I mean not doing much in the selling it. They've had the market share for a long time and relied on that fact along with the fact that Exchange was such a poor performer compared with Domino. Meanwhile Exchange has been improviing and growing marketshare and has been pushed strongly by MS. IBM hasn't been pushing Lotus Notes as much. If a company currently has Notes/Domino installed, they're probably going to stick with that combination although a few are moving or have already moved to Exchange. MS is really selling the Domino to Exchange migration path.
Domino and WebSphere have some overlap in terms of directory services and application development. WebSphere is IBM, Domino still kinda is Lotus, so WebSphere wins in the sales channels.
IBM doesn't want to provide easy solutions to their clients because confusion could lead them to an easier solution somewhere else. Back in 2003 they combined the teams but I haven't heard much regarding easy integration and fairly well defined roles for each.
Domino isn't really a driver for WebSphere sales. Had it been, there might have been a bigger push to get Domino out there. WebSphere isn't a driver for Domino roll-outs either. When WebSphere first came out I spent a couple of days at IBM's Palisades campus in NJ for a Notes sales pitch. The company I was at were Notes users. They kept trying to pitch WS as well but it was hard for them to answer questions as to when to use which for what. I've been trying to follow Notes/Domino development and WS integration since then from time to time but it's not worth it for me to look it up every day.
I don't think IBM has been doing much with Lotus Notes lately. A little bit with Domino server but even there it conflicts too much with WebSphere. I think IBM really dropped the ball with lotus notes.
The article mentions the "linux operating system" a lot. There is no such thing as the linux operating system. There are operating systems that use the linux kernel. There isn't even a tangible GNU/Linux operating system in my opinion. There are operating systems that use the linux kernel and other GNU software. Each distro uses a different kernel with different patches and different versions of libraries and other software that's important for the distro. So RHEL != SuSE != Debian != Slackware.
"Microsoft, especially likes to include patches for all of the subsystems of Redhat into it's security counts, and then compare that to patches for just windows "
By just windows... they don't mean just the windows kernel but everything that comes on the installation cd. It doesn't make sense for them to compare windows to a kernel. They should make it clear that they are comparing Windows XXX to Red Hat YYY though in the assesment. Even if they say Windows is better/more secure/whatever than Linux when they are really comparing it to RHEL, it only makes things WORSE when people on the OSS side start talking about Linux as an operating system. It just reinforces the position that Linux = Red Hat. That's not good. They mention IBM and HP as supporters of the linux operating system but that really talking about RHEL and sometimes SuSE for them.
There's the Linux Standard Base specification which is close to a "linux operating system" but ISV's aren't certifying to the LSB they are certifying to specific distros.
One of the main points I'm trying to make is how is it fair to compalin when someone like microsoft mixes the meaning of linux up but not when other people refer to the linux operating system? I don't think it is. Just how it is bad for Google to punish certain types of search engine tricks but use them themselves on their sites.
It's not just bad when Microsoft does it... it's always bad in my opinion. Comparing windows to the linux kernel is bad, comparing windos the the linux operating is bad too since it's not a tangible thing, although Unix to Linux kinda makes sense. That last link is a short article worth reading. It gives some insight on how the big guys push linux. It reinforces another point I'm trying to make. People love IBM but if you've ever had IBM try to sell you linux, it's usually because they know you want linux, they push much harder with AIX in the *nix space. I'm not saying that's bad... but then complaining when people like sun do the same thing doesn't make any sense.
Big corporations are involved in Linux now and there's a lot of big money at stake for some people with linux, open source, and related work. The whole XXX is our friend and YYY is our enemy is going to hurt the community. Things were different when people were just passing around floppies or ftp'ing things to each other. I think it's unwise for the OSS community to look at what red hat, suse, sun, ibm, etc are doing and evaluate the action... not just take it for granted that they're a 'friend'. There's a set of principles and beliefs that go with f/oss software that helps keep it what it is. There are even legally binding licenses to insure that happens. Just because corporations start using open f/oss doesn't mean they all of a sudden pick up all those principles and it especially doesn't mean that they have to stick to them if they decide it's not in their best interest. It's good that big business has jumped onto f/oss but it's in their interest to try and control it as much as possible and you see that happening quite a bit.
I agree to a point. Within the method body, I don't comment too often, but there was one bit of advice I got once that I try and use as much as possible. Before you write your method, write the javadoc comment for it. It gives you a minute to think about what it actually is going to be doing and lets you think about the design before you start working on the implementation.
