M$ builds gliders with the same skill, hype, and complexity with which they build operating systems. Here is the result. My conclusion? It is wrong to assume one of the wealthiest corporations in the world can do anything right. They seem to always release beta-software. Have you noticed? Diversity is good.
Archive for August, 2010
For years AMD denied the netbook was real, continuing to crank out high-end chips and imitations of high-end chips. Now they are getting serious about low-end power consumption and better performance. It’s about time. Intel has taken a huge lead with the “Atom” line. In 2011, AMD’s new chips promise to compete well with Intel’s for less than 1 watt. Another sign of a sea-change at AMD is that the Bulldozer and Bobcat chips will be more modular, an approach ARM has used for years. This should mean AMD will have the capability of re-using what works and patching or replacing what does not with much less time and energy wasted. That’s the right way to do software and hardware.
Other moves by OEMs to push the borders of what is a netbook should mean the netbook market will be revitalized. AMD’s chips set to arrive in 2011 could make 2011 a great year for netbooks. I still think ARM will continue to expand its role because the smaller instructions and instruction-set mean less bandwidth to and within the CPU. The same thing applies to internal storage. You need less if your code is more dense. There is nothing sacred about x86 instructions and ARM does not carry that baggage.
Of course M$ is a huge corporation with turnover of staff, marketing campaigns and aggressive salesmen so it changes as time goes on and in response to changes in the economy but, for a long time M$ has been intolerant of competition. They were never content to coexist but needed to eliminate competition going to great ends to do so. Their huge revenue streams pays for lots of strategies to eliminate competition far beyond what is needed or possible in companies competing properly in the market. In particular they have gone to great lengths to prevent GNU/Linux competing on the desktop through FUD, funding SCOG, patent-threats, etc.
Too much water has passed under that bridge for me to believe there is much possibility for M$ to change in its relationship with GNU/Linux. They may tolerate GNU/Linux because almost all their customers use GNU/Linux but they still would not let GNU/Linux enjoy an opportunity on netbooks, notebooks or desktops. They called GNU/Linux a cancer, derided the great programmers who created the Free Software, threatened patent-suits and pressured many to pay licence fees for GNU/Linux and paid OEMs not to install GNU/Linux. Then retailers had no stock and consumers could not choose GNU/Linux. They pressure those few who do sell both OS to eliminate the possibility of comparing the price of identical units with the two OS. They fouled the XML ISO process with pressure-tactics instead of technical arguments. They short-tracked a 6000-page specification. I used to work on a 90-page specification and we did not bother to short-track. What is the point of forcing a broken spec on the world? The efforts to eliminate competition never cease but they constantly change form.
Some believe M$ now “loves” Free Software. It isn’t happening. Those salesmen at M$ will say anything to improve their market position and undermine the competition. That isn’t going to change.
There is a decent article at ITwire about migrating a small accounting office to GNU/Linux. The guy took nine months to do a slow/thorough job one person at a time. That makes sense from the business point of view (minimal/gradual disruption) but is very inefficient use of IT skills. Small does not get economy of scale very well.
My present school was migrated fairly gradually and was larger but took only a few weeks not months so my pace would be seen as much more rapid. The differences are that my staff are mostly younger and their IT needs are not specialized. The usual GNU/Linux desktop does the job fairly well and we may add useful apps but none of the additions are “mission-critical”. TFA describes accommodating several specialized apps. The guy did use thin clients which are the best use of IT IMHO. Accounting in particular does not need a lot of power in the CPU. If there is speed required it would be in managing data which is mostly limited by algorithms and storage.
As usual one of the commenters says “7″ is faster than XP. In what universe? Here, “7″ is slower than XP even when running on five years younger hardware.
Like Munich TFA describes a very gradual approach which requires much more work than is necessary. Lots of migrations are done very rapidly, say over a weekend, for projects that size. Extremadura did 80K PCs over a weekend. Things did not break because they had very little before and whatever they got was far superior. That’s an “easy” migration/leap. Having to treat every PC as unique in a system is much more likely in a small system because there is not a lot of redundancy whereas in a larger organization there may be groups of 50 or more users who can be migrated together.
A good article on the subject neglects a few obvious facts. The downward curve on many search terms at Google does not signify much because the number of search terms keeps growing as the web grows. New people are joining the web all the time through aging and affordability. The number of websites keeps growing. Naturally searches that target particular websites decrease in proportion. This indicates maturity of the subject searches not a decline in interest level. Further, although Ubuntu is the top curve, there are many distros whose total effect is similar. So, Ubuntu is the most popular but the popularity of GNU/Linux is much larger than the popularity of Ubuntu. Ubuntu naturally is at the top because it is the only GNU/Linux distro properly advertized. Ubuntu/Canonical does not wait for hits. They have deals with Dell and others to distribute the OS. They have free CDs. They have a thriving community which also pushes the distro. Other distros have fans but Ubuntu has salesmen and they target business and consumers. Suse and RedHat mostly target businesses. Choose your distro and run with it. If you need more help, there is a website for that. If you still have questions, try LinuxQuestions.org.
