Submitted by Attachmate now owns Novell and therefore, by extension, also owns SUSE and openSUSE. With Oracle currently doing everything in its power to thoroughly destroy what’s left of Sun’s open source commitments, scepticism abound about the future of SUSE, and more specifically of openSUSE. Attachmate’s CEO has answered some questions about the future of SUSE and openSUSE, and as far as words go, it’s looking good.
Attachmate CEO Jeff Hawn talked to ZDnet, and has tried to address some of the concerns the community may have about openSUSE’s future. According to Hawn, the community involvement with openSUSE and other open source projects withint Novell is crucial, and Attachmate intends to continue these projects.
“SUSE sponsorship and participation in key open-source projects is a fundamental element of the business. This commitment is driven by a desire to contribute to and collaborate with the community in a way that fosters the success of open source technologies overall and creates the greatest value for our customers,” Hawn states, “The openSUSE project is a great example of vibrant and healthy collaboration. SUSE sponsorship and participation in projects like openSUSE creates great value for the community and also for SUSE customers who benefit from the innovations and advancements we create together.”
Furthermore, Hawn confirms that Novell/SUSE will continue their membership of the Open Invention Network and the Linux Foundation, which I’d say is good news as well. As for Novell’s and SUSE’s commercial offerings, it would seem that Attachmate is not planning on changing much, other than splitting Novell and SUSE.
“SUSE was acquired by Novell some time ago, and we see tremendous potential in this technology,” Hawn explains, “Our hope is to bring prominence to it by giving it individual branding as a separate business unit. By separating both Novell and SUSE, we can give each of these brands the focus they need to meet the needs of their specific customers and ensure that they are successful.”
In the end, only actions are what matters. Let’s hope openSUSE will continue to flourish under Attachmate rule.
I tend to get nervous though when companies split up business units like this, but my fear comes from what I remember with Palm when they separated their hardware and software divisions. This is a completely different case and in fact Suse/OpenSuse was on its own entity at one point before Novell stepped in. I fear when businesses separate acquisitions into subsidiaries like this, they are merely separating high risk business units away from the main company’s financials so that it’s easier to close shop and not affect the main business financials if it fails.
But like Thom said, compared to Oracle’s way of doing things, I really appreciate how Attachmate is allowing Suse/OpenSuse to operate like business as usual.
Edited 2011-04-29 22:40 UTC
SUSE is the profitable part, Novell is the high risk business. Netware, even if it still considered the best Unix by some IT pools, is dead. eDirectory lost to MSAD, the network part is unseen in recent serveur room. Whats left? Novell tried to save itself by becomming a new RedHat, but failed to deliver in most case and the “Open” part of SUSE came too late (SuSE 10.2 -> OpenSUSE 10.2). The way they handled locals partners to is a failure too (welcome to our big family! We promise you contract if you help us push our product deeper in your city, now, eat our dusts). Just like with eDirectory, it seem that of their “open” project are missing some bits of code to make them work fine on other Linux (Mono with the codecs, Suse studio, built system and others).
Overall, I hope a SUSE separated from Novell with enough financial backing/trust from the parent company is a good thing to get rid of the old mentality, even if SuSE never had a clean record in the first place.
Netware is not Unix, not by a long shot.
Here is a quote I fully believe and completely backup:
Novell’s Directory Services is a competitive product to Microsoft’s Active Directory in much the same way that the Saturn V is a competitive product to those dinky little model rockets that kids light off down at the playfield. — Thane Walkup
I’m an NCE from the 2.x through 5.x years… Bindery was very good… NetWare v4.0 and 4.01 were complete and utter shiitake. I went through upgrades and nobody really understood a “directory service” about 4.10 they got it right. 4.11 was superior and it just went on from there. NetWare was by far the best File and Print services system at the time and GroupWise v5.5 turned out to be a far better e-mail system than Exchange ever was (and still it IMO… Exchange sucks in comparison). I just weep at the MSAD success, while NDS/eDir just laid down and fell over.
Its really STILL is a much better Directory Service, just that nobody uses it for anything more than an LDAP server on steroids now.
Oh well. lets move on. MSAD is shit, it has so many limitations “Per Forest” and “Per Tree” its not even funny. Collisions can and do still happen all the time. The replication between servers is shit (compared to eDir’s method) and MSAD’s replication is crap compared to slapd and slurpd.
Ok, so enough of MSAD being a crap… YUM, I love its taste… as long as I can wash it down with significant amounts of Dark Rum.
gfolkert – You are 100% correct.
I’ve had to use both heavily and MSs directory services are not anywhere near as good as Novells.
The problem with Novell is that they always assumed that the best technology would win and that everyone would play fair.
MS _can’t_ play fair. When you don’t have as good of a product you can’t play nice and expect to win. Especially when your product isn’t _anywhere_ near as good as the other companies (and I agree about it still not being anywhere near as good).
So what does MS do?
Lots of things. FUD of course but MS also plays to CEOs egos. I’ve met too many CEOs that believe that being able to hang around someone rich makes them feel a lot better. It’s like having a trophy wife. Hey, if that guy can marry _that_ totally hot woman then he must be really great even though he isn’t good looking.
That’s how the place where I work moved from GroupWise to Outlook/Exchange and are moving all services from Novell to MS (it’s literally on the timeline for 8 years – plus some things are moving to Linux – but not SUSE – from Novell).
Outlook/Exchange is horrible compared to GroupWise. It needs more servers and even doubling them is not enough to have Outlook clients “live”. Instead most companies don’t have Outlook in “live” mode like GroupWise runs with no problems. Instead they (and we) use cache mode. Pretty pathetic.
