The author of this article, Tom Adelstein, articulates in unique detail the importance of Sun’s JDS more effectively than Sun itself has done. And he explains within the context of .Net and its limitations how the scope of JDS extends beyond the GNU/Linux community’s parochial horizons.
How to Misunderstand Sun’s Linux Desktop Strategy
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
That was a great article.
It was not a great article. The only thing great about it was the title: “How to Misunderstand Sun’s Linux Desktop Strategy” because it did a very good job at misunderstanding basically everything.
My head hurt just reading it. It was a horrible mishmash of confused ideas with no coherent voice or strategy. Of course in this way, it is indeed much like Sun’s Java Desktop strategy.
It also seemed to pull numbers right out of thin air. Where are the sources for OO.o providing “95% interoperability”? What about the “400 million Windows 98 users”? These figures are simply made up. The latter does not make sense at all given Google’s stats for desktop useage (http://www.google.com/zeitgeist) Indeed not much of the article makes sense at all.. it is just a long, deranged rant, full of disjointed ideas.
I couldn’t get past the marketing hype where he berated MS for the misleading usage of .NET, but completely ignored the fact that Sun is using the Java name for a desktop that has nothing to do with Java. What a farce. Note: I am a Java developer by trade.
I second that. My head hurt after reading it too. If there’s a nother site that better explains .NET vs Java I would be interested in reading it.
Adelstein spends some of the article alluding to the fact that he is unbiased. I usually find than when an author mentions this in his/her piece, bias exists in the article.
Here is the source:
Hal Varian, professor of information management, economics, and businessat UC Berkeley did a project with his son, Christopher, entitled http://www.acmqueue.com/modules.php?name=Content&pa=showpage&pid=55… .
MOXIE reveals that out of 50 random MS Word (.doc) files downloaded off of the Internet with Google, 90% loaded perfectly. This was using OOo 1.0.1 — and we know anecdotally that OOo 1.1 compatibility is strokes better.
Interestingly, StarOffice 6.0 in the study got a 93% perfect score on the same MS Office files. This meets expectations because SO6 has additional fonts which reduce the chance for font substitution irregularities which account for most of the file format incompatibilities in my own experience as a Win2Lin desktop migration suport and training specialist.
SO7, which is based on OOo1.1, also has noticably better file format import results than SO6. All tolled, this suggests that “95% interoperability” is an accurate statement in the assumed context of ‘most’ documents that are created in a government workflow.
Now additionally, the article above, “How to Misunderstand…,” mentions the OOo 95% compatibility within the context of MS Office inter-version incompatibities which are far more significant a problem than OOo success handling other formats. In fact — and I know this firsthand — OOo has far more success handling MS Office file formats than MS Office itself. For instance, the OOo Discuss Mailing List hears weekly of examples where OOo has filtered a broken MS Office document and made it palatable for MS Office again.
File incompatibility has been used as an excuse to avoid migration or adoption of OOo. It is the pot calling the kettle black.
Sam Hiser
Marketing Project Lead
OpenOffice.org
It was a horrible mishmash of confused ideas with no coherent voice or strategy. Of course in this way, it is indeed much like Sun’s Java Desktop strategy.
Sun’s Java Desktop strategy is not a mishmash of confused ideas. Java Desktop has a very clear target in mind. And that target is Microsoft. That target, is the average desktop user. That target is also Longhorn and its managed code environment on the desktop. Sun’s answer is Java Desktop + Tiger. They are shipping a desktop with a fully functional office suite for less than $100. That’s an very good bargan. Furthermore, they in the process of negotiating deals with Walmart, and Circuit City to sell systems preloaded with Java Desktop, and they have gotten a deal with the Chinese government to deliver 200 million Java Desktops to China.
Java Desktop does not target Linux. In fact, it helps Linux a great deal.
The target is clear, It is the average user. And it is Longhorn. The difference? Java Desktop is available now. Java 1.4 already delivers performance that many benchmarks have shown to be over twice as fast as .NET. And Tiger will be available long before Longhorn. Furthermore, Tiger will deliver lightening fast performance on Java Desktop, as well as on Windows.
The other difference? Bill Gates has stated that by the time Longhorn ships (2006), he expects that the average PC will have a dual core processor running at 6 Ghz, 2 Gb of RAM, and a terabyte of hard disk space. That probably gives some indication of the the hardware requirements Microsoft expects that Longhorn will require for acceptable performance. (And given that any code that uses the Windows API is entirely managed code… ie: all new apps developed for Windows are .NET apps, I think those requirements are probably about right).
