On Thursday, March 15, 2012 3:39:01 PM UTC-4, Mark S Bilk wrote:
> Desktop/Laptop Linux is a great SUCCESS - How to get started
> On Mar 14, 3:56 pm, Torre Starnes <
torre.star...@gmail.org> wrote:
> > 1. Organization.
Linux development, and OSS development are considered business models that IT departments in many corporations and consulting organizations have been trying to emulate. Linux developers have been able to collaborate more efficiently, coordinate more efficiently, and develop more software in less time, with more flexibility than more traditional "waterfall" approaches used in more traditional Microsoft style "New release every 3-7 years" approaches.
Linux development has been organized so that kernel, libraries, user interfaces, and application development can all be done independent of each other, with each component being enhanced and improved, often dramatically, while maintaining backward compatibility and consistent user interfaces. The practice of "publish early and often" has made it possible for the product to evolve more quickly as well, even if there are changes at the binary interface, the availability of source code for libraries and API "wrappers" has made it much easier to release applications for newer systems, even new processors, without any more effort than a recompile.
> > 2. Realizing early on that fragmentation will kill the project
> > quickly and reacting to bring the community together.
> See below.
> > 3. Providing what the consumer wants to pay for.
The irony is that the Linux/FOSS community has been able to produce things that consumers might pay for, and then "Give it away", rather than demanding huge "up front" royalty payments that can total up to thousands of dollars per user, and then offer support plans for an additional 10-20% per year, thousands of Linux applications are provided in a Linux distribution, with customers being able to try the various applications and tools, and then fund the support for the products they like the most. Many Linux users purchase support contracts from companies like Red Hat, and many corporate customers purchase these support contracts "in bulk" getting very competitive pricing.
> > 4. High quality applications.
The common WinTroll tactic is to compare a free application like GIMP to a $1500 product such as Adobe Photoshop Creative Studio and try to claim that GIMP is poor quality software - because it has to be compared to a $1500 commercial product to appear to be inferior.
GIMP has features that Microsoft ShovelWare wouldn't think of including, and has features that you wouldn't find on the $40-$100 versions similar products.
> > 5. Ease of use.
WinTrolls love to point to the command line interface and claim that this is the "real" Linux interface. They will cite complex shell scripts and claim that these are too difficult to use.
One area where Linux has traditionally been weak is the difficulty of CONFIGURATION. Most WinTrolls who pan Linux don't actually USE Linux, they rant and rave about how hard it is to install Linux and get some chip for which you much download a proprietary driver from Broadcom or ATI or NVidia, knowing that the specific chips in question are covered by nondisclosure agreements with Microsoft - which means the vendors can offer binary-only drivers and support software, but can't provide source code, even if they wanted to.
For years, Linux advocates have suggested that if Linux were pre-installed, or if the WinTroll had been able to find out - by reading the product description, which machines were "Linux Ready" or "Linux Compatible", they could have easily installed Linux and had a working system without all that pain. Microsoft was aware of this, and worked very hard to limit the availability of such information - especially in the retail setting and the product descriptions used in media advertising and vendor web sites.
When companies like ASUS and Acer started shipping NetBooks with Linux preinstalled, they were remarkably easy to use. Furthermore, they were inexpensive, and because they didn't need as much storage, could use solid state disk and run for several hours on a very small and light battery. Furthermore, the netbooks could be plugged into a full sized monitor and keyboard and used like a full sized computer.
Microsoft contacted ASUS and Acer and suggested that it could prove that at least 10% of these "Linux Netbooks" were now running Windows XP, and that unless these companies could PROVE that the customers had purchased appropriate licenses (at $300/machine for FULL versions), Microsoft would expect them to pay for all of the licenses being used by these "pirates". Microsoft offered to "settle" by offering them a price 5% of the price of a full license, on condition that ALL future Intel devices be shipped with OEM licenses, which required that Windows be preinstalled prior to shipment.