Don’t get me started about pushing applications (totally different subject than email). With Novell when you tell apps to push you can have them push right now, that second, to as many computers as you want (OK, there are limits). With MS the app doesn’t install right away. Instead it install anywhere from four to 24 hours later. And that is “normal”. How pathetic.
I could go on and on and on and on and on about how MS is nowhere near as good as Novell. Lemmings will be lemmings though. It’s like the merchant who convinces the king that only true royalty can see the material. Since the king doesn’t want to appear to be of bad (non royal blood) he lies and says he can see the material and the man sells the king the material.
People talk about Steve Jobs “distortion field”. What they don’t understand is that Bill Gates has a bigger distortion field and they have no idea how he is serving them **** for years and they are eating it up like crazy thinking they are eating very good food.
In some cases Sabon, we are getting very good food. As a long time J2EE developer (more than 10 years) who has been working with .NET for the past year, I really have to say that Microsoft produces a far better product when it comes to the whole Web application stack that they provide (ASP.NET, IIS, Visual Studio.NET, SQL Server). The fact that it is all designed to work seamlessly together produces a platform that is more robust and less prone to weird problems than any Java stack I have worked with. I’m convinced most of the problems with Java come from cobbling together technologies from different vendors to achieve the same kind of functionality that Microsoft gives you in one seamless stack that is all designed to work together.
And as far as development environments go, Visual Studio.NET is a better product than any of the Java development environments available.
Obviously, I’m not trying to suggest Microsoft is perfect. But when it comes to development of enterprise level applications, well, I can only say I wish I had tried the MS stack sooner. I spend a lot less time pulling my hair out than I do with J2EE / Java EE.
Edited 2011-05-01 01:29 UTC
There is the problem. Yes ASP.Net works when on a pure MS stack. Mono legally we don’t know since ASP.Net is not covered by the community promise.
Really lot of cases I find ASP.net worse than PHP. At least PHP I have cheaper server rental than java or .net.
When it comes to web techs. PHP is Number 1. Java is Number 2. And .net is tail of the hunt. Both Java and PHP has lot of premade applications to get you going with.
Do you have a link that supports that statement or are you just making those numbers up?
I actually saw figures supporting this just the other week while looking at different web app frameworks. Can’t remember where though but you don’t have to be a genius to know that PHP is the most widely deployed one and that .Net isn’t even on the map when it comes to public, high-volume sites. It’s a good guess that Ruby On Rails is more popular than .net too.
Could be different for Intranet stuff but Java is pretty entrenched there so it might be a toss-up.
The only .net web app of any stature that I can recall right away is Dekiwiki.
Edited 2011-05-01 11:22 UTC
PHP might be the most widely deployed Web framework, but when you are saying that, you are including things like say, Mom and Pop’s Diner, which consists of 5 pages. Maybe a dynamic menu, and an email contact form. The vast majority of mission critical enterprise Web applications are written in either Java, or .NET. (Think e-bay, your bank’s online banking Web site, online stock trading, etc).
The vast majority of PHP users are irrelevant when comes to business strategy planning at Red Hat or Novell. That’s because most PHP shops are very small, and don’t purchase support contracts (either because they can’t afford them, downtime doesn’t cost them very much, or their applications and OS use cases are simple enough where problems they can’t solve on their own are unlikely to crop up). Because of that, Red Hat and Novell can’t make any money off of them. But given the types of sites that are built on Java and .NET (stock trading, auction sites, etc.) these companies can lose hundreds of thousands, or even millions of dollars a minute when their site is down. For these types of users, running without a guaranteed 24/7 instant phone access to support contact is not an option. And that’s where Red Hat and IBM make their money.
Well, Dice.com has 7 times as many Job openings for .NET developers as it does for Ruby developers. So I think it’s a good guess that you are wrong about that. Anytime you are doing more than a simple CRUD app (and admitidly, probably 90% of small Web Sites out there are simple CRUD apps), the magic of Ruby on Rails disappears, and perhaps even starts to become a liability over time. Ruby on Rails still hasn’t really gotten any traction in large scale environments (and I’m not really sure PHP has either.)
Well, if you’ve ever ordered anything from Barnes and Noble online, you’ve used .NET. If you’ve ever ordered a computer online from Dell, you’ve also used .NET. And of course, Hotmail (or Live Mail, or whatever they call it these days) is built on .NET).
When you consider “applications” of stature, you need to considering vertical market applications that power billions of dollars of financial transactions every day. There’s a good chance you use these applications on a daily basis without even realizing it, either directly or indirectly (if you visit an ATM, there’s a good chance the back end software that handles your transaction is a service written in either .NET or Java, for example). A lot of times, Java and .NET don’t appear as popular as PHP because it’s an “out of sight, out of mind” kind of thing. People just aren’t aware of how much both of these technologies impact their lives because they don’t see them running behind the scenes.
Also, believe it or not, most of MIT’s applications are built in .NET.
Edited 2011-05-01 15:25 UTC
You don’t have to run a pure MS stack to use ASP.NET of course. You can use it with Oracle, MySQL, DB2, or any other database for example. But again, there is something to be said for having a single point of contact when something goes wrong as it eliminates any ambiguity as to who’s responsibility it is to support the issue.
And that’s fine, if your biggest concern is cheap hosting space. But if your biggest concern is that you lose you hundreds of thousands of dollars for every minute of downtime you have, then having a well supported stack where you can get a support team on the phone 24/7 that knows the entire stack inside and out, because they are the guys that wrote the whole thing becomes a lot more important.
Well, PHP might be the most popular as far as sheer number of PHP Web applications deployed. But again, keep in mind that many PHP Web sites are very small, and PHP might be used for nothing more than providing an email contact form or something.