Java 1.4 is here now. Tiger will be available soon, and will perform better than Java 1.4. Linux is here now, and performs better than Windows. And the combination of Tiger and Linux will leave Longhorn in the dust on the desktop. And it will do it while Longhorn is still nothing more than vaporware.
That’s Sun’s strategy. And if the China order is any indication, it poses a real threat to Microsoft on the desktop. 200 million installations has got to but a significant dent in Microsoft’s desktop market share.
Sounds to me more like a SUN propaganda and a Java ad more than anything else.
I’ve met Sam Hiser on a couple of occasions (NYLUG, LWE ’03) and I think he’s a great guy. I also agree with the general compatibility statistics he quotes, and have seen MS Office version incompatibility up close and personal.
But “parochialism”? The author says, “Linux developers and advocates tend to value the concepts of the Free Software Foundation and its GPL license as the only way to work.” What’s misguided about that? Adelstein suggests there’s a lack of vision regarding the huge market of Windows 98 users that can be converted. Huh?
He reacts negatively to the dislike people have for marketing tricks such as calling a Linux distribution a “Java Desktop System”. Tom, it isn’t about competing with smaller distros — the name is just ridiculous, and a shallow (if understandable) marketing ploy.
Good for Sun that they’re doing beneficial things while trying to stay profitable. Boo to IBM for only supporting Linux as a server OS. Beware C# and any Microsoft “standards”. Yes, that’s all well and good, but who wrote the “parochial” quote?
It doesn’t especially fit the article. Eugenia, are you editorializing here?
(And given that any code that uses the Windows API is entirely managed code… ie: all new apps developed for Windows are .NET apps, I think those requirements are probably about right).
What I mean by that, is that any new apps developed for Longhorn will be .NET apps. The current Windows API will be replaced with the WinFX API in Longhorn. WinFX API is entirely managed, the current incarnation of .NET is nothing more than a subset of the WinFX API. In other words, .NET exists now so that programmers can write programs for Longhorn now, and so that by the time Longhorn finally ships, programmers will be acustomed to its API and be ready to deliver applications for Longhorn.
And as I said, that means all apps written for Longhorn are managed code that runs in the .NET framework.
And Java performance leaves .NET in the dust. And Tiger will be even better.
Desktop apps that vendors write for Java Desktop will also run on other versions of Linux. And with numbers like 200 million installations, I think vendors will take notice and start writing apps for Java Desktop.
It’s a little disconcerting that many folks still don’t understand why it’s not called the “Linux Desktop System”. Sun has no intention — that I can see — of advancing Linux with their JDS. This release of the JDS is about laying a foundation for the Java/desktop framework. It’s an attempt to compete (pre-emptive strike?) with what Microsoft is planning with .NET. It’s not just a “branding scheme” — it’s not about taking a Linux distro and slapping “Java” onto it so people will have the word pounded into their brains. At least that’s what I got out of the article.
John
“It’s a little disconcerting that many folks still don’t understand why it’s not called the “Linux Desktop System”. Sun has no intention — that I can see — of advancing Linux with their JDS.”
Linux will advance because of JDS though. And Sun does advance Linux by helping fund GNOME and Ximian and such.
If JDS does well, and so far it is doing very well, than app vendors are going to take notice and start writing apps for Linux. Device makers are going to start shipping Linux drivers standard with their devices, etc.
No matter how you look at it, a successful JDS helps Linux.
Sun is a positive force for Linux and open source technology. They use Java because it is a fexible technology for tying together a product line, and they are doing the right things as a vendor. Their product line is based on Linux and Solaris.
I would rather Sun make changes, not to Linux, but to Java, and that’s what they are doing. If Sun started to make major changes to Linux, than that takes away control from the open source community. Linux is the platform that many vendors can base a product line on, and that’s the primary difference between Linux and all other operating systems. Linux supports fair competition by remaining neutral and stable.
Some guy opening 50 random .doc files from google with only a %10 failure rate doe not equal interoperability.
Throw in some spread sheets, some power point, an exchange server, and some visio diagrams and you are playing another ball game.
Before I had MS office at home I tried using OOo just for reading some of this stuff from work and my success rate was not pretty. I could parse allot of the stuff, but I don’t think a single document was correctly formatted.
OpenOffice.org and StarOffice have the best MS Office compatibility of any suite I have ever used.