The Zegeniestudios.net site recommends Mint, a couple of Ubuntus and Debian for me, so I am OK… 
OpenOffice.org is one of the flagships of desktop installations of GNU/Linux. It is one of the most active projects in FLOSS. At 450000 it almost certainly has the largest community and possibly installed base (GNU/Linux or FireFox may be ahead in numbers…).
The featureset and ease of use of OpenOffice.org are mature and first-rate. With such a large community of supporters and installations, it should be around for a long time. The question arises about how acquisition by Oracle might affect it in the future. While Oracle has a lot of influence, the project is open and a larger base of contributors do operate. A recent article questions the intentions of Oracle for this product because on that other OS dependencies on non-free software exist (Visual Basic). The same doubts could be raised about several other projects Oracle acquired from Sun: MySQL and Java being the most prominent. VirtualBox is important and has no clear alternative as a GUI. The recent suit against Google for patent violations is alarming. FLOSS needs to be above that fray to stay healthy. If the owner of a project is willing to sue other members of the community using software fairly (an assumption) FLOSS will fizzle. Such actions essentially renounce Free Software status.
On the other hand, these major projects including OpenOffice.org have no clear alternative except forks and the patent-sword can still be wielded against the forks or alternatives. Postgresql and MUMPS ( non-SQL/RDBMS) could be used but a patent-troll can still attack on the basis of functionality. With OpenOffice.org that is true as well but most of the technology of office-suites is long standing and not subject to patents. The exception could be Java. The database component of OpenOffice.org depends on Java. It could be replaced by a dependence on Python or other scripting/interpreted language but that would be a major disruption.
I do use alternatives of OpenOffice.org from time to time. KDE Office and some components of GNOME work (GNOME also depends on OpenOffice.org). I use LyX for some writing projects. KWord includes the database-merge capability so useful for writing students’ reports. Everything else in OpenOffice.org, I have good substitutes like GNUmeric, phpMyAdmin, Scribus, Inkscap, Dia, etc. but they are not so well integrated. Could we survive a catastrophe with OpenOffice.org? Yes, but it would be a major disruption. In order to minimize disruption it is important to explore options long before a crisis emerges. I have frequent opportunities in my teaching but others will have to make a determined effort to explore GNU/Linux for functionality outside of OpenOffice.org.
Oracle still sells StarOffice, officially so they may continue that and leave OpenOffice.org alone one way or another. We shall see in good time.
How does a monopoly sustain itself? It can produce better products than anyone else or it can cheat, messing with competition and providing “inducements” to stick with the game-plan.
Obviously, without producing better products, M$ would not be able to maintain exclusivity on retailers’ shelves and OEMs product-lists unless they provided inducements. As we recently saw in SEC v Dell, it is not OK to keep those inducements secret if they are a substantial/material fact that could affect investors’ decisions. We know retailers and OEMs margins are tight so the inducements are material. M$ certainly has not produced better products than any other software house on the planet.
What has recently changed is that the SEC is paying huge rewards for whistle-blowers. How many thousands of people in the food-chain are in the know about how M$ keeps the monopoly going? How many have been laid-off or have become disenchanted with the magic kingdom? This is fair turnabout. M$ sends the BSA after its customers. It is fair that “partners” should spill the beans.
I look forward to the revelation that the emperor has no clothes in the coming year. While M$, itself may not have a technical violation all kinds of partners may have failed to disclose sufficiently to describe the risks to investors of dependency on M$.
Dell did OK last quarter but it took in only $2.9 billion for consumer PCs. Imagine if those PCs had shipped with GNU/Linux and they had been able to pocket another $50-$100 per PC. That would have been another $100-$200 million revenue. Compare that to a $21 million loss.
I can do the maths, Dell. Can you?
That other OS is holding you back from profiting from your labours. You are wasting time working for M$ instead of yourself. WAKE UP!
Also, you could sell many more PCs to consumers if you offered what they want, small, cheap computers and GNU/Linux so they are free of malware. How about fixing your website?
OMG! My Job is Threatened
Published by August 20th, 2010 in Linux in Education and technology. 2 CommentsI thought I had job security because I was the only one in the building who knows what
cd scripts;./all somecommand
does to every PC in the building. Nope. I was wrong. A system like mine can be automated so that I don’t need to be here. If a geek is needed a backdoor can be left open. I could be obsolete once the system is set up… Fortunately, I have a day job which pays the bills.