Also commenting accessible members and methods makes coding on big projects a lot easier, as well as reuse further down the road more managable. Reuse especially should be a big factor which is a big reason why we should use these high level OO languages to begin with.
With modern IDE's it's nice when you can start typing, and if you have a question about the implementation of a method, you can ctrl-space to have the java doc comment come up for the classes in your project. This is great, especially if you're working with other people's code as well. The docs say it does this this and this... no need to try and try and read their code to figure out what it does. Unless of course it doesn't do what it says:)
Just knowing what parameters it takes, what it returns along with a descriptive name is usually enough but you can't but everything into the name like it may return an empty array instead of null when it doesn't find anything. Things like that help a lot.
"If you're having trouble with RH, grab SuSE. If you're having trouble with SuSE, grab Mandrake. If you're having problem with Mandrake or any of the many other commercial dists, grab Debian or Ubuntu."
And you think that translates into an "enterprise ready" game plan!?!??!!?!?
The article talks about Linux being an enterprise class OS. It's not an OS. Not all Linux distributions are enterprise ready. If you install Oracle on Debian, are you going to get support from Oracle or Debian? If you install WebSphere on Mandrake, will you get support? Even if you install Fedora and grab all the SRPMS for RHEL to setup a Samba server, is Microsoft going to help you if you're having problems connecting your windows desktops to it?
An operating system for a server doesn't do anything except provide a base to install what it is that will run the server tasks in enterprise deployments. If you can't get the software stack supported on the operating system, it's a mute point. People choose linux because they want to cut down their deployment and support costs and spend more money on the part that actually does the work. Even if you set up all that stuff your self and get it to work, the time and effort in getting it done and keeping it up to date has to factor in.
CentOS, is NOT a free RHEL. It is built from the SRPMS but you won't get support for it from people that support RHEL and it is not something that Red Hat is providing. The people that make CentOS (which is very good) take advantage of the GPL and build their own distro based on RHEL. Red Hat doesn't seem too happy about this.
I wasn't talking about OpenSolaris. OpenSolaris is different from Solaris. Apparently there's a pilot program for OpenSolaris with developers and about 50 or so ISV's. You can't even compare OpenSolaris to Linux. OpenSolaris is an operating system. Linux is a kernel. Just curious, what are these restrictions on OpenSolaris anyway? That it's not GPL'ed? That doesn't prohibit you from including GPL'ed tools with it, you just can't mix OpenSolaris and GPL code together. You can't take GPL code and put it into BSD'd code without making the BSD code GPL either. But you're right, it still hasn't arrived yet. Right now, it's still yet to be delivered, but Solaris 10 is out there.
Solaris 10 is free as in beer. You don't have to pay to deploy it, you just have to pay for support if you choose to but nothing is stopping you from downloading it and installing it on a bunch of systems. If you buy something like Oracle that is certified for Solaris 10, you can get support from Oracle even if you're not paying for support from Sun. If you buy support from Sun and install software on it, you can get support from Sun. Try to get support from Oracle for CentOS, SAP, Websphere, Peoplesoft, etc.
Linux IS free. Linux is not an operating system. Most people don't want kernels, they want operating systems. Without talking about specific linux distros, this article is just osdl marketing fluff. Don't get me wrong. It's not that I don't think RHEL or SuSE Enterprise Linux are enterprise ready, or that even the newest linux kernel is. But when you talk about linux os's in general, like this article, the argument doesn't work. A kernel alone is pretty much useles. This is talking about how good the kernel is, implying the operating systems that are built on it will have all the same qualities which isn't true.
This article itself does the same thing. Talking about how the Linux operating system is this, or that or how the linux generated 36 billion in revenues. Linux is a kernel you can download for free. Nobody pays to use linux, they pay to use a linux based OS. Most popularly Red Hat or sometimes SuSE. It's not just Sun and anti Linux people doing it. This article is talking about operating systems, not kernels. Some people may consider this nit-picking, but based on your comment, I'd think you agree with my point. Linux is Free. Linux based Operating Systems aren't in all cases. (Fedora != RHEL). And just because, Linux the kernel can do something, doesn't mean these Linux-based OS's are shipping with that kernel and the user space software on top of the kernel can handle these load.
What happens if one day Hurd comes out and it kicks everybody's ass. Or some other free kernel and the big vendors decide to switch. Someone like Red Hat can still be Red Hat, but not be Linux.