I can share documents with coleagues, including documents that use Word’s revision tracking, and they work fine in OO. My coleagues don’t even know I am not using MS Word.
Most PowerPoint presentations work fine too, except for the cutesy animations that don’t belong in a presentation anyway because all they do is distract from your content.
BTW, one more great thing with OpenOffice.org
I submitted a bug report to OpenOffice.org about the Impress module. I had a response in 2 days. The bug was fixed within a week.
Try that with Microsoft and MS Office.
I’d like to see a third and a fourth enterprise competitor base a product line on Linux. I guess that Novell is going to make some waves, however only Sun is focused on the Linux desktop. All of the other vendors that deploy their products on the Linux platform are concentrating on the enterprise. I guess that the only desktop efforts are Fedora Core, JDS, and this OSDL team of 10 programmers who want to establish some standardization of some sort.
The “Five myths about ‘Wronghorn'” link on the article are very coherent to me.
The main quastion on this article is:
Microsoft will support some type of open standards??
Microsoft is lying about your “cross-platform code”??
How standards Microsoft will show to us??
Win32 will become fully incompatible with new WinFX??
How much of the “five myths” of longhorn are true??
How Microsoft will manage your customers after they descarted 27% of windows user base(the windows98 user base, data of http://www.google.com/press/zeitgeist.html)??
Is true the Microsoft are focing the migration of old platforms like windows 98??
Some explanation about the use of Google zeitgeist as statistic: the win98 user base are in majority SOHO and corporation/gov legacy systems, and good part of them are off-line machines doing only simple tasks or only have access to internal LAN doing simples tasks too.
Yes, this article sounds like a great trolling to me too, but the questions have some sense.
What happens to QT, KDE, and QT apps. Surely they would require them to get any serious bleeding of average windows users. There simply aren’t enough gtk/gnome/java apps for a diverse user base. I’m sure if they keep thier release schedule slow enough then someone will setup an optional QT/kde addon for it. Personally though, I would prefer something similar in design to Mandrake get a high marketshare, it is fast and simple/easy, without losing configurability, and it is the most DE/toolkit neutral “easytype” distro out there. I will continue using gentoo either way though.
What about the “400 million Windows 98 users”? These figures are simply made up. The latter does not make sense at all given Google’s stats for desktop useage
This looks at people with internet access. Many companys don’t give this out for “normal” users, and it is them who get the hand-me-down/cheap-n-chearfull 95/98 boxes.
“What happens to QT, KDE, and QT apps. Surely they would require them to get any serious bleeding of average windows users. ”
Not really, when you consider what most people use their computers for. Most people use their computers to surf the Web, do email, write letters, and maybe edit the occasional vacation photo.
In that case, Java Desktop has all they need. Mozilla for Web and email, StarOffice for writing letters, and GIMP (a GTK app) for editing their vacation photos.
The .NET runtime is also buggy. After installing it, my XP box started to hang on shutdown. After I uninstalled it, the problem went away.
what programming language should I learn next? I really wanted to start helping out on some local open source projects(a mini OS and a 3D game engine),so Id like to be able to learn rather quickly.
C hands down. More than likely it’s what the OS you want to help with is written in. Plus once you have learnt C; C++, and Java will come easily.
Jared
I wouldn’t say C. Learn to walk before you run and try Pascal. It’s limitations may grow irritating, but it helps getting the initial concepts driven home. There are two free Pascal compilers I know of, search around, and find a good tutorial.
I’d go with Java. C and C++ are easy to learn after, and you don’t have to worry about choosing a toolkit, or the evil sides of C(++).
I second the Java sentiment. C wouldn’t be hard to learn afterwards. My university switched to teaching Java after I passed throught the programming classes. And it does seem like a good way to learn.
C is still THE language for OS’s and maybe 3d games, but I think its easier to learn (the C language) after java rather than before. I wouldn’t bother with Pascal. From what my professors told me, Java is a step forward in teaching languages, as well as in general.
I would third that. Java also has some nice development tools, and looks nice on a resume.
Learn C and xlib next, that will teach you how to build GUI applications.
I have been a full time programmer for a few years now, personally the best advice I can give you is that you learn a scripting language first, as they are fast and easy, try learning how to script with PHP, then you should be able to hop on to C, thats how I did it.
Don’t learn java first, learn it once you are confortable with that style of language — PHP, then learn java.
it’ll damage you forever.
http://www.joelonsoftware.com/articles/fog0000000319.html
“Learn C and xlib next, that will teach you how to build GUI applications.”