There are a lot of pre-written applications in PHP, yes. But that only matters to me if I am doing something like trying to set up a blog, or wiki, or forum site. PHP has done a pretty good job at producing “generic” horizontal reusable applications like that.
Both Java and .NET on the other hand, tend to be used a lot more for developing vertical market specialized applications that by their very nature, are not reusable because of their specialization. And are not “generic” like a blog, or wiki.
Edited 2011-05-01 15:30 UTC
I’m Replying here because of the stupid “hide nested comments” here on this site.
I knew you were just making stuff up, because Java is almost not being used any more and nobody is using ruby etc.
The data from the top 1M websites are as follows:
PHP: 76.2%
ASP.NET: 22.7%
Java: 3.2%
ColdFusion: 1.3%
Perl: 1.1%
Ruby: 0.6%
Python 0.2%
Source:
http://w3techs.com/technologies/overview/programming_language/all
The number they have for Java is total BS, and any amount of research would prove it is obviously wrong (as would any understanding of how Java powered Web sites are set up). The reason it is total BS is because there is no way for you to tell that a Web Site is using Java. Unlike PHP, which typically can be discovered because Apache will report that it is running mod_php if you query it for what modules it has installed (a security hole btw, that server admins should disable, but most of them don’t) most Java sites just report that they are running Apache HTTPD server because Apache proxies requests for dynamic content to a Java application server behind the scenes. So any attempt to tell how many Web Sites are using Java is completely flawed.
Here are some statistics that are not flawed however:
* Java is the #1 language on Sourceforge.
* IBM, Red Hat, Oracle, VMWare, and HP are all strongly behind Java.
* 100% of Fortune 1000 companies use Java.
And here are some current numbers from dice.com as far as the demand for various languages. From most demand to least demand:
Java: 15911 job listings
.NET: 9084 job listings
PHP: 2996 job listings
Python: 2447 job listings
Ruby: 1478 job listings
Coldfusion: 335 job listings.
Java is clearly the dominant platform. The only platform that is even on the radar as a potential threat is .NET.
Also, according to Simply Hired, demand for Java developers increased 10% between September of 2009, and January of 2011. A 10% increase in job demand (even though we went through the worst recession since the 1930s) would be a good trick to pull off for a language “that is hardly ever used anymore” as you put it. Demand for PHP jobs also increased 10%, which means that PHP made no gains on Java at all in that period.
.NET was stagnant and saw no net increase in job demand.
So yeah, I’m afraid I have to call that statistic you posted exactly what it is. BS. And if you read the disclaimer at the site you got it from, and if you knew how most Java application servers are invisible because they run behind an Apache instance acting as a proxy, you’d know why those statistics are BS.
Edited 2011-05-02 06:19 UTC
I would also point out that everything I said about why the Java number is BS also applies to why Ruby on Rails is under-represented. Because like Java, Ruby on Rails is typically configured to use an application server such as Mongrel that runs behinds an Apache instance. Once again, Apache proxies requests for dynamic content to the Mongrel server. Because of that, as with Java, it’s impossible to tell that a site is running on Ruby on Rails.
This method of deployment is radically different than the method that PHP uses of embedding the PHP interpreter inside Apache.
I was not aware that sourceforge is only used for creating publicly accessible websites.
Likewise IBM, Red Hat, Oracle, VMWare, and HP only business is to create publicly accessible websites.
What was being discussed was web technologies,and these days java in LOB applications is almost exclusively an intranet technology and is also disappearing there for new projects.
Java is however straddling a number of technology segments, like LOB, embedded and mobile and will of course hold out for a long time before it disappears completely, kind of like cobol.
Ah. So that’s why e-bay is powered by Java. And why IBM and Oracle’s blogging sites are powered by Java. And why Google App Engine is built on Java. And why e-trade is built on Java. And why Twitter’s back end is built on Java. And why a lot of open source software sites use the JIRA issue tracker, which is built on Java. And why Wells Fargo’s online finance applications are built on Java. And why L.L Bean is built on Java. Because no one uses Java to build public Web sites anymore. I get it now.
Got a citation to back up that claim? I’m quite certain the reason Google spent millions of dollars to develop support for Java on Google App Engine is NOT because no one is writing new Web applications in Java. I’m also quite certain they didn’t develop GWT because no one is writing new Web applications in Java.
You also completely failed to address my point about why attempts to determine how many Java Web applications are out there are flawed. Because unlike PHP, you usually can’t probe a a Web site and find out that it is using Java.
There is also something else you are overlooking, which is that we have hit the limits of Moore’s law. The free ride that developers have been enjoying of faster CPUs has come to a halt. Intel has been warning developers that this was going to happen, and that developers were going to have to start writing their applications to be concurrent as we start adding more cores because of the fact that we can’t make them go faster anymore. Within a few years, Intel will be producing desktop CPUs with 32 cores.
What are these massively concurrent applications going to be written in? Python? Ruby? PHP? Concurrency is notoriously difficult to do well in any of these languages.
Once again, enter Java. Several new functional programming languages, such as Clojure and Scala, have been gaining a lot of popularity in recent years, specifically because of their concurrency features (immutable data structures, software transactional memory, etc).
So again, your argument that “no one is using Java anymore” is flawed. Java has been well ahead of the concurrency curve for years, running and scaling well on systems that have thousands of cores. Python, Ruby, and PHP have a hell of a lot of catching up to do if they want to compete with Java in this new world where developers cannot rely on faster CPUs anymore, and are forced to deal with writing massively concurrent applications.