"The whole Solaris 10 for free thing is to persuade people to stick with Sun even if in reality a "free" Solaris is anything but when you slap on support costs."
So? That's not any different than Red Hat. Show me where I can download RHEL on red hat's site without paying and with out just a trial period.
"Geez! With an attitude like yours, I have to think that you've run into the very small vocal minority of Linux supporters who are a little imbalanced."
With an attitude like yours, I have to think maybe YOU are one of those imbalanced linux supporters considering I just posed a question and didn't give my opinion. Or maybe you're just a groupthink kinda guy.
The article is about enemies of linux claiming a lacking in linux to promote their own platfor which is what my link describes.
The point is, the article doesn't name any specific examples of what is being said and who is saying it. How do we know we can't make the same argument with those enemies when we don't know who they are or the context in which they are saying them? This article is no better than thte FUD it seems to be dismissing. If you take parts out of that link I sent you can make an article look like IBM is crapping all over Linux. Too many things that get called FUD these days are similar.
So much of this news on open source is just crap. Nothing more than opinion peices with no real content. Unless someone with a development or administrative background reports something I don't care about it too much when it comes to software or hardware. There used to be a time in open source software where if someone said "your x doesn't do y" or doesn't do it as well, someone would SHOW that it does instead of claiming "So and so is using it, it must be good". Now linux and open source has their own sensational pr and marketting divisions.
You even say something similar yourself at the end of your posting
"But I do see that there are people who would prefer to see Linux and its user base curl up and die. It is those people that I have a problem with."
Who are these people and what is it that makes you think that and how is it different than this IBM article?
"but management wants someone to turn to if a problem should arise"
Sun has StarOffice which is the commercial version built off the OpenOffice.org codebase. The sell it for a lot less than MSOffice and they provide support for it as well.
A bunch of open source projects have a company behind them just for this reason.
"Thirdly, we don't cut off the old file formats. So we maintain backward compatibility with the old Office file formats. I've got a bunch of customers who are using StarOffice to import their old Office documents and then export them to Office XP. Now go figure--we're the migration tool."
"However, I've always wondered if sun's motives for funding open office were a bad thing. (Apparently they just want to make Microsoft mad.)"
I don't think that was Sun's only motivation. Most people think of Sun as only a server vendor. They really started out as a workstation company and still make a lot of workstation products. They were very good machines for workstation type functions such as CAD, EDA, simulations and other engineering/mathematical applications. Typical IBM PC's couldn't handle the type of workload these workstations did.
As PC's and the collaboration and office tools used them became more prominent (Windows, Office, groupware), people that only used workstations were at a disadvantage because they couldn't run these Windows applications on their workstations. Then PC's started to get more powerful and were able to handle some more of the work that you'd normally get a workstation for.
Sun at one point had a PCI x86 card that you could insert in your workstation to run windows in solaris. Not sure if they still have it, but it shows how important running these windows only applications had become. If you needed a workstation, you also needed a PC for the "regular" stuff. This made the already high cost of workstations more expensive because they couldn't handle everything the PC could.
So, the goal to "make Microsoft mad" isn't the only reason. The reason was, that you shouldn't be locked into any particular platform to be able to function in most organizations. With an office suite that can read and write to the defacto company standards that runs anywhere you want it to run, you were freer to choose the platform that made more sense for you, without having to have two computers.
This is probably the most compelling reason that Sun did what they did with Star/OpenOffice, not to just tick someone off. It's not just good for Solaris users, it's good for people that want to run any platform they choose. Including Linux users.
Imagine a company that can give it's engineers high end workstations running unix, it's call center and admin staff linux or some thin client based on a *nix, it's public relations and design groups Macs, etc. Or you can choose whichever you waht that makes you more productive still while being able to read and write documents sent from others in and out of the company. This is a very important thing for someone that doesn't sell windows based machines.
That's why projects like evolution and the various connectors are important as well. I feel it's a shame IBM never went all out with LotusNotes. It had a lot of good things going for it. Maybe if they opened sourced it they wouldn't have gotten slammed in market share by exchange like they did. It also would have given everyone a very mature, well known, widely deployed groupware product. I wonder if it's even still a viable option to do such a thing anymore.
Interestingly, from a while back, there's this blog by Tiemann. And this report of the blogwar (Red Hat opens losing propaganda offensive against Sun) With Sun being one of the biggest contributors to open source, Red Hat being the biggest commercial linux distro, and OSI well... well at least it's not run by esr anymore... it would be nice to see the opensource/sun feuding end.
