER going to take those prejudices away. They're just plain zealotry that focuses on the wrong things. Please stop "proving" that OS A is better than OS B by comparing them. Try and focus on taking away those old misconceptions that are in people's heads.
So what your saying is that he's your average Slashdot reader.
"I use a Mac because I'm just better than you are."
I sincerely doubt that the general /. reader dos not know of the alternatives. Actually I'd bet most of us has made up our minds and are more than willing to engage in a flame war to show our view.
:p
But lets face it:
Windows vs Linux is like my butt! Its divided, and no mater how shity one side thinks the other is, it's not getting rid of it.
www.aleo.no
The part of this study calling the success of the Linux admin "unsupported" is ridiculous. It is supported, by the Linux admin and any other nix admin worth their salt. Almost every time I've had to call for support on Win and WinApps and yes even Linux, their first suggestion is the ever popular - uninstall & reinstall. Maybe the companies should hire real admins who know what their doing instead of installation jockeys who know how to use a touch tone phone. If I'm the admin of a system, all support stops with me. If I really need to call someone else in to fix MY system, I"m no longer an admin but just an operator.
l " distro of linux as opposed to something more generic and stable(the whole techie reason for linux) like gentoo? I was a staunch SuSE supporter until it started suffering from emessitis:).
On the use of SuSE......Why would you use the "use-to-be-great-but-now-has-been-ruined-by-novel
And finally my reason for linux versus windows------Applications Shouldn't Change the O/S or cause it's demise, hence different terminology "Applications" and "operating system". Windows O/S and also windows apps do way too much undocumented behind the scenes things that can go wrong. Perhaps it is not so much a problem with windows and win apps as much as it is the fault of their not fully-documenting exactly what different things(dll's, ini-files, registry entries, etc..) are, where they're installed. This info is so poorly lacking in the win environment that many times a marginally knowledgeable person knows more about a product than the people you call for support. Anyway...I digress.
--
L8R,
guitardood
For whoever cares: I've been programming(low-level machine code through current high-level langs) & administrating multiple systems on various O/S's for the last 20 years.
-- L8R, guitardood
Instead of basing your opinion of what is written on the reputation of the writer, why not think about what he has written and judge for yourself?
slashdot broke my sig
how do you really count number of patches? microsoft sometimes combine multiple vulnerabilities into 1 patch. so, naturally, microsoft patches tend to be fewer than linux ones.
No, it is not one of those stallmanian rhetoric about Linux being the kernel, but Linux is still only the kernel.
It is impossible to compare Linux to Windows because they are not the same thing!
Windows is a Windows distribution, and the only one, except if you argue that XP Home, XP Pro and 2k(3) are different enough to be considered different distributions but I don't.
Linux is a general term used to described thousands of distributions.
In a sense you could compare Windows XP Pro (just to be clear) to RedHat EL4 WS, Windows XP Home to Ubuntu 5.10 or Fedora Core 4, and the server version to server version of different distribution, because at the end of the day, in the real world, people do not build their Linux from scratch, even if they could (and this is THE point of OSS).
It is not fair for Windows, and I'm not particularly pro-Microsoft, to compare it to LINUX as a whole. When you will be able to take ALL software from one distro to another without any change, yes you will be able to. Now I'm not saying this is what has to happen, because I think people are fine with choice, but for the sake of comparaison, Windows need to be compared to A Linux distribution like Linux distribution are compared to each other, it is not about the kernel, it is about everything arround.
It might be sad for some idealist, but look at commercial OS, Windows XP and Mac OS, 95% of the time, you will download an app, any app, install it and it will just work, this is this level of usability before all the eye candy that people want. Now I know you can't expect Linux apps developper to produce ready to use package all the time, considering the mess it is out there, but something has to be done, static binary might be the way - who cares about space when you have convenience and a 80 GB drive - some kind of BETTER standardization might be another, leave the package management to those who want it, leave the source for the geeks.
Real people want what is supposed to be an operating system, a system that make the link between the machine and the apps, and yes you can call it a appliance if you want, some people do not really care about tinkering if their machine works with the apps they want. Put some standardization on document format, so everybody can access information, and leave the do it yourself computing to those who want it. How many people use a car, how many people can or want to fix it or build their own? How many people use a TV, how many can or want to fix it or build their own? How many people use a computer? How many people can or want to fix it or build their own?
Because geeks thinks that computer are easy to tinker with does't mean that people want to tinker with theirs!
>Do you really know enough about low-level kernel resource management and subsystems to judge that
>the unix architecture is superior?
The unix architecture, doesn't point to the kernel. There are many completely different kernels in the unix world. It's propably the system architecture that the gp is talking about, and yes, the tools are relevant. You see, the same tools are available for windows as well, but they are not nearly as usefull.
PHP and perl have nothing to do with those tools either, although perl is another example of the different usage i was talking about.
It is true that those tools were made for unix, not windows, so you really don't expect them to be as usefull. You would expect windows to have native tools though, which they don't, after 15 years in the market...