Edited 2011-05-03 14:57 UTC
I take that back. There is one Java IDE that is as good as Visual Studio.NET. IntelliJ IDEA. It’s the only one that can compete with Visual Studio.NET.
Eclipse has been stagnant for years now and has been showing virtually no innovation. And Eclipse WTP is terrible when it comes to Javascript development.
NetBeans had a lot of promise, but it seems Oracle is intent on discontinuing development and support for half of it’s features. So I don’t know what the future holds for NetBeans anymore.
IntelliJ Idea is awesome though, and can effectively compete with Visual Studio.NET quite well.
Have you ever used AD? We are currently using Netware 6.5 and AD and I could tear my own eyes out every time I have to touch the Netware box. Console One is a joke It’s slow, creaky, and barely usable. Takes minutes to load on my quadcore system. AD users and computers loads almost instantly, and has never crashed on me, can’t say the same of Console One.
The Novell Client for Windows is one of the buggiest pieces of shit I have ever seen.
Zenworks is a monster that is clunky and virtually different every version. It tries to do too many things, and ends up doing none well.
SIPS sucks as a name service, it’s nonstandard, and seems to cause problems every other day.
Compared to eDirectory, AD is stable fast, and the administrative tools are built right into the OS, It uses DNS as a name service, so it is easily administered. GPOs are really nice, useful way to push policies, so useful, that Netware even supports them, instead of doing it itself.
The world has moved on to better things, maybe you should learn the new technologies, instead of lamenting a crappy product from a company that forgot how to compete.
I’d sure like to know why more of my AD policy doesn’t actually take affect on the client nodes. The central AD policy provided a security template rule to not run removable media on insert yet there I was having to add a local GPO on each machine before they stopped auto-running removable media. I’m also still writing long login scripts to cover things that should simply be policy rules.. and heaven forbid I have WinXP and Win7 trying to run the same login script without presenting a nice big error to my user. I’m still having to manage Exchange separately from AD (changes in Exchange flowing into AD instead of changes in AD properly affecting Exchange).
I can’t compare with Novell as the last time I worked with that was Netware prior to eDirectory becoming a separate product. I’d sure like to see MS-LDAP be a little more effective at what it claims to do though.
Yes, Microsoft has always made sure of that though various changes in the underlying GSSAPI interfaces, etc. Novell could certainly win an Anti-trust suite against them on the matter if they ever tried. Sadly, they never have tried and Attachmate probably won’t try either.
I’m sure that in the 10 years XP has been supported, or the 5 years that Vista has been supported, or the 2 that Windows 7 has been supported, Novell might have been able, at some point to put out a workable client. They haven’t been able to.
This argument is just excuses, and nothing more. The api hasn’t changed between SPs, just major releases. Nice troll though.
I’m sure that in the 10 years XP has been supported, or the 5 years that Vista has been supported, or the 2 that Windows 7 has been supported, Novell might have been able, at some point to put out a workable client. They haven’t been able to.
This argument is just excuses, and nothing more. The api hasn’t changed between SPs, just major releases. Nice troll though. [/q]
Actually, MS did change it between XP SP2 and SP3. I believe they may have adjusted it between SP1 and SP2 as well. I was researching at the time SP3 came out (trying to hook into GSSAPI for authentication), and the changes were quite substantial.
Can’t say anything per Vista/Win7 – though they had a dramatically different GSSAPI than XP did, from what I understand while Vista was in beta.
MS has been known for a long time – going all the way back to NT4 – for sabotaging the various systems that try to use the GSSAPI to do things like eDirectory/GroupWise so that their Domain system (now ADS) looks to be superior. Many just gave up and just piggy-backed on DC/ADS authentication (essentially capitulating to MS) – Novell with eDirectory was about the last ones to continue it to the depth they did.
Ok, so SP2 came out when? 2004.
SP3 came out in 2008, that meant Novell had 4 years to build a stable Client for Windows, and they couldn’t do it. They had 3 years before that. They’ve had 3 years since and the Novell Client for XP still sucks.
Argue all you want, but they haven’t changed it enough that it explains Novell’s incompetence. once in 11 years (XP) is not a valid excuse.
Vista came out in 2006, and they haven’t got it. Windows 7 came out 2 years ago, still haven’t got it stable.
so that’s 7 years for Xp to XP SP2, 3 years for SP3, 5 years for Vista and 2 for Win7.
Admit it, Novell sucks.
FYI – you’re just looking at the known public updates to the API. Remember, there is also the Windows Update mechanism too and Microsoft does push a lot of updates out through there that they do not publicize. So as to subtle changes that can break a lot of things like eDirectory there is more than ample opportunity for ways in which Microsoft has been known to use to break GSSAPI on Novell.
Now, you could argue that some of those changes were not intentional as Microsoft does not have a very good track record of pushing out patches via Windows Update that always keeps a fix in place. All to often one patch will undo another patch, and they’ll repatch in a third patch. (Several public examples of this throughout Microsoft’s history – e.g. WMF security flaws).
So whether it was intentional or not, it has likely happened in ways that Novell or anyone other than Microsoft (and in some cases even Microsoft) can account for in order to release a stable add-on product – which is what eDirectory/GroupWise is for Windows.
This is not a matter of making excuses for Novell. It is a matter of simply pointing out the history and behavior of Microsoft – known anti-competitive behavior at that.
Oh my god, can you make any more excuses? Novell has done a horrible job, the client sucks. It’s slow, it’s fickle, it just sucks.
Show me evidence, not tinfoil hats and conspiracy theories. Novell lost, and it’s their fault. They let NT and AD eat their lunch, and really poor quality software was a big part of the problem.