Oooh – you sadist!
Seriously though, Python is a good language to learn. It has a clear and simple syntax, yet is far from being a toy language. If you are not familiar with object-oriented programming (which is necessary with Java), is can teach you lots about that.
In addition, you can compile a lot of Python code to Java bytecode using Jython.
You can clearly tell that this is a paid-for advertorial by Sun because no-one else so far has discovered what the “Sun Java Desktop” has to do with Java.
It starts to become hypocritical when he calls Microsoft’s .NET strategy confusing. I don’t see the difference really, Microsoft calls completely unrelated things “.NET” and Sun calls completely unrelated things “Java”.
“It starts to become hypocritical when he calls Microsoft’s .NET strategy confusing. I don’t see the difference really, Microsoft calls completely unrelated things “.NET” and Sun calls completely unrelated things “Java”.”
It has to do with the future. And the future is .NET and Java. The future is managed code for all applications.
Sun is basically pushing Java as the language of choice for those applications, while Microsoft will be pushing C#, or managed VB, or managed C++.
Java is a good first language to learn.
As far as C, yes, if you want to to OS level work, C is a good language, but you are not going to learn C quickly.
The traditional wisdom has been that one should learn C before C++ for example, because that allows you to learn programming without also having to grasp object oriented concepts at the same time. However, many, including me, who has been programming for about 15 years, are really starting to question this logic. So are many Universities who are starting to teach Java as a first language.
Basically, the idea is that one of the reasons object oriented concepts were so hard to grasp, is that programmers had to unlearn a lot of ways of thinking about data and procedures. It seems that learning object oriented concepts comes more naturally if you learn about them right away when you start learning programming.
I would definately not start with C as your first language. Start with Java.
“I find it crippled by technical design faults and a design centered on a single operating system”
dude must be smokin’ some kind of drugs
“Both “versions” of .NET’s lack a clear technical advantage”
there is a lot technically advantage and besides that a lot other neat things like syntax sugar etc. I happen to find advanced .NET things like Reflection, Attributes, CodeDom etc. extremely useful and I use it regularly, try to do such things with Java. It simply lacks a lot things or it’s pain to use. I never wrote a single line Java so I might be wrong but after all I’ve read Java lacking so many things without even supporting essential features like properties, enums etc. I really can’t take that language serious or bother to take a closer look at it. Generally I’m opened minded towards new plattforms and languages. Again I might be wrong but after endless nights reading sites like OSNews I think the Java plattform seem to be a poor excuse of a programming model compared to .NET/Mono
“there is a lot technically advantage and besides that a lot other neat things like syntax sugar etc. I happen to find advanced .NET things like Reflection, Attributes, CodeDom etc. extremely useful”
I can do it in Java. And I can do it with less code than C#
“It simply lacks a lot things or it’s pain to use. I never wrote a single line Java”
If you’ve never wrote a single line of Java, then how can you say it is a pain to use or lacks such thing? Probably shouldn’t critisize it unless you have tried it.
So you find those things useful in .NET huh? Do you also find the fact that .NET is over 200% slower than Java on average (and far more slower than Java on some tasks) to be useful as well?
I guess ill start out with some scripting(PHP/PERL) and then move up to Java.Afterwards ill give C/C++ a try.
Thanx all
Complete and utter rubbish.
” you also find the fact that .NET is over 200% slower than Java on average (and far more slower than Java on some tasks) to be useful as well?”
If you’re going to troll, at least put a little effort into making it believable.
I teach classes on J2EE (enterprise java) and on ASP.NET. I have studied both technologies and have come to appreciate certain qaulties of both. I also dabble in various open source projects. So I am interested in all the issues presented in the article.
I believe the article is just plain FUD. Don’t pull out your guns and call me a Java hater. I’ve used Java for 7 years and I am a sun certified Java programmer. And don’t call me a Linux hater, I’m not.
But I do have to disagree with Mr. Adelstein on .NET. The article is a jumble of ideas without any clear logic or presentation. It keeps hiniting that ‘.NET’ is scary but has no evidence. I agree that there are some things about .NET that could be scary – but the article does not focus on those. Instead the article focuses on:
1) ‘Java Language Conversion Assistant’ – so what, there is nothing evil about a conversion utility. There is also a VB 6.0 to VB.NET conersion utility, should I be shaking in my boots? If you don’t like .NET then make a .NET to Java utility but don’t keep repeating ‘Java Language Conversion Assistant’ and expect a rational developer to be upset.