Washington bullets will simply be known as the "Bulle
Applying updates directly on production ?! This should not be done on whatever OS. It's not the "Windows way", it's the "stupid way". That said, even under the hypothesis that the argument is not flawed then doing things the proper way would have taken too much time, leading Linux to a greater TCO anyway. What's missing from the entry is the only meaningful study : that there is no absolute best operative system, and every single case is a different story which should be deeply analyzed by itself.
I've been an IT administrator in a company that was funded by Microsoft. We were actualy given briefings quarterly showing 'studies' that prooved that Windows was better. Kind of like what McDonald's restaurants started doing after the movie 'Supersize me' blew their cover. They 'proved' to their employees that the company is doing the 'right' thing. Pretty much what Microsoft does even for mere end-product affiliates.
.NET framework (or other development technologies built to target Windows Developers). Lovely Idea. However, The amount we 'can' know about .NET framework without referring to a hacker's manual, is basicaly the amount Microsoft want's us to know 'safely'. So that someday when we need a better solution, We need to go back to microsoft and pay more. It would be silly for such a big corporation to PROVIDE a versatile solution if it wants to make money. Why wont Intel overclock their CPU's and send them off with a bigger heat sink before marketing? The cost? (it would be a mere 5 dollars over the original). Would you pay 5 dollars extra (over a 3.4 Ghz) for a 3.8 Ghz machine?. I definately would. (Do not say it is unstable, almost all of my home pc's run on P4 3.4 Ghz overclocked systems at 4.01 Ghz safely, and I do most of my office work on them). Same reason, why would Bill Gates unlock all the possibilities of Windows all at once for the hackers and programmers to explore?. Why not keep them coming back for more.
I was the person in my IT department who suggested the team move to Linux, because I was sick of having to 'read' Microsoft manuals of their software when they 'launched' something new. It is true, Microsoft basicaly assumes that its 'end-user' even if its a Software engineer by training, is basicaly stupid. Explaining to the person who said 'Windows any one can run, linux is for specialists'. It does not end there.
Let's say (like in my case) I have a particular e-commerce solution to handle and I want my application and (OS) to be tailored to that solution. Let's also assume Windows DOES provide such a solution and it works great. Patches are seemless, updates are a breeze, I could deploy it with my eyes closed. Everything great so far. Let's say now though, my company starts dealing with another company that has a different e-commerce application working for them. Or my companies demands change. It wont be then a simple matter of 'upgrade' or 'download a patch to fix'. It would be a matter of making the program work for me, without having to pay thousands of dollars and relicensing new software?. Microsoft is basicaly a strictly 'product based business' NOT a solutions provider. There are alot of people who claim 'Microsoft has developed several seemless integration options' Such as the
If you are going to have a 'technical' debate on Windows vs. Linux, i'd pose this question: When you have a dual processor Xeon system for your main file servers, and you want to use all that processor power and high pipeline bandwidth 'only' to ensure data security and smooth retrieval. If there is any one who has worked on powerful machines and used both Linux and Windows would understand when I say that 'a trimmed linux distribution' can deal alot better with raw hardware pottential than Windows OS can.
Bottom line is, I switched to linux to 'free' my company from the Microsoft bond.
It is TRUE, given the 'right' set of solutions, Microsoft OS and Linux distributions BOTH perform well. In some situations Microsoft has a clear victory, in others Linux rules the day. There never can be ONE study of ONE solution to proove LINUX is better or WINDOWS is better. There can be common sense that says on the long run, I'd rather know what i'm doing so that I can build upon it. Rather than having to call teacher Bill Gates for help.
And you have very little of it with Microsoft. You do things the MS way, or you are fucked. This is not a 'delusion', stigma, FUD, or misinformation. It is a business model; a very succesful, and well marketed, business model. This is the core interest of Microsoft and the essence of Windows as a whole:
"How do we keep people from making a choice to use something else"
This is the thought process behind your Exchange server, Active Directory, Roaming Profiles, Office documents, OS patches, and Tech support. All wrapped up in a really sexy desktop.
Linux is about choice. Linux is about standards. Linux is about YOU deciding what's best for YOU and then having the freedom to do it and contribute back to the whole process. That is what Linux is about.
You make the choice.
Join the Slashcott! Feb 10 thru Feb 17!
You've hit the nail on the head.
/usr} that all I really needed to do was reinstall the networking stack -- just extract some files from an archive and overwrite the corrupt ones. The trouble was, I didn't know where to begin looking for what files I needed to do that with! So I ended up having to reinstall all of Windows. What a waste! That's like having a whole new fitted kitchen installed, just because the sink waste pipe is blocked!
I recently had to fix a Windows machine {beancounters run some legacy app for compatibility with group HO, we've not hacked its protocols yet} that had been hit by a virus. Post-disinfection, the network hardware was undetected. I knew {from past experience with mucking up Linux boxes in various interesting ways -- let's just say, don't ever run out of space on
The thing is, I seriously doubt there are many Windows people who could tell me just what files I would have needed to replace. There are no doubt one or two gurus out there, but I'd stake money that they also know a little bit about at least one other non-Windows OS too. You could just about train a monkey to reboot a Windows machine, which is always the first line of attack and works just too often. I've seen people reboot Linux boxes and get surprised / disappointed / angry when the problem did not go away -- well, why should it? What did you change? In fact, I would say that if rebooting a poorly Windows machine is enough to cure it, then that indicates that Windows must be losing track of its own state somehow somewhere; and doing it in enough different ways never to be really sure which is the dominant one. In any case, with the Windows box, there probably would be only one service which would need restarting; if you could even do them separately, that is.