Obviously you are not any where near aware of Microsoft’s standard practice of operation. Nothing I have said is conspiracy theory/etc; but the same tactics Microsoft has used against numerous competitors from Netware to WordPerfect, and numerous others. The Antitrust trials in the US and EU slowed them down, but still did not stop the behavior – just made them try to hide it harder, and use some projects (IE) to faint improval. Yet, their tactics with WP7, SCOG (yes, they’ve been publically related to the SCOG efforts), and against Novell (of late – via CPTN and Attachment) are just the latest examples.
But them, you probably won’t both to go read all the condemning documents provided by the Comes v. Microsoft trial (http://www.groklaw.net/staticpages/index.php?page=2007021720190018) either.
No, Microsoft won’t ever change because all those practices are too ingrained in their culture. To change, they’d have to get rid of probably every manager for the first 3 or 4 tiers of management – which would just completely destroy the company. Replacing Ballmer would help, but his lieutenants would probably keep it up and subvert anyone who takes his place from trying to actually change – it’s just how they’ve been bread in the organization.
So please, go learn about Microsoft and how the really do business; why they are in the position they are in; and why most (like yourself) don’t think there are better products on the market – why products like NetWare and WordPerfect (that were dominant at one point) failed as organizations adopted Windows. And why Microsoft has been subjected to and found guilty of Antitrust violations twice already; settled with Novell over WordPerfect (in Novell’s favor mind you), and why Barnes & Noble is leveling Antitrust violations in the lawsuit brought by Microsoft against them (and numerous others).
No tinfoil here. Just reality when it comes to Microsoft.
Uh, yeah. Actually it is. My company is a Microsoft Business Partner, and Microsoft is very open about sharing information with us when we aren’t sure how to make one of our products work well with Windows. They have people specifically dedicated to helping developers integrate with Windows. Microsoft has worked with Sun to make AD and Java directory services compatible with each other and work well together. They have been very helpful in the development of Mono, answering questions and clarifying issues when the Mono developers aren’t sure about some part of .NET, etc.
Sorry, but everyone else has managed to make their software work well with Windows. Novell is the only company that seems unable to get things to work correctly.
The days of them being able to withhold information from third party developers are largely gone. That’s one of the things they got nailed for in court. For several years now, Microsoft has been very open about sharing information with developers. Novell has no one to blame but themselves that they can’t things to work well.
Edited 2011-05-03 14:36 UTC
Netware was a better product, until Netware 4. It then began to get worse, with each successive version. You can continue to tell me that changing APIs is MS’s way of doing business, but I believe that in this case, it is not the problem. By the time they had something workable again, Netware 6, it was too late.
The problem began when they rewrote netware in the 4 series. Netware 4 sucked, and you won’t find anybody who’ll argue that. It continued to be directionless, with multiple administration interfaces, shitty coding and crappy documentation. The Netware 4 fiasco happened at just the right time for MS and NT to take the business world by storm
Christ when you need to get a Netware Engineer to custom code patches to make your production operating system work, YOU ARE NOT DOING IT RIGHT. You don’t see anybody doing that anymore, not for a very long time.
Anyway, it comes down to proof, those APIs are public, and they haven’t changed all that much since Win2k. They changed quite a bit in Vista which came out over 4 years ago. Show me proof, or shut up.
I use Linux, I use BSD (when I can, I love it), I use Windows but if I had my way, I’d never use Netware again. It’s rickity, it’s administration tools are fragmented and hard to use, and it’s fragile.
I had to get a network engineer to remote in and apply a patch, because of “trade secrets”. My damn server kept crashing because of a buggy SCSI card driver. What the hell is that? That is no way for an OS vendor to be acting in 2009(when it happened). They should have issued the patch for everyone and be done with it.
There is a reason Netware lost the war, and it’s not MS, and MS hasn’t been able to get away with that kinda crap for a very long time, but keep living in the past, perhaps you’re happier there.
It’s definitely not the problem. In fact, In fact, people who make the “Microsoft is constantly changing APIs” claim are just parroting back crap they have heard from someone else, and have obviously never programmed against the Windows APIs.
The reality is that win32api has changed very little since for over 15 years. It’s hardly changed at all since 1995. WinFX. And even then, if a win32api developer were to look at code written for Windows 3.1, he would find most of it quite comprehensible. Microsoft tried to keep the API as similar as possible in order to make the transition for Windows 3.1 programmers easier.
The win32api has been far more stable than the Linux API has been over the years. That’s for damn sure. How many times has Linux changed the message bus we are supposed to use? Or the sound system we are supposed to use? First it was OSS, then ALSA… GNOME and KDE both had proprietary messages busses. Until they scrapped them and switched to a common one, etc.
That depends on which API you are referring to. Some have, like GSSAPI, have changed quite dramatically. Others change in very controlled manners (e.g. DirectX), while yet others never change any more or changes are introduced while maintaining the old APIs (most user space APIs).
However, if you look at the kernel space, then changes happen quite a bit more frequently. XP->Vista was a major kernel space change, Vista->Win7 was a minor one. Certain functionality – like GSSAPI – is a lot closer to the kernel than others.
Yes, Microsoft has stabilized the majority of the user space APIS – enough that most developers will never know a change occurs. However, when it comes to integration of competing products then Microsoft has done things – even selectively against specific products – to make those products not work as well. This is well documented in numerous court cases.
And – FYI – I have directly programmed to the Win32 APIs. They suck.
The only difference is that Microsoft makes the changes internally, behind closed doors. For instance, Microsoft moved the sound system and video system in Vista to user space. No one outside the development team really had input to how that was being done. They just did it.