2) ‘The confusion of the .NET name’ – there is confusion but the article does not address the confusion. Instead the article seems to be saying ‘is web services part of .NET or not’ and the clear and definate answer is yes, it is. I can only guess, since the article was so poorly written, that the author means to scare us – but about what I cannot say. Are we supposed be scared that web services are in .NET or are supposed to be worried that web services may be removed from .NET in the future? I really don’t know what his concern is. My opinion – web services in .NET is good and they won’t go away. If he thinks web services are bad then please explain, and if not then what is his problem with .NET having web services. And why does he use the ‘naming confusion’ topic to stir up a reaction when he does not present a conclusion as to how that naming can really hurt the industry.
3) ‘Only a slim chance exists that someone will ever run C# code on Linux, Solaris, OpenBSD’ – this is a complete lie and a slap in the face of the Mono project. Again, this is FUD not logical discussion of the topic.
4) ‘How effective or important do people see the need for computer languages to interact at the Application Programming Interface (API) level with Windows and .NET? As a programmer, how does that benefit me?’ – this statement is so stupid it is beyond words. The API is (among other things) how one accesses the mouse, keyboard, video, audio, and files. Humm, how important is it that a computer language be able to do IO? Gee, pretty darn important mate.
This is where the FUD lays – if he had just said ‘Why make a version of Cobol that uses the .NET API?’ I would have agreed, but he is using language the looms larger attempts to leave the reader worried.
My opinon – Java was right to promote one language and MS is silly to think that Cobol.NET or Perl.NET or JS.NET are useful. I dislike all .NET languages I’ve looked at except for C#. I teach only C# and warn people that VB.NET is an illusion – it is not VB, it is a new API and therefore disimilar from the old VB in many ways. C# is simple and easy to learn, if you have to learn something new you may as well learn the best language for the job. And I think that what .NET did to JavaScript is a offense to the language. The beauty of JS is in its simplicity and that it is independent of any external API.
So altough I agree that the idea of multiple languages is mostly corny, I do not agree that the CLR is a bad idea. The CLR is a better design because it looked at many different languages as references during the design.
Again, this is just FUD. We are supposed to believe that the CLR is bad because the designers looked at many languages during the design phase?
5) ‘How will Microsoft maintain the thousands of language versions in the market? Who will they decide should get any air: Linux, GNU?’ – MS has no intention of supporting thousands of languages, again FUD through exaggeration. And Linux is not a computer language – what point is he trying to make? Perhaps another jab at Mono? Boy, I guess he sure hates Ximian? But if he does, he should just come right out and say so instead of making lot of little jabs.
6) ‘A library tagged with a sign that it might contain “unsafe” code doesn’t instill much confidence in me.’ – this is so *&#$* ignorant it goes beyond words. A statement like this makes me wonder if he is even a programmer at all. For the record: Java can call native methods. The practice of calling native methods is exactally the same as ‘unsafe’ in .NET. That is that they are both unsafe. And so is all C++ and C code and so are many other languages. Safeness just has to do with how the program can access memory. C or C++ can access memory in a bad way and can mess a system up. If the author hates ‘unsafe’ code so much then he not only hates all native methods in Java but he must also hate Linux since Linux uses C/C++. Once again this is just pure FUD – there is nothing wrong with providing the ability to have an unsafe method. What that allows the programmer to do is access legacy APIs and libraries that have not be ported to .NET yet. True those APIs may have memory leaks, but those are the same legacy APIs we have been using in Windows programs for nearly a decade and that does not seem to the problem the author has…
7) ‘Using the dot-Net branded Common Language Runtime, Microsoft will deprecate the entire Win32 platform. So, your beloved applications just won’t run any more.’ – obviously instead the author is quite worried about losing access to the Win32 API. Well, two points a) you can still access the Win32 API using ‘unsafe’ methods, b) yes, MS will deprecate the entire Win32 API (thank the heavens, I lothe that API with a passion). That does not mean the ‘beloved’ application will not work. Microsoft will support the Win32
API for at least a decade, probably longer. At the very least they will do so with VirtualPC [may bundle with future OS for seamless legacy support], which they purchased recently, but more likely they’ll just keep the Win32 API tucked away in the OS in a ‘protected’ memory space so that all the bugs in Win32 are kept away from the safe .NET apps and away from the core operating environment.