I was horrified to read, “The days of DIY system administration are rapidly coming to a close.” All those lovely GNU tools about to be replaced by automatons. Sigh. Change is a given in IT. Fortunately my system is small enough my home-made configuration works well and it will take some effort to implement puppet or one of the other automatic systems.
Then it occurred to me that such automation does not replace me but allows me to scale. I can manage 1000 PCs as well as 100. Have to push 1 PC per student plus additional labs… What am I thinking? I can automate the whole thing and retire.
Usually that other OS has some bug that multiplied by the number of PCs running it gives a pretty scary number. This week it seems M$ has out-done itself. Almost every application on that other OS is a vector for a newly disclosed vulnerability in how executables load in that other OS.
“We calculated that there are about 100 billion instances of this class currently exposing users,” he said, explaining Acros came up with that number by assessing the market share of individual applications that contain the bug, then multiplying it by the global installed base for Windows.”see TFA
We should abandon IT if this is the level of insecurity we must experience or change OS immediately. GNU/Linux was designed from day one with security in mind, copying decades of sound practices in UNIX operating systems. Use GNU/Linux. You will be safer.
The design of that other OS has been a Kludge from day one, tacking on layers of features on top of a copy of CP/M which was a poor man’s imitation of an operating system compared to UNIX. You can use GNU/Linux which has all the standard features of UNIX operating systems and none of the baggage of that other OS, maintaining reverse-compatibility of vulnerabilities. All the features added to that other OS in the last decade merely make this current issue easier to exploit. Conveniences for users become conveniences for intruders. Within days exploits of this vulnerability will ramp up and the sluggish response of the Wintel ecosystem to work around the vulnerability instead of fixing the root problems will give the bad guys billions of points of failure in the world’s IT. We cannot afford to continue. If what the world has been doing does not work, it has to change.
Really. What are you going to do about this problem? Worry until M$ issues a patch? Worry until every ISP updates every app on your systems? Worry until you can update all your systems and deal with backups? Assume everything is compromised and re-build images and deploy? You cannot fix M$’s problems. They won’t let you. You have to choose a different plaform. That other OS is not working for you. Choose almost any distro of GNU/Linux or MacOS or FreeBSD and your life will be better.
The state of Free Software is good and getting better every day. I like to describe status in numbers but others are more qualitative. Take Eben Moglen, for instance. He gave an inspiring presentation on this topic at LibrePlanet 2010. You can download the video here (60 MB, ogv, 48 min.). Some key points I found surprising although I knew the facts. I had not connected the dots:
- Free Software has come from being ignored and ridiculed to being required by everyone. The world of IT now depends on Free Software.
- Free Software quickly solved the problem of “write once – run everywhere” when no one else could.
- Free Software is being defended against software-patents by large businesses that have tons of software-patents. Patent-trolls hurt all software businesses.
- Free Software is the best software because it has no commercial interests yet businesses love it because it is so good and costs so little.
- Free Software database, MySQL, was a key component of the purchase of SUN so that Oracle could defeat the scalability of M$’s SQLserver.
- Free Software will be even stronger in the future because the low cost of entry and the four freedoms (run, examine, modify and distribute) mean almost every youngster interested in IT will know about it and take that knowledge into adulthood.
According to Gartner, virtualization is in the early stage of mainstream acceptance. Thin client desktops have been around a while but still are used by a small fraction of clients. The movement to various kinds of virtualization will carry thin clients along. Thin clients can be part of the hosted virtual desktop service but the cheapest way to go is with thin clients connected to a UNIX OS such as GNU/Linux. X has been around a long time, is mature, and is being actively developed. There are other protocols which can be used with that other OS as well but X is pretty good for what I do.
If you read TFA you see repeatedly that the acceptance of such and such virtualization was delayed by M$’s baroque licensing. Even now they try to squeeze every cent out of every client. Life is so much simpler with GNU/Linux. Connect it and forget about it. It’s that easy. With Debian GNU/Linux, I can make any PC a thin client of a newer more powerful machine by doing a minimial installation on the client (just X, not a whole GUI) and put “X -query newpcip” in /etc/rc.local and reboot. On the newpc, I can use GNOME, the default desktop, and run gdmsetup as root and enable tcp connections, remote login and perhaps add “Welcome to NewPC!” to the login screen and reboot or restart gdm. It’s that easy. No purchase orders. No choosing a vendor. No fiddling around. I do have to create accounts if I am adding users…
10349
8015
107
2
0
14332 
7212 
7166 
2724 
2035 
1232 
198 
7 
0 
0 
0 
0 