Conversely, Linux had no sound system at one point. OSS was introduced. When they changed, the community banded together to make ALSA. OSS is still maintained – there are even modules in ALSA to support OSS interfaces.
KDE had DCOP for a long time – for KDE2 (October 2000) and KDE3. They only moved off of it to DBUS (standardized by the community via Freedesktop.org) with KDE4 (January 2008, equivalent release date). That’s hardly frequent.
GNOME went from Orbit to Bonobo (up to 2.4) to DBUS in about the same timespan.
I’m not saying FLOSS doesn’t change some things frequently – internal driver APIs for the Linux Kernel change nearly with every sub-release, but they also maintain them if you get your driver in the official Linux Kernel tree – but it’s no where near as bad as you make it out to be.
Not to mention that many FLOSS technologies can co-exist. D-bus and Bonobo can run very well right next to each other, and a simple program/daemon can convert between the two to enable integration. A deprecated FLOSS API can be re-instantiated by brining back the necessary support libraries; while a deprecated API from MS is gone forever (e.g. Win16 on Win64 systems).
I’m not saying it doesn’t suck. It does. It’s a real pain in the ass to program with. But is it any worse than say, programming Xlib directly? I’m not sure it is. Both are low level APIs that force you to deal with a lot of low level details.
Sure, parts of the community banded together to create ALSA. But as an application developer, I don’t remember anyone asking me if I was OK with that change or not. And I don’t remember having any input on it. Linux development is mostly done in an ivory tower just like Windows development really. It’s more open because multiple companies are involved and making the decisions. But the average Linux user / application developer has no say in which APIs change and how much they change. It is basically Red Hat, and a few other major players that make all these decisions.
Again, standardized by “the community” you say. But again, I don’t remember being asked if I was OK with having to re-write my applications to use the new message bus. I also don’t remember being asked if I was OK with GTK 2 completely breaking backward compatibility with GTK 1.
It’s worse than it is on Windows. The vast majority of Windows applications written for Windows 95 still run on Windows 7. How many Linux applications can say that without a recompile, and often code changes? Or without lots of cursing by end users trying to find out which versions of older shared libs they need, and how to get them?
That’s true. But Linux APIs break backward compatibility on a far more regular basis than Windows APIs do. Yes, you can install the older libraries (in most cases) to resolve that problem. But that’s a hassle. It’s one of the big reasons that not many commercial application vendors support Linux. Breakage because of changes in shared libraries is just far too common.
Edited 2011-05-04 17:35 UTC
If you used NetWare, continued to have problems, and keep using it, then the software isn’t your biggest problem. The lack of understanding of the platform and it’s limitations is your problem. Blind support of something that doesn’t work is probably a failure by yourself or whoever decided not to address the problem(s). Netware has essentially been a dead end for years. It was never a good application platform and the design of it made issues like your scsi driver worse than with other operating systems. Also, the core of the OS hasn’t been rewritten since version 3.x. Components have changed over the years like adding tcp/ip support, but it’s essentially the same kernel design without good memory protection in versions 3,4,5,5.1,6, and 6.5. Novell said netware was not going to be touched outside of some patches years ago. Their life cycle was no secret.
We still run some Netware servers, mostly because they run fine and we rarely have to touch them. If they were a constant hassle, I would have probably done something about it by now.
The NT/2000/XP client was ultra buggy in the early 4.x series. Since then it has been fine. The latest client for Windows 7 works fine too. Maybe you need to take a look at your environment and fix the issues.
The tools are mediocre. I complained to Novell for years about it. But, unless you’re using groupwise, or older versions of everything else, imanager works fine enough for everything else. Having to use Console 1 for the latest version of groupwise is absurd and lazy. Their model of splitting their software from interacting natively with eDirectory, started with zenworks, is their best bet. I hope they do it with groupwise soon.
You have to stop looking at Novell as netware, because those days are long gone.
I am just wondering where you get your information from. Show me anywhere that mentions that Novell Suse Linux business is profitable?
You can go back as far as 2009 through today and Suse Linux has not made a profit for Novell. If you find something to the contrary, post it here as an update.
Ok, the numbers are not that nice, bu the whole division still do better than most of Novell offering
http://www.novell.com/company/ir/qresults/
Novell do only 15% of its revenue from licences and 85% on services and subscriptions. The license money from the Linux division is non disclosed, probably bellow 1%, but the SUSE service account for 18% of the revenue, 54% if you take into account the whole ecosystem built on top of it (including NetWare virtualization solutions). But yea, look like the last time they were very profitable is when microsoft bought licences pack for HyperV, but it can be said for Novell too, not just the SUSE division.
In the right hands, any small project can flourish. In the wrong hands, even Microsoft could disappear virtually overnight. Played correctly, SuSE could replace Ubuntu as the entry level distro (when I started using Linux, it was in fact recommended to me over Ubuntu, which at the time was 5.04, Hoary Hedgehog or Warty Warthog, I forget which). Played correctly, it can serve the higher level users at the same time.
With Canonical too busy telling users what they want, rather than asking, Attachmate could find themselves with a lot of new users. Exactly what value this would be to them, how they want to monetize, I don’t know, but there’s definitely potential for becoming the dominant entry-level distro; and increasingly, people aren’t leaving the entry-level behind. They get comfortable there. I’d personally like them to utilise some of the good work done on Plasma and release a good tablet build of SuSE alongside their desktop offerings. Even if only a handful of people use it now, it’s a growing market, and will be increasingly important in the coming years. A strong focus on feature parity between x86 and ARM will be important, too, I think. With Windows 8 being able to run on ARM processors, we’ll probably see a lot of ARM netbooks, tablets, and even light desktops out in the world, and from my brief (by no means thorough) Googling, ARM can be a pain in the ass for Linux users.