(cont…)
(…cont)
8) ‘Java uses a different approach toward language interoperability. The Java platform uses focused, task-specific languages. Instead of promoting a collection of overlapping, general-purpose languages, the original Java Language programmers recognized that some tasks deserve special languages of their own. … For example, we access databases with the Structured Query Language (SQL) and change XML documents using XSL Transformations and XML Path Language. Java programmers solve quick systems-related problems using various shell scripts.’
– Really? I suppose ‘the original Java Language programmers’ created SQL, XSLT and XPath? Not! And I suppose that .NET doesn’t support these languages? Of course .NET supports them. You cannot hold up XSLT and SQL as an example of why VB.NET is a bad language. That just isn’t logical. So guess what, ‘Java uses a different approach toward language interoperability’ is just plain misleading. More FUD.
Indeed, not only is it the case that Java developers did not invent SQL, XSLT and XPath; but it is the case that they added the support for these languages due to market demand – support for those languages does not have to do with some grand vision ‘task-specific languages’ but instead with filling holes in Java functionality that are required to operate in the business community.
The best example the author could have used would be EJB-QL. EJB-QL is Sun’s replacement for SQL in J2EE – specifically when using Container Managed Presistence and Container Managed Relationships. EJB-QL is an example of Sun putting forth a vision for ‘task-specific languages’ rather than just adding on support for existing standard languages. However, EJB-QL has a very limited and focused use – only in an enterprise Java container.
So I believe the authors attempt to differentiate Java and .NET is generally inaccurate.
9) ‘Java and the languages it supports can run everywhere.’ – again, I’ve developed in Java for 7 years and am a java certified programmer and I teach Java classes. I have no overt dislike for Java. But Java does not cut it as a user interface. Swing and Java are too slow. The author cannot just say ‘run everywhere’ and assume that we all should just live with the speed issue. when you can honesty say it ‘runs quickly everywhere’ then you have a point. I teach Java on servers (J2EE) and there it performs well enough. But then we are comparing apples and oranges – .NET is not intented for just server applications. The fact that .NET runs GUI apps faster than Java is not a disadvanage of .NET.
10) ‘… stuck with Windows unless Novell can fulfill the promise of Ximian’s MONO …’ – and why wouldn’t they? does he have some secret knowledge about the project falling apart? or is he just trying to create FUD?
11) ‘Java developers grow architecture supported by most of the industry. That may change when the Win32 standard disappears.’ – ah, so now you are saying that Java depends on the Win32 API? I thought he said Java runs every where – not just Windows. Then it must not be linked to the Win32 API too deeply, right? Hmm, for an open source guy he really seems to like Win32.
12) ‘Java developers can often download “early-access” releases of new technologies’ – yeah, and we also can download the Longhorn SDK right now, 4 years
before Longhorn is released.
13) ‘One might consider the saying that those who forget history may be destined to repeat it and that everything is the same as everything else only different.’ – repeating sayings, without clear and logical argument, is not constructive – in fact it seems like FUD to me.
well this is enough knee jerking on my behalf. i’m just upset to see an article like this being linked with open source. it gives open source a bad name, and Sun and Java a bad name. the real concerns with .NET are Microsoft’s licensing systems (both developer and end user) and Microsoft’s patents/intellectual property. But rather than discuss these concerns in a clear and thoughtful way the author prefers to focus on:
a) .NET name confusion
b) taking shots at .NET (e.g. unsafe code)
c) throwing around the term Web Services
d) continuely denying the existance and success of Mono
e) claiming Java is superior without rational evidence
f) attacking the CLR for being a well thought out design, one that could accomadate some variation in language syntax
g) threatening that Win32 won’t work anymore – which is pure speculation. nearly everyone in business knows that this would be a death kiss for Longhorn, of course Win32 will be supported and for a long long time.
Unprofessional and misinformed is the best way to summarize this article.
“3) ‘Only a slim chance exists that someone will ever run C# code on Linux, Solaris, OpenBSD’ – this is a complete lie and a slap in the face of the Mono project. Again, this is FUD not logical discussion of the topic.”
Uh huh… Tell me. Have you bothered to look at the amount of the .NET specifications that Microsoft has actually made available to the public? It’s barely enough to write a C# compiler. Mono is a joke. And cross-platform .NET is a joke.
“But Java does not cut it as a user interface. Swing and Java are too slow.”
I guess you have never heard of Eclipse? Or other native widget toolkits for Java?
And as far as Swing being slow, apparently there are a lot of people who don’t agree with you that it is not usable for UIs. Take a look at Swing Sightings on the Java web site.