Of course, since Windows XP needed more storage and memory than Linux, Acer and ASUS, as well as HP and Dell, had to sell netbooks that had magnetic hard drives, and extra batteries, which not only increased the cost, but meant that the laptops could only run about 2 hours on battery, or that bigger and heavier batteries were needed.
When Microsoft began insisting that Windows 7 be installed on NetBooks, it pretty much killed the market. There was no profit left, the costs were higher than the retail prices, and prices were eroding. OEMs and Retailers fought to make a little profit by offering AntiVirus software, and trying to get customers to pay THEM for the MS-Office licenses, as well as offering support contracts and extended warrenties. A $250 Netbook could end up costing a customer as much as $700 with all the "extras", and many customers just opted not to get the "extras".
Android has flipped that on it's head. With Android, OEMs could use a chip that COULDN'T run Windows, assuring that pirates would not be able to install Windows on it. Furthermore, since the hardware was known (part of the ARM modular fabrication), the drivers could be compiled in, reducing the amount of "probing" - which meant that the Linux system could be fully powered up in as little as 10 seconds, rather than the more typical 3 minutes for a fully functional Windows system.
Since there was no configuration required, Android devices had all the best features of Linux, virtual desktops, true preemptive multitasking, client/server based applications with servers either on-device or via the net, and the reliability and security of Linux. Furthermore, Google offered a secured repository from which users could download both free and purchased applications without fear of viruses and malware.
Eventually, some Trojan horse applications were written and distributed through "direct" channels, but even these could be detected and the user could be alerted that they were taking a risk. In most cases, the applications were limited in the damage they could do, because most users didn't even know HOW to "root" their machines. They didn't need to.
> > 6. Filling a void instead of creating one.
Not sure where that is going. Linux has created many new markets and industries. Linux helped power the early internet servers, POPs, dial-up servers, and still powers many of the servers provided by Internet site providers. Microsoft still maintains a presence in the server market, but has never been able to deliver as much "Bang for the Buck" as Linux or UNIX. The UNIX market has been shrinking in terms of number of servers, but that's because the server market has shifted from hundreds of 2 processor or 4 processor stand-alone servers to servers with as many as 128 cores connected to SAN storage with 32 drive modules buffered by gigabytes of cache memory, and configured into virtual machines which can share resources. A single UNIX server of today can do the work of as many as a thousand of the old 4 processor Solaris servers with directly attached SCSI drives.
Linux has also shifted to "The Cloud" - with virtual machines being strategically distributed across blades that can have as many as a thousand processor cores in a 19 inch rack, connected to storage arrays of hundreds or even thousands of drives, with cache sufficient to read a gigabyte - 7200 times per second.
Linux on the "Desktop" or Tablet, or SmartTV, has created new demand for new applications. Applications like Google Desktop have made it possible to share spreadsheets in real-time, making it possible for many users to make updates as often as needed, and to always know, immediately, what the current status is, and to maintain accurate statistics. IBM, Amazon, and Zoho have similar products which can be used in either public or private clouds.
> > 7. Quality.
Quality has always been a cornerstone of Linux. When Linus first published his 10,000 line kernel, and said, "take a look and tell me what you think", he got a LOT of feedback, and very often, comparing the tiny little kernel which ran on an 80386/SX processor with as little as 1 megabyte of memory, with Sun SPARCStations and SPARCServers. Linux had to compete with BSD, AT&T System V, and OSF/1. Tannenbaum and Linus held a very public discussion of the Mach Microkernel vs the Linux kernel.
What Linux had that the others didn't was the ability to debug and test a kernel using an inexpensive PC, and running the kernel to be debugged an a stable kernel as an application. Linux had made extensive use of the Memory Management Unit, which made it much easier to switch between the host kernel and the target kernel without needing $millions worth of complex test and diagnostic equipment. In addition, when the kernel did crash, it was possible to capture everything in a core dump, and debug the crash, tracing back to the root cause of the problem - all using cheap hardware and free software.