I’m interested to know what problems exactly linux on ARM encounters. I suspect that it has to do with the much more proprietary hardware ecosystem, but am a lot more interested in what actually happens than what I think happens
Edited 2011-04-30 08:30 UTC
Lack of optimisation, because each ARM vendor makes his variant of the base design they licence from ARM. Texas Instruments opens well its OMAP variant, but nVidia is far slower and less open.
Add on that the fact there is a far larger variety of hardware on the ARM plaform (wifi and graphical chipsets, touchscreen, accelerometer, camera, and bootloader – chen it is not locked), and the net result is so much fragmentation and target hardware so small that we are largely dependent on the vendor to get enough code and data to adapt a functional linux.
As always, we are largely dependent on the community to code enough code and get enough data to adapt the Linux kernel.
Toshiba AC100 hacking efforts are an early bird of what will come, I presume.
I don’t see Canonical as being competition to Attachmate. Novell basically had four main competitors: Oracle, Microsoft, IBM, and Red Hat. Of these five vendors, Novell was probably the weakest because they are the furthest from being able to provide a full stack solution. By that, I mean they are the furthest from being be able to provide a single vendor point of contact for support.
In the open source world, we often like to tout freedom of choice as one of the benefits. However, if you are running a mission critical application that you need vendor support for, building different parts of your stack on software provided by different vendors can get dicey when it comes to support. If something breaks in your data access layer for example, who’s problem is it? The database vendor’s? The application server vendor’s? The ORM framework vendor’s? If your entire stack is from one vendor, there can be no ambiguity about who’s problem it is when something goes wrong, and who is responsible for supporting it.
Of the five vendors I mentioned above, currently, only Microsoft, Oracle, and IBM are capable of providing a full stack:
Microsoft has .NET, MS SQL Server, IIS, and Windows Server 2008. Oracle, thanks to the Sun purchase, has two Java EE application servers, (Glassfish and WebLogic), Oracle and MySQL databases, and Solaris. IBM has AIX, a Java EE application server (WebSphere), and the DB2 database.
RedHat comes close, in that they provide a supported OS and a Java EE application server (JBoss). But they need to buy up a database vendor. Red Hat should have bought MySQL, but in the past they have been hesitant to provide their own database because they wanted to appear database vendor neutral. But now that Oracle owns Sun, that’s not going to fly anymore. Oracle is obviously going to push the Solaris / Oracle combo, and discourage the Linux / Oracle combo.
And then we have SUSE, which is sort of left without any kind of stack at all.
Alas, it is too bad that VMWare didn’t buy Novell. That would have raised some eyebrows and probably caused a little bit of nervousness at Oracle, Microsoft, IBM, and Red Hat. A VMWare purchase of Novell would have been very interesting indeed.
Edited 2011-05-01 01:05 UTC
How can you make money by from making desktop Linux distributions, good Sir?
Red Hat makes money by selling support for their server distro, and if Canonical will ever make a buck, that will be because they followed Red Hat steps and released a server version of their distro.
Have been using openSUSE since 11.3 release. I like the default install and its standard configuration and selection of packages. For the needed extra non-free or patent potentially encumbered stuff there are packman repositories.
KDE 4.x (now 4.6.x) is finally running smoothly. Actually, it is running so well that I see less and less “reasons” to run xfce now. Just some tweaks and voila, a clean and pleasant desktop. My favorites applications on linux are mostly kde4/qt4 now. Here is my list:
– Dolphin (default file manager) is the best of the bread to day-to-day use now (though I fire up mc from my favorite terminal emulator when a job needs it);
– amarok is nice (after tweeks);
– k3b is the best burner of FOSS;
– smplayer is damn good or video;
– digiKam for photo management;
– qtoctave and freemat (the last from svn). Damn good for most of engineering things I do;
– qtiplot is also very useful for plotting stuff;
– luckyBackup (hard lessons we never forget);
– kate, kdevelop, Qt Creator and kdiff3 for most of the programming things. And KDbg is back!
– scribus and texmaker/lyx (cant make my mind between them) for most of more-than-just-text stuff;
– UDAV for mathgl stuff;
– goldendict is the best free dictionary (to me, of course);
– yast2 for system management. Good, not spectacular.
My main missings are:
– CAD (damn, Autocad is just too good! Have been playing with Bricscad lately to see if I may stop dual-booting);
– image editing (have high hopes for krita from calligra fork, using gimp, which is very nice and full of plugins by now);
– vector stuff (also high hopes for karbon. I like inkscape, anyway).
And there are some stuff that I pretty much doubt I will move from, as it is nice to have them on both MS-OS and linux:
– LibreOffice;
– Mozilla Firefox;
– Mozilla Thunderbird.
What I dislike is the akonadi/nepomuk (the last I always disable, the first can not go away for some reason) things. I dont have use and would like to get rid of them or, at least, leave them on a very frozen state on a deep dark cave somewhere on the confines of the universe.
The forums and bug reports are also good and helpful.
I also have been playing with openSUSE:Education-Li-f-e lately and confess I am impressed so far.
I really hope that Attachmate will do the right things, from a FOSS/community perspective, and succeed on linux/desktop field.
Edited 2011-04-30 14:08 UTC
I’d say the most frequent crashout on my system is Dolphin. It’s a very nice file manager but it frequently bombs out when I’ve lots of directories open. The novelty of opening a second directory and selecting some files only to see all my Dolphin instances disappear has long since worn off. If it’s a flaky plugin taking the whole thing down then fair enough; fix the plugin and fix Dolphin so a crashing plugin doesn’t kill the entire thing.