And if you meant Java itself was slow, .NET is quite a bit slower.
“12) ‘Java developers can often download “early-access” releases of new technologies’ – yeah, and we also can download the Longhorn SDK right now, 4 years
before Longhorn is released.”
No you can’t. All you can download is a subset of the WinFX API, which from the tests I have done when deciding whether to go with .NET or Java on my latest project, is slower than a turtle. The .NET runtime engine is horrible. At is several times slower than Java.
“d) continuely denying the existance and success of Mono”
Mono exists. But is it or will it be successful? No. If for no other reason than that philosophically, a lot of open source developers will refuse to use it because it smells to much like Microsoft.
“g) threatening that Win32 won’t work anymore – which is pure speculation. nearly everyone in business knows that this would be a death kiss for Longhorn, of course Win32 will be supported and for a long long time.”
No one is threatening that Win32 won’t work anymore. The threat is that Microsoft will support Win32 by mapping the Win32 API to equivlaent WinFX function, or that somehow or another, they will force Win32 apps to undergo all kinds of runtime checks, and thus make them run slow as a dog. That will undermine competing strategies on Windows like Java that will then have to go through the Longhorn runtime checks, as well as its own runtime checks.
If you think that that this isn’t all about Microsoft tightening its stranglehold on the industry, and if you think that Microsoft won’t do something to intentionally undermine their competition, than you haven’t been paying attention.
The reason it is called Sun Java Desktop is because while most corporations will most likely avoid anything called Linux like the plague (despite the efforts of their admins to convince them) they will buy SJD because in it they see 2 things they already use on a daily basis : Sun and Java. It is all about smart marketing at stupid management.
My answer? As many as you can! Seriously
It shouldn’t take more than a week to pick up a single language (you can learn Python in a day, for example). After you learn several, you can specialize in a couple (you’ll know which ones you like). Do not learn languages based on which ones you think are popular — not unless you want a job in programming, anyway. A lot of very cool software (mldonkey, emacs, etc) has been programmed in obscure languages. And you can access GTK+ from pretty much every language under the sun, so that too should be a non-issue. Now your choice of problem-space is going to limit your choice of languages somewhat — C++ is a good choice for 3D (most of the OSS 3D engines are in C++), while Python is a good choice for GUIs or XML-processing programs.
9)”Java and the languages it supports can run everywhere. – again, I’ve developed in Java for 7 years and am a java certified programmer and I teach Java classes. I have no overt dislike for Java. But Java does not cut it as a user interface. Swing and Java are too slow.”
>> There are smarter ways to make the Swing code faster – i-e:- Swing worker thread (ofcourse depending on your application), profiling and lots of stuff, leveraging flexibility of GCs in hot-spot, etc. The *quintessential high-point* of Java platform is WORA and not about speed –
why do you care when the world is embracing 64-bit procs., 3+GHz….
“The author cannot just say ‘run everywhere’ and assume that we all should just live with the speed issue. when you can honesty say it ‘runs quickly everywhere’ then you have a point. I teach Java on servers (J2EE) and there it performs well enough. But then we are comparing apples and oranges – .NET is not intented for just server applications. The fact that .NET runs GUI apps faster than Java is not a disadvanage of .NET.”
>> Of-course .NET GUI apps will run faster than Java – the semantics of the CLR are different from that of JVM friend, as u would ‘ve known. The MSIL is far-more low-level (layer-wise) than Java Byte Code – as told by a MS rep. in a talk, when I questioned him;
“Java Language Conversion Assistant’ – so what, there is nothing evil about a conversion utility. There is also a VB 6.0 to VB.NET conersion utility, should I be shaking in my boots? If you don’t like .NET then make a .NET to Java utility but don’t keep repeating ‘Java Language Conversion Assistant’ and expect a rational developer to be upset ”
>> Java -> .Net = Java + .Net extesnions (say, CLR compliant) And for security purposes I deploy this in Linux,
Solaris – I can’t assure it will work fine unless, say mono’s CLR is 100% compliant with MS’s..this dilemma will make me to go for thw suggested *migration path* to switch to .NET/windows. Another thing is we have this tool ‘cos C# ~= Java (semantics, syntax) wise.