Also important was that Linus and the Linux kernel team realized that they needed to provide API compatibility with ALL variants of UNIX even if their internal implementations of those APIs were radically different. Because of the way Linux used the MMU, operations that used to take milliseconds could be done in nanoseconds. By managing how libraries were mapped into the MMU, and using position independent code, more of the libraries could be shared from process to process, meaning less memory was needed and/or much more could be done with the amount of memory available.
Linux also realized that the biggest bottleneck to performance wasn't the CPU or the Memory, but the hard drive. As a result, he encouraged the development of more efficient storage file systems, ext2, ext3, Reiser, ext4, all designed to take advantage of superior memory mapping and the ability of Linux to use the "free pool" of memory dynamically, using it for disk cache, reading whole i-nodes a cylinder at a time rather than a 4k "cluster" at a time, managing "fragments" of 1 megabyte or more at a time, and reducing the fragmentation as storage was freed up.
This made Linux a very powerful server platform, and often competed favorably with smaller UNIX systems.
Linux also had to compete with the UNIX Security model, but contributors had already seen a gap between the Unix security model and the demands of a secure Internet server. Linux adopted "Pluggable Authentication Modules", allowing a range of security options that could efficiently scale from a few users, such as a desktop environment, to direct integration with corporate LDAP servers which could be configured to securely manage thousands, even hundreds of thousands, of individually identifiable users. As a result, new server applications could simply use the Linux security model to extend their own security, rather than having to create custom security models. By using LDAP servers, users could be organized into groups that could be given more restricted access, and new customers could self-register and be given controlled access - rather than just having to make everything public to everybody.
Of course, that efficiency, stability, and security meant that Linux systems could be scaled down to smaller systems as well. Linux was easily ported to PPC, MIPS, and ARM chips, making it possible to implement "system on a chip" technologies. Modern "Linux on a Chip" technologies can cost as little as $2 per device in OEM quantities, making Linux on a Chip a viable alternative to discreet electronics. We are now seeing Linux on a Chip even in SDHC formats, allowing your camera to send pictures to your tablet or cell phone via WiFi.
Linux "Appliances" based on the "Linux on a Chip" technology began to proliferate into many new devices. GPS devices, MP3 players, digital picture frames, WiFi hubs, Routers, IP-Phones, DVRs, Cable Modems, Cable Boxes, Digital converters for conventional TVs, Smart TVs, even several of the computers in cars, now run Linux. Even places where BSD is preferred, libraries, tools, services, drivers, and applications originally developed for Linux are an integral part of more and more UNIX systems including BSD, AIX, Solaris, and HP_UX systems. Other vendors, such as Silicon Graphics, have switched to the Linux core kernel and libraries, and added their own User Interfaces on top of these.
Mac OS/X is BSD based, but uses many of the same code elements used on modern Linux systems. In many cases, the Linux code has been enhanced and security hardened, but since the Linux enhancements are often only released under GPL or LGPL licenses, Apple has to negotiate to get these upgrades and fixes into their own product, which can often get a bit expensive.
Ironically, Linux often exists on Windows systems as well. Many applications now make use of the Cygwin libraries, which allows programmers to use a standard API that can then permit them to use the same source code on Windows, Linux, and Mac, as well as Android, iPhone, and iPad devices.
Even many Java based applications are now developed and tested on Linux first, because Linux distributors offer VMs that are free of Native Mode code, assuring that any code written to run on the Linux JVM will run on any other Java JVM, including Linux, Android, Windows, Mac, iPhone, iPad, BlackBerry, or WebOS.
> > 8. Financial backing.
In 1996, when NT 4.0 was first released, InfoWorld gave Red Hat Linux version 4.0 the nod as product of the year. Suggested that it was a "Tie" between Red Hat and NT 4.0, probably to keep from losing Microsoft's economic support. Later that year, Red Hat released 4.1 and 4.2, which passed NT 4.0 very quickly in terms of performance, security, stability, and even ease of configuration.