Nepomuk. this is becoming detestable. When ram maxes out, swap use grows and my machine crawls too a halt, there is one consistency. Not the vm’d server with similar uptime to the host (groupware and various monitoring apps; it’s not sitting still). Not the pentarget VMs or related heavy load tasks on the host os. The 72 hours of Ophcrack or John burning away on multiple cores.. anything productive I’m doing?.. nah.. it’s Nepomukservices. Even Firefox on Windows causes less grief leaking memory everywhere and crashing out.
This is Debian 6; stability and security are core project goals. But, I can accept that other distributions may be shipping a more stable Dolphin build. Dolpin may be falling below Nepomuk seems to be a steaming pile everywhere it lands out of the cow though. I’d be happy to see both significantly improved or at least enough so that I’m not shipping crash reports so regularly.
(and with that, I’m off to disable Nepomuk. Even if other apps are using it’s index library.. I’ll take the performance drop over the Nepomuk crap happily. It’s had it’s chance until I see an updated package.)
the good news for you is that this will indeed be solved in the upcoming version of Dolphin. The problem is that the file information plugins can crash Dolphin. While many bugs in those have been fixed, they still aren’t perfect and Dolphin will use them in a separate process to prevent this issue in the future.
Sweet! I’ll be watching for a package update in Debian’s repositories; hopefully it’s not a version update that will get stuck in Deb Testing rather than being included into Deb Stable.
Now, if someone would just fix the Samba “resource unavailable” nonsence. I have one machine that has no issue reading Samba shares off my NAS and another machine build from the exact same package list and config script which consistently returns the “unavailable” error. Dolphin displays it as “password incorrect” when it’s not and smb4k just repeatedly asks for the uname/passwd regardless of what config settings I change. Since they are both relying on mount.cifs, I blame the “unavailable” error.
Samba itself is a necessary evil in it’s own right which I’d drop like a hot potato if Windows wasn’t so NIH objectionable to any other network sharing standard. (why on earth is there still no native SSH in Windows and what I would do for a native sshfs for windows cause any of the implementations tried so far suck).
But that news about Dolphin is happiness. Thank you for that.
Yeah, CAD apps are a weak-link in the Linux chain. We’ve tried various ones (QCad, Cycas, Rhino, and another one I can’t recall the name of) but haven’t found one that works well for those trained on AutoCAD. Until we found ProgeCAD.
It runs via Wine, sure, but it runs exceedingly quick and stable on our Debian 5.0 boxes. It’s AutoCAD compatible, and the interface is very similar to AutoCAD. Our AutoCAD-trained teachers love it.
I remember when Attachmate had only a few employees.
Back in the early 80s the company I worked for used Attachmate to connect to mainframes. I think there was less than 10 people that worked for them at the time.
I wish them good luck (in a non sarcastic way) with Novell and SUSE. I hope that they can do a lot more with Novell’s technology and products that the post Ray Noorde CEOs have done with it.
Ray was a true visionary. Like my other post here, he thought that the best technology would win just because it was best. And it did for awhile. MSs non-technology methods (FUD – suck up to CEOs – the emperor has no clothes – technics worked and the best products lost out. Hopefully Attachmate can find a way to resurrect Novell, assuming that they bought Novell for more than just patents.
I give SuSe about 2-3 years, then it is probably shut down or castrated.
The Board of Attachmate is colluded. In about 6 months, expect the first release of the new SuSe distro under Attachemate leadership to be crap.
They will announce “changes” which include old world thinking such as the closing of the web site to only customers and a large precipitous drop in programmer contributions.
(They will basically lay everyone off to increase revenue.)
Next up for the scrap heap: RedHat.
-Hack
You might be right about SuSE. But I don’t see Red Hat going to the scrap heap anytime soon. They have too many customers, and too many business partners. Red Hat is pretty much the dominant unix OS in the the enterprise today.
I do see Red Hat as an attractive acquisition target though. Possibly either IBM or VMWare might be interested in buying them.
Edited 2011-05-01 16:57 UTC
The rule is simple, kick all the Americans out and leave it to the Germans to recoup the true spirit of SuSe. Novell burned SuSe’s credibility and added layers of wasteful marketing and management.
As much as I hate saying it…I agree with you.
If you know anything about Novell, you’d realize how funny the wasteful marketing remark is. Despite having a marketing division, when was the last time anyone saw any Novell marketing? I remember some fish ad during the superbowl ten years ago, but since then….?
http://www.deseretnews.com/article/705371845/Employees-say-hundreds…
> With Oracle currently doing everything in its
> power to thoroughly destroy what’s left of
> Sun’s open source commitments
I hate this type of bullsh** that comes from people that have no idea whats going on in the most important post-Sun open source communities. OO.org was nether one of the most important ones from Sun perspective !
Glassfish, OpenJDK, Netbeans, MySQL, .etc. … those are the most important projects from Sun. And Oracle is putting more resource in to then than Sun did !
Oracle just stopped investing in projects that don’t interest then. And why shouldn’t they ? Open Source is not about sponsoring geek’s …
Well, I think there is some concern in the community about some of the projects you mention. For example, Glassfish competes with their commercial app server WebLogic. So one has to wonder how much effort they are really going to want to put into Glassfish.
The same is true with MySQL. How much effort are they going to really want to put into MySQL given that it competes with their flagship product, and they can’t make nearly as much money off MySQL as they can off of Oracle DB?
As far as NetBeans, they seem content on making it a Java only IDE (unless the community takes over the Ruby on Rails and Python modules). Oracle doesn’t want to deal with this modules anymore and has discontinued development of them