“‘The confusion of the .NET name’ – there is confusion but the article does not address the confusion. Instead the article seems to be saying ‘is web services part of .NET or not’ and the clear and definate answer is yes, it is. I can only guess, since the article was so poorly written, that the author means to scare us – but about what I cannot say. Are we supposed be scared that web services are in .NET or are supposed to be worried that web services may be removed from .NET in the future? I really don’t know what his concern is. My opinion – web services in .NET is good and they won’t go away. If he thinks web services are bad then please explain, and if not then what is his problem with .NET having web services. And why does he use the ‘naming confusion’ topic to stir up a reaction when he does not present a conclusion as to how that naming can really hurt the industry.”
Yes, its quite true that .NET is nebulous, (UD of FUD) ad Bill Gates himself said in a interview (news.cnet.com, I can’t remember the date) that he would a grade “F” for their progress in convincing the masses abt .NET. If you want the application layer be web-services aware why bake the so called .NET into the OS (here’s where the confusion is) – since MS is known for rich, developer friendly libraries. Sun, say could afford to bake JVM in desktop/servers and make all Java apps. look native.
I am a regular user of Linux, Java/C programmer. I greatly APPRECIATE this article – a true unabashed insight to dispel the misinformed notions of JDS..Indeed JDS is a foundation for a wave-of Java technologies to invade this desktop. (Project looking-glass being one of those *technologies*). It will dawn on those OOo/JDS bashers here later, that JDS will slowly, steadily and surely be populated with good Java apps. and usher a new desktop era – Java APIS/SDK to OOo etc. – I am not blah..blahing in a dream..please:)
My first programming language was javaScript because it was available on my school pc. It was a great start but programming isn’t about the language but about getting a job done. I personally learned C++ at University and taught myself Java at the same time as an exercise. If you are going to be a programmer you should be able to do both.
The funny thing about programming is when you learn something about one language it tells you something else about another one. And different language types like procedural(c) and object oriented(java) offer different advantages based on the job at hand.
So my basic advice is learn them all. At the same time even. You will then be able to make a judgement call about what works for you, not what works for the posters here. (We all think differently you know).
Last figure I heard from Sun was 3m.
What version of OOo did you use, Jim? Have you tried 1.1?
All in all, the article seemed pretty clear and well-put to me. Okay, so the guy is a Java advocate, but he clearly states some very rational reasons for his view. Which is the point, surely?
I’ve just scrolled through this. Jeepers, the whole point of calling it “JDS” is to leverage the Java brand and spread it across their product line. People know Java, they don’t know Linux. Why not leverage the marketing power of one to the advantage of another? Why spend money trying to establish a *new* brand?
As for the future, who *cares* whether JDS, Red Hat or SuSE/Novell/Ximian is a player, isn’t the more important thing that people are *moving* away from Microsoft?
Some people here really don’t have a clue AT ALL about running a business or marketing products. Here is a hint for the geeks and so-called “business gurus”, stick to your local school cake stalls and stop making commments on issues you have absolutely NO idea about.
> Some people here really don’t have a clue AT ALL about
> running a business or marketing products.
I do have a clue about making IT purchases though and this Sun marketing crap doesn’t impress me.
> And Sun does advance Linux by helping fund GNOME and Ximian and such.
I find that hard to believe, why would Sun want to fund the development of Mono?
>>> And Sun does advance Linux by helping fund GNOME and Ximian and such.<
I find that hard to believe, why would Sun want to fund the development of Mono?<<
I don’t know anything about Mono, but Sun does have their hand in a lot of open-source projects, including GNOME and Ximian, and not to mention Mozilla.
Scroll to bottom:
http://foundation.gnome.org/
http://www.ximian.com/about_us/partners/profiles.html
He speaks of misunderstanding Sun, but he horribly misunderstands IBM. IBM is NOT a consumer desktop company. They are big iron. HARDWARE. They’re push for linux is not to give the usr a fancy-schmancy desktop. It’s to make linux run nicely on their hardware, so they can sell more harware. If the customer chooses their desktop elsewhere so be it, they want the backend services running on their gear. Maybe he should look into their strategy. This strategy has been obvious for decades, and is still unchanged.
“I find that hard to believe, why would Sun want to fund the development of Mono?”
mono != GNOME
And I don’t think Sun has much to worry about from mono.
> Some people here really don’t have a clue AT ALL about
> running a business or marketing products.
I do have a clue about making IT purchases though and this Sun marketing crap doesn’t impress me.
Great, now, while you’re at it, why don’t you go into some REAL facts and REAL details about the advantages of you chosen dogma over SUNs product range?
Thank you for cutting through the FUD.