There was a lot of talk about cutting down on the number of OSI approved licenses after Sun's CDDL was approved. There was a little talk last year but nothing as big as what's been happening now. Is there any conflict of interest now that an exec of RedHat runs OSI with regards to what OSI stamps as OpenSource and what licenses they decide to trim? There has been a lot of bad mouthing back and forth between Red Hat and Sun lately. Could they use their position to cut out the CDDL?
I agree with you on how open source seems to be changing since the big guys stepped in.
"With respect to the above code, OOP is good for fairly static hierarchies. OOP is bad for a system in constant change."
Another thing... there's the whole reflection api in Java where you can inspect what methods it has at runtime. You don't have to call account.fancy1() account.fancy2() directly. You can loop through all the methods in the function that start with "fancy" and call them.
I had to use the reflection api once in a class that merged a text file with various beans. The text file contained tokens like $${token} and would call bean.getToken() where token could be any string so $${email} would call getEmail(). This was handy as the bean and text file were subject to change but the parser class didn't have to. A lot easier than writing something like if (token.equals("email") bean.getEmail(); for every possible value of token.
The reflection api used to be very slow. You still take a performance hit for it but it's gotten a lot faster with each JVM release.
instead of defining FancyAccounts as classes.. define them as interfaces... more code but you keep the code for each type of account in the class where it belongs. So you'll have something like.
abstract class Account { float balance; abstract void doSomething();//not a fancy account, no fancy function. }
problem is you have to define fancy1 in both classes and you have to worry about keeping them the same if they have to be.
C++ has the edge in an example like this since you could just define the interfaces as classes with the methods in them and use multiple inheritance. Your C example could be replicated in Java with
"With respect to the above code, OOP is good for fairly static hierarchies. OOP is bad for a system in constant change."
That's not really true in this example. In Java, if I wanted to add another Fancy interface and class that has a completely different doSomething(). Add the class file to the application and all the calls to doSomething on it will work. Update the class that feeds Accounts in to the system and all you have to do is recompile one file in addition to adding the new class. With C you'd have to recompile and link everything.
" Perhaps you could name a case where Java is faster than C"
Sure.
The Ackermann test, Java wins by a small margin.
Also look at takfp. The interesting thing about takfp can be seen here in the side by side results The slope of the red line (Java) is not as steep as N increase so even though it starts slower, it wins in the end and continues to do beat ther performance of c and c++ even though it's start time was a lot slower (these numbers include startup time). Saw this in some other tests too, like reverse-file. This is mostly in comparison to c++ but it shows how JIT compiling could help.
I didn't go through all the gcc parameters but each program was compiled for the machine it ran on. In shrink wrap software, or even in the data center, you're not going to compile the code on each machine and it has to be tuned to the lowest common denominator. With JIT you don't have to worry about that as the VM will do the necessary optimizations for you. The benefits can be seen in some of these tests.
There are a lot of cases where it's close as well. In other tests I've seen Java in some cases outperformed C++ and C under large usage when the gcc binaries were optimized for the lowest common denominator. Anyway... the potention for VM's JIT compiling and all that is not just theoretical. It needs to be improved, but that's no reason to give up on it.
In other test, the lines of code are dramatically different. Would I rather write a 178 line spell checker in C and have it run in about half a second? Or write a 23 line Java spellchecker that runs in about 2 seconds? When you go past just number crunching and into real applications like web applications, you'll see a big savings in time and development.
There's a place for high level languages like Python, Ruby and Java and a place for languages like C. The real power language seems to be OCaml. It seems to do really well on both loc, memory and cpu time.
This was really cool to look at, thanks for the link. I looked up some other stuff as well. I'm tired of hearing how Java is slower than PHP since in my own benchmarks I haven't seen what people are talking about. Most of the side by side graphs look very much like this where it seems that PHP doesn't scale under load compared to java, c or c++.
I've seen it on other sites as well. Something about table widths being set to 100% or something. On some sites, the main text table cell doesn't show up until there's a reload. The same ctrl- ctrl+ fixes those too or a reload. It's really annoying.
Sun Microsystems just barely held onto its number three position, with $1.365 billion in sales after a 5.1 percent decline in revenue compared to the fourth quarter of 2003. Part of Sun's revenue decline is due to the shift toward X86 server, but most of it is due to Sparc customers buying less iron or
paying a lot less for what they do buy.