In 1997, Linus Torvalds suggested, at the prodding of Caldera CEO Ransom Love, that "Linux take the Desktop". Linus and the kernel team, as well as the library teams, worked on ways to improve support for desktop and laptop systems, with better shutdown, recovery, sleep, and hibernate capabilities. Linux had had many of these features even in the early 1990s, but they not been emphasized because of the focus on Linux as a server.
Meanwhile, Linux as a server had shifted from "that little box hiding in the corner, with not screen attached", to a very strategic part of the corporate IT infrastructure. In January of 1997, most CIOs didn't even know they had Linux running in their data centers, about 17% said they had Linux servers. One year later, the number was more like 90%, and many of those who had tried to get rid of Linux discovered that their e-mail, intranet, web servers, and security - stopped working, or became very buggy when they tried to run the same functions on Windows.
By the end of 1998, even IBM was realizing that they could not compete head-to-head with Linux, and decided to "join them". IBM invested $1 billion in Linux support, promotion, and support. IBM had tried to get 64 bit UNIX from SCO, but even after $millions in seed capital and giving them thousands of lines of kernel code from their UNIX and Mainframe systems, SCO couldn't get a working version. Meanwhile, Linux was running on 64 bit chips, including SPARC, PPC, MIPS, Alpha, and the prototype Itanium chips. Linux was delivering the goods, and doing it brilliantly well.
The final push for IBM's adoption of Linux was when some IBM employees in the research lab had created a mainframe emulation package called Hercules, and some Linux hackers, for grins and giggles, ported Linux to run on Hercules. IBM found that the code also worked on REAL IBM Mainframes, and in 1999, began tweaking VM into ZVM, which was able to very efficiently run Linux images, without some of the overhead traditionally needed to support MVS images and OS/390 UNIX Images. By the end of 1999, IBM was showing a Mainframe that could run over 10,000 Linux servers, all handling the load of a 2 processor PC.
IBM had been trying to figure out how they were going to phase out the mainframe, especially since there where still thousands of corporate customers who were still using MVS, CICS, IMS, and other "legacy" applications. Linux had not only made it much easier to very efficiently interface to these applications, it had also created a whole new market for the mainframe.
That investment has come back at least a hundredfold, and IBM now considers Linux a strategic part of it's product offerings and product line.
HP, Sun, and Dell have also come around to embracing Linux, and Oracle, prior to purchasing Sun, had also not only adopted Linux, but had made Oracle Linux a strategic part of it's own offerings.
Linux applications and services were also designed to use smaller components that could be combined on a single PC, or distributed across a broad array of servers, into clusters. The NASA Beowulf Cluster, originally based on 5 servers, had begin to grow into hundreds or even thousands of servers. Distributed databases and distributed applications, servers that could search other servers, extended the ability of Linux to handle very large and complex environments. Google's search engine has expanded to thousands of servers all over the world, and can provide the most relevant results in less than a second.
Using Web Services, Dynamic Discovery, message passing interfaces, and other distributed computing, JMS Publish/Subscribe, and so on, Linux has made it possible to achieve unimaginable performance and remarkably low cost.
> > 9. Accountability.
> > 10. Pride in the product.
>
> Desktop Linux satisfies all of these criteria. Tens of millions
> of people around the world are using it. It provides security,
> power, multiple virtual desktops, flexible file systems,
> modularity, processor independence, open source code, and zero
> cost. Most of these advantages are not available in Microsoft
> Windows or Apple Macintosh.
>
> The only reason why Linux is not the most popular operating system
> is that Microsoft has used many different coercive and deceptive
> operations for 15 years to try to destroy it:
>
>
http://cosmicpenguin.com/linux/MICROSOFTS_WAR_AGAINST_LINUX.html
Microsoft as a long and well documented history of using every means necessary to keep ALL competitors off the desktop. They have successfully managed to beat Apple, Mac, DR-DOS, GEM, several variants of UNIX, Sun, HP, Dec, IBM, OS/2, Warp, BeOS, and hundreds more that never even made the papers. Byte magazine used to cover the rise and fall of hundreds of would-be competitiors to Microsoft.