But I don't think closed-source software vendors particularly like the idea of low-level field maintenance tools. It's like electronics manufacturers who would rather have you replace a whole PCB just because one fusible resistor has gone open circuit {like it was designed to, but it costs a few pence to unsolder it, solder in a new one and see if it was caused by a real fault or just unlucky}. These people want us to have to install a whole new kitchen for a blocked pipe, and maybe they'll try and sting us extra for their fancy KlogPruf(TM) technology while they're about it. But not too clog-proof for the plumbers and the manufacturers of drain-cleaning products {evil bodges though they be} still to turn a profit on the deal, obviously.
Je fume. Tu fumes. Nous fûmes!
Examples:
1) Build/maintain a web server that can handle 10,000 static pages a minute on a budget of $5,000.
2) Build/maintain a web server that can handle 10,000 dynamic pages a minute on a budget of $10,000.
3) Build/maintain a database server that can handle 10,000 transactions per minute on a budget of $20,000.
Then extend the maintenance of those over 2 years, 3 years, 4 years, etc. Include hardware/software upgrades.
Each team will come up with different approaches. I'm sure everyone here knows about the MindCraft "study" where a single Windows box with 4 processors and 4 NIC's was setup as a web server. But the logical approach would have been to setup 4 smaller, less expensive, boxes with better redundancy.
But almost everyone I talk to who complains about the "architecture" is really complaining about tools, like shells and programs (e.g., PHP, Perl, etc.), and is relatively clueless about the OS architecture (disclosure, I think the modern NT kernel is pretty well-designed, even if the default tools are sub-par.)
.NET, etc.) and data sharing layers (OLE, COM, etc.), and the only philosophy I've seen to it is "make it easy to put data on the screen for the user". Hence spawning a process in Windows is almost as complex as firing up a full-blown GUI application with menubar, dialogs, etc. (yet spawning a thread is as fast as Unix). However, getting the GUI OUT is a harder problem. Did Microsoft finally ship a Windows server that didn't require a graphics card?
/dev, /var, and data directories were read-write, all else was read-only, so that in the event of power failure the system would be able to reach fsck at least. The first such system was put to sea and lasted over 8 months with zero reboots until Katrina knocked it out, all the while continually notifying a hardware watchdog circuit at 0.5 second intervals that it was still alive. The system configuration part took roughly three weeks, including testing, mainly due to IRQ issues with the proprietary hardware (USB, ethernet, PCMCIA, wireless, 12 serial ports, A/D, DIO, hardware watchdog).
I'll try a stab. As others pointed out, the "Unix architecture" is still a debated term, despite various books on the subject. Fundamentally, Unix is the POSIX-like C API (fopen(), brk(), fork(), etc.). However, many Unix purists I know say the tools are also part of the architecture, e.g. sh, passwd, chsh, ls, rm, etc., since the "philosophy" of the architecture is simple processes doing their one thing well and the system making it easy to string the tools together. From this base the other layers can be stacked on -- but they remain optional to the functioning of the computer.
The "Windows architecture" seems to be less philosophical and more pragmatic (in a commercial sense.) We have the WinAPI with layers above it including programming languages (MFC, VB,
I recently had a project that I think illustrates the difference in "philosophies" on a more technical level. I had two "embedded" computers that would be deployed on a sea platform: one was a PC/104 (100MHz 486) running TinyLinux in 32MB RAM and the other was a embedded Pentium 266MHz board running Win2k workstation in 64MB RAM. Yes, apples and oranges, absolutely. The vendor who supplied the Windows-based instrument had never put one of their instruments underwater OR under external automation before; they actually thought adding an entry to the Startup folder would be sufficient to make it all work.
On the Linux PC, I already had an image from the vendor that included a kernel and BusyBox and essential filesystem (16 MB). The system would boot and drop me to a shell on COM2, with ssh server listening on eth0. I compiled separately and added to the system just the utilities I needed: rsync, rsh/rlogin, inetd, Perl, crond, etc. I also recompiled the kernel, shaving 1MB off the memory use. Last, I re-partitioned the system such that only the
The Windows PC was already set up and needed only a "little" software to integrate with the Linux PC. Like I said, the data acquisition software had never been designed for automation. I used Kermit95 to manage the serial port link between the two systems (best $50 we spent by far). However, I had to write programs in C to start the program asynchronously (it would hang Kermit95's remote host command otherwise) AND stop the program by sending a specific series of key presses and mouse clicks (because it couldn't respond to anything else like a signal or status file). I also had to download third-party software to list and kill processes (PsList, PsKill). Testing was a royal pain and getting around the vendor's naivete ultimately cost us another three weeks.
You aske