Also interesting...
Sun's X86 server revenue grew by 360 percent, hitting $152.5 million for 2004.
" What I would like to know, is are they counted as Unix or Linux servers?"
My guess would be Linux as reports claim that the majority of Sun's x86 servers are ordered with linux preloaded. Compared to 4.something billion in linux server sales overall, it wouldn't make that much difference in the numbers. Sun's still a small player in x86 but they're growing fast. They're more in the 64 bit x86 space. I read somewhere that they're the largest buyer of AMD Opteron chips.
That's not necessarily true in all cases. In that (or a similar report based on the idc study), it mentioned that Sun's shippments where up higher than their revenues. Not sure about the other unix vendors but Sun has been dropping it's prices.
This is a pretty interesting rebuttal from one of the OpenSolaris developers to what an HP Exec said about the system. May address some of your open development concerns. I thought it was great to see that the OpenSolaris movement already had an outspoken developer that can fly off the handle and curse. Seems like they're off to a good start to being just a regular open source joe:)
Slashdot Mirror
By that logic, wouldn't writing a patch that powered off your system once it started booting would be a real fix too :)
I mean not doing much in the selling it. They've had the market share for a long time and relied on that fact along with the fact that Exchange was such a poor performer compared with Domino. Meanwhile Exchange has been improviing and growing marketshare and has been pushed strongly by MS. IBM hasn't been pushing Lotus Notes as much. If a company currently has Notes/Domino installed, they're probably going to stick with that combination although a few are moving or have already moved to Exchange. MS is really selling the Domino to Exchange migration path.
Domino and WebSphere have some overlap in terms of directory services and application development. WebSphere is IBM, Domino still kinda is Lotus, so WebSphere wins in the sales channels.
IBM doesn't want to provide easy solutions to their clients because confusion could lead them to an easier solution somewhere else. Back in 2003 they combined the teams but I haven't heard much regarding easy integration and fairly well defined roles for each.
Domino isn't really a driver for WebSphere sales. Had it been, there might have been a bigger push to get Domino out there. WebSphere isn't a driver for Domino roll-outs either. When WebSphere first came out I spent a couple of days at IBM's Palisades campus in NJ for a Notes sales pitch. The company I was at were Notes users. They kept trying to pitch WS as well but it was hard for them to answer questions as to when to use which for what. I've been trying to follow Notes/Domino development and WS integration since then from time to time but it's not worth it for me to look it up every day.
I don't think IBM has been doing much with Lotus Notes lately. A little bit with Domino server but even there it conflicts too much with WebSphere. I think IBM really dropped the ball with lotus notes.
"Microsoft, especially likes to include patches for all of the subsystems of Redhat into it's security counts, and then compare that to patches for just windows "
By just windows... they don't mean just the windows kernel but everything that comes on the installation cd. It doesn't make sense for them to compare windows to a kernel. They should make it clear that they are comparing Windows XXX to Red Hat YYY though in the assesment. Even if they say Windows is better/more secure/whatever than Linux when they are really comparing it to RHEL, it only makes things WORSE when people on the OSS side start talking about Linux as an operating system. It just reinforces the position that Linux = Red Hat. That's not good. They mention IBM and HP as supporters of the linux operating system but that really talking about RHEL and sometimes SuSE for them.
There's the Linux Standard Base specification which is close to a "linux operating system" but ISV's aren't certifying to the LSB they are certifying to specific distros.
One of the main points I'm trying to make is how is it fair to compalin when someone like microsoft mixes the meaning of linux up but not when other people refer to the linux operating system? I don't think it is. Just how it is bad for Google to punish certain types of search engine tricks but use them themselves on their sites.
It's not just bad when Microsoft does it... it's always bad in my opinion. Comparing windows to the linux kernel is bad, comparing windos the the linux operating is bad too since it's not a tangible thing, although Unix to Linux kinda makes sense. That last link is a short article worth reading. It gives some insight on how the big guys push linux. It reinforces another point I'm trying to make. People love IBM but if you've ever had IBM try to sell you linux, it's usually because they know you want linux, they push much harder with AIX in the *nix space. I'm not saying that's bad... but then complaining when people like sun do the same thing doesn't make any sense.
Big corporations are involved in Linux now and there's a lot of big money at stake for some people with linux, open source, and related work. The whole XXX is our friend and YYY is our enemy is going to hurt the community. Things were different when people were just passing around floppies or ftp'ing things to each other. I think it's unwise for the OSS community to look at what red hat, suse, sun, ibm, etc are doing and evaluate the action... not just take it for granted that they're a 'friend'. There's a set of principles and beliefs that go with f/oss software that helps keep it what it is. There are even legally binding licenses to insure that happens. Just because corporations start using open f/oss doesn't mean they all of a sudden pick up all those principles and it especially doesn't mean that they have to stick to them if they decide it's not in their best interest. It's good that big business has jumped onto f/oss but it's in their interest to try and control it as much as possible and you see that happening quite a bit.
Also commenting accessible members and methods makes coding on big projects a lot easier, as well as reuse further down the road more managable. Reuse especially should be a big factor which is a big reason why we should use these high level OO languages to begin with.
With modern IDE's it's nice when you can start typing, and if you have a question about the implementation of a method, you can ctrl-space to have the java doc comment come up for the classes in your project. This is great, especially if you're working with other people's code as well. The docs say it does this this and this... no need to try and try and read their code to figure out what it does. Unless of course it doesn't do what it says :)
Just knowing what parameters it takes, what it returns along with a descriptive name is usually enough but you can't but everything into the name like it may return an empty array instead of null when it doesn't find anything. Things like that help a lot.
And you think that translates into an "enterprise ready" game plan!?!??!!?!?
The article talks about Linux being an enterprise class OS. It's not an OS. Not all Linux distributions are enterprise ready. If you install Oracle on Debian, are you going to get support from Oracle or Debian? If you install WebSphere on Mandrake, will you get support? Even if you install Fedora and grab all the SRPMS for RHEL to setup a Samba server, is Microsoft going to help you if you're having problems connecting your windows desktops to it?
An operating system for a server doesn't do anything except provide a base to install what it is that will run the server tasks in enterprise deployments. If you can't get the software stack supported on the operating system, it's a mute point. People choose linux because they want to cut down their deployment and support costs and spend more money on the part that actually does the work. Even if you set up all that stuff your self and get it to work, the time and effort in getting it done and keeping it up to date has to factor in.
CentOS, is NOT a free RHEL. It is built from the SRPMS but you won't get support for it from people that support RHEL and it is not something that Red Hat is providing. The people that make CentOS (which is very good) take advantage of the GPL and build their own distro based on RHEL. Red Hat doesn't seem too happy about this.
I wasn't talking about OpenSolaris. OpenSolaris is different from Solaris. Apparently there's a pilot program for OpenSolaris with developers and about 50 or so ISV's. You can't even compare OpenSolaris to Linux. OpenSolaris is an operating system. Linux is a kernel. Just curious, what are these restrictions on OpenSolaris anyway? That it's not GPL'ed? That doesn't prohibit you from including GPL'ed tools with it, you just can't mix OpenSolaris and GPL code together. You can't take GPL code and put it into BSD'd code without making the BSD code GPL either. But you're right, it still hasn't arrived yet. Right now, it's still yet to be delivered, but Solaris 10 is out there.
Solaris 10 is free as in beer. You don't have to pay to deploy it, you just have to pay for support if you choose to but nothing is stopping you from downloading it and installing it on a bunch of systems. If you buy something like Oracle that is certified for Solaris 10, you can get support from Oracle even if you're not paying for support from Sun. If you buy support from Sun and install software on it, you can get support from Sun. Try to get support from Oracle for CentOS, SAP, Websphere, Peoplesoft, etc.
Linux IS free. Linux is not an operating system. Most people don't want kernels, they want operating systems. Without talking about specific linux distros, this article is just osdl marketing fluff. Don't get me wrong. It's not that I don't think RHEL or SuSE Enterprise Linux are enterprise ready, or that even the newest linux kernel is. But when you talk about linux os's in general, like this article, the argument doesn't work. A kernel alone is pretty much useles. This is talking about how good the kernel is, implying the operating systems that are built on it will have all the same qualities which isn't true.
What happens if one day Hurd comes out and it kicks everybody's ass. Or some other free kernel and the big vendors decide to switch. Someone like Red Hat can still be Red Hat, but not be Linux.
When you're talking to the type of enterprise customers Sun (even Oracle, IBM, etc) usually talks to, Linux does mean Red Hat.
So? That's not any different than Red Hat. Show me where I can download RHEL on red hat's site without paying and with out just a trial period.
With an attitude like yours, I have to think maybe YOU are one of those imbalanced linux supporters considering I just posed a question and didn't give my opinion. Or maybe you're just a groupthink kinda guy.
The article is about enemies of linux claiming a lacking in linux to promote their own platfor which is what my link describes.
The point is, the article doesn't name any specific examples of what is being said and who is saying it. How do we know we can't make the same argument with those enemies when we don't know who they are or the context in which they are saying them? This article is no better than thte FUD it seems to be dismissing. If you take parts out of that link I sent you can make an article look like IBM is crapping all over Linux. Too many things that get called FUD these days are similar.
So much of this news on open source is just crap. Nothing more than opinion peices with no real content. Unless someone with a development or administrative background reports something I don't care about it too much when it comes to software or hardware. There used to be a time in open source software where if someone said "your x doesn't do y" or doesn't do it as well, someone would SHOW that it does instead of claiming "So and so is using it, it must be good". Now linux and open source has their own sensational pr and marketting divisions.
You even say something similar yourself at the end of your posting
"But I do see that there are people who would prefer to see Linux and its user base curl up and die. It is those people that I have a problem with."
Who are these people and what is it that makes you think that and how is it different than this IBM article?
What IBM? Does this make IBM an enemy of Linux as well?
Look in the title bar of your browser. There is key word spamming in it.
Sun has StarOffice which is the commercial version built off the OpenOffice.org codebase. The sell it for a lot less than MSOffice and they provide support for it as well.
A bunch of open source projects have a company behind them just for this reason.
I don't think that was Sun's only motivation. Most people think of Sun as only a server vendor. They really started out as a workstation company and still make a lot of workstation products. They were very good machines for workstation type functions such as CAD, EDA, simulations and other engineering/mathematical applications. Typical IBM PC's couldn't handle the type of workload these workstations did.
As PC's and the collaboration and office tools used them became more prominent (Windows, Office, groupware), people that only used workstations were at a disadvantage because they couldn't run these Windows applications on their workstations. Then PC's started to get more powerful and were able to handle some more of the work that you'd normally get a workstation for.
Sun at one point had a PCI x86 card that you could insert in your workstation to run windows in solaris. Not sure if they still have it, but it shows how important running these windows only applications had become. If you needed a workstation, you also needed a PC for the "regular" stuff. This made the already high cost of workstations more expensive because they couldn't handle everything the PC could.
So, the goal to "make Microsoft mad" isn't the only reason. The reason was, that you shouldn't be locked into any particular platform to be able to function in most organizations. With an office suite that can read and write to the defacto company standards that runs anywhere you want it to run, you were freer to choose the platform that made more sense for you, without having to have two computers.
This is probably the most compelling reason that Sun did what they did with Star/OpenOffice, not to just tick someone off. It's not just good for Solaris users, it's good for people that want to run any platform they choose. Including Linux users.
Imagine a company that can give it's engineers high end workstations running unix, it's call center and admin staff linux or some thin client based on a *nix, it's public relations and design groups Macs, etc. Or you can choose whichever you waht that makes you more productive still while being able to read and write documents sent from others in and out of the company. This is a very important thing for someone that doesn't sell windows based machines.
That's why projects like evolution and the various connectors are important as well. I feel it's a shame IBM never went all out with LotusNotes. It had a lot of good things going for it. Maybe if they opened sourced it they wouldn't have gotten slammed in market share by exchange like they did. It also would have given everyone a very mature, well known, widely deployed groupware product. I wonder if it's even still a viable option to do such a thing anymore.
Interestingly, from a while back, there's this blog by Tiemann. And this report of the blogwar (Red Hat opens losing propaganda offensive against Sun) With Sun being one of the biggest contributors to open source, Red Hat being the biggest commercial linux distro, and OSI well... well at least it's not run by esr anymore... it would be nice to see the opensource/sun feuding end.
I agree with you on how open source seems to be changing since the big guys stepped in.
Another thing... there's the whole reflection api in Java where you can inspect what methods it has at runtime. You don't have to call account.fancy1() account.fancy2() directly. You can loop through all the methods in the function that start with "fancy" and call them.
I had to use the reflection api once in a class that merged a text file with various beans. The text file contained tokens like $${token} and would call bean.getToken() where token could be any string so $${email} would call getEmail(). This was handy as the bean and text file were subject to change but the parser class didn't have to. A lot easier than writing something like if (token.equals("email") bean.getEmail(); for every possible value of token.
The reflection api used to be very slow. You still take a performance hit for it but it's gotten a lot faster with each JVM release.
instead of defining FancyAccounts as classes.. define them as interfaces... more code but you keep the code for each type of account in the class where it belongs. So you'll have something like.
//not a fancy account, no fancy function.
abstract class Account {
float balance;
abstract void doSomething();
}
interface FancyAccount1 {
void fancy1();
}
interface FancyAccount2 {
void fancy2();
}
class FancyAccount extends Account implements FancyAccount1 {
void fancy1() { };
}
class FancierAccount extends Account implements FancyAccount1, FancyAccount2 {
void fancy1() { };
void fancy2() { };
}
problem is you have to define fancy1 in both classes and you have to worry about keeping them the same if they have to be.
C++ has the edge in an example like this since you could just define the interfaces as classes with the methods in them and use multiple inheritance. Your C example could be replicated in Java with
"With respect to the above code, OOP is good for fairly static hierarchies. OOP is bad for a system in constant change."
That's not really true in this example. In Java, if I wanted to add another Fancy interface and class that has a completely different doSomething(). Add the class file to the application and all the calls to doSomething on it will work. Update the class that feeds Accounts in to the system and all you have to do is recompile one file in addition to adding the new class. With C you'd have to recompile and link everything.
Sure.
The Ackermann test, Java wins by a small margin.
Also look at takfp. The interesting thing about takfp can be seen here in the side by side results The slope of the red line (Java) is not as steep as N increase so even though it starts slower, it wins in the end and continues to do beat ther performance of c and c++ even though it's start time was a lot slower (these numbers include startup time). Saw this in some other tests too, like reverse-file. This is mostly in comparison to c++ but it shows how JIT compiling could help.
I didn't go through all the gcc parameters but each program was compiled for the machine it ran on. In shrink wrap software, or even in the data center, you're not going to compile the code on each machine and it has to be tuned to the lowest common denominator. With JIT you don't have to worry about that as the VM will do the necessary optimizations for you. The benefits can be seen in some of these tests.
There are a lot of cases where it's close as well. In other tests I've seen Java in some cases outperformed C++ and C under large usage when the gcc binaries were optimized for the lowest common denominator. Anyway... the potention for VM's JIT compiling and all that is not just theoretical. It needs to be improved, but that's no reason to give up on it.
In other test, the lines of code are dramatically different. Would I rather write a 178 line spell checker in C and have it run in about half a second? Or write a 23 line Java spellchecker that runs in about 2 seconds? When you go past just number crunching and into real applications like web applications, you'll see a big savings in time and development.
There's a place for high level languages like Python, Ruby and Java and a place for languages like C. The real power language seems to be OCaml. It seems to do really well on both loc, memory and cpu time.
This was really cool to look at, thanks for the link. I looked up some other stuff as well. I'm tired of hearing how Java is slower than PHP since in my own benchmarks I haven't seen what people are talking about. Most of the side by side graphs look very much like this where it seems that PHP doesn't scale under load compared to java, c or c++.
Here's an article looks like CS2C has made some changes to JDS and is calling it NeoShine?
I've seen it on other sites as well. Something about table widths being set to 100% or something. On some sites, the main text table cell doesn't show up until there's a reload. The same ctrl- ctrl+ fixes those too or a reload. It's really annoying.
No and yes. From http://www.itjungle.com/breaking/bn022405-story01. html
Also interesting...
" What I would like to know, is are they counted as Unix or Linux servers?"My guess would be Linux as reports claim that the majority of Sun's x86 servers are ordered with linux preloaded. Compared to 4.something billion in linux server sales overall, it wouldn't make that much difference in the numbers. Sun's still a small player in x86 but they're growing fast. They're more in the 64 bit x86 space. I read somewhere that they're the largest buyer of AMD Opteron chips.
That's not necessarily true in all cases. In that (or a similar report based on the idc study), it mentioned that Sun's shippments where up higher than their revenues. Not sure about the other unix vendors but Sun has been dropping it's prices.
This is a pretty interesting rebuttal from one of the OpenSolaris developers to what an HP Exec said about the system. May address some of your open development concerns. I thought it was great to see that the OpenSolaris movement already had an outspoken developer that can fly off the handle and curse. Seems like they're off to a good start to being just a regular open source joe :)