Microsoft has often gone to the very edges of the law to defend it's control of the desktop market, especially the OEM distribution channel. Starting in it's second year, in 1977, Microsoft realized that they could not count on retail consumers purchasing their product, especially when superior products were available at lower cost from more aggressive competitors. Bill Gates quickly realized that the only way he could be assured of a reliable source of revenue, was to insist that the OEM pay for the licenses, in advance, based on a substantial but predictable increase over previous year's sales.
But getting the OEM to pay wouldn't be easy. It would require a bit of extortion, threatening to sell the strategic product to a competitor, along with enhancements that the competitor offered, that wouldn't be available in your product. To get Altair to pay $150,000 ($50 per machine for 3,000 machines - expected sales), Bill threatened that if they did NOT pay, he would port Micro-Soft BASIC to the SWTP machine - which supported a bootstrap ROM. Altair paid, but insisted that Gates agree not to port to the 6800 chip used by the SWTP machine. - Extortion. The same tactic still goes on today, with Microsoft threatening to put one OEM out of business by offering technology to competitors and excluding the reluctant or uncooperative OEM. Ultimately, they have to surrender. Without the OEM license, the OEM can't sell the line, but the OEM license mandates that the software be installed on EVERY machine.
Of course, as we know, Gates DID port to another machine, the Commodore PET, which had to run a 6502, similar to what Apple used. This allowed Gates to get around the agreement not to port to the 6800, but at the same time, gave him the ability to put his own customer, Altair, out of business, in exchange for an OEM license agreement with Commodore. It was fraud, Altair had assumed that Microsoft (Gates) could be trusted, and soon found that Gates had outmaneuvered them in the contract negotiations. Would they have agreed if they had know that Gates was planning on bankrupting them - even as they were signing the contract?
Microsoft pulled similar maneuvers when they stiffed Commodore to get Radio Shack, then stiffed Radio Shack to get IBM, and several years later, stiffed IBM to produce Windows under their own name.
And the extortion didn't stop either. Dell was threatened with bankruptcy when he began offering PCs powered by SCO Unix. AST, ALR, and NEC were busted when they attempted to offer both OS/2 and Windows machines.
And when the Mac came out, Microsoft began making vaporware announcements of Windows. It would take nearly 8 years for Microsoft to deliver a marginally functional version of Windows, Windows 3.1, and even that was inferior to Apple's Mac offering in terms of stability.
When Sun threatened to capture Desktop market share - capturing almost 15% of the corporate desktop market as "Workstations" using both Sun consoles and X11 terminals, Microsoft again used VaporWare - announcing Windows "New Technology" which would be "A Better UNIX than UNIX". Microsoft wouldn't actually deliver the goods until about 9 years later, with Windows 2000, ironically one of Microsoft's most successful product launches- in terms of units deployed, but a failure financially - because Microsoft had to give it away to keep corporate customers from switching from NT 4.0 to Linux on desktops.
In 1994, Microsoft was more worried about OS/2 Warp than Linux, but Microsoft was also acutely aware of Linux, and concerned that a PC doing the same things that had previously only been possible with a $35,000 workstation, could be a big threat to Microsoft. Microsoft could compete against Sun, but a freely distributed operating system that could give PC owners the capabilities of a SUN, using a PC that couldn't even run the latest versions of Windows anymore - that was a big problem.
Microsoft used the Windows NT Hype for most of 1993 and early 1994, but when NT was released, and NT 3.1 and NT 3.5 sales never materialized, Gates scrambled to announce "Chicago" as "NT Lite". He announced that it would be out "Real Soon Now" and then announced it as "Windows 95", indicating that it would be out in early 1995. Linux was getting better and better, the Internet had gone from a research network to a key strategic offering, and more and more people were finding out about Linux through the Internet.
When Microsoft finally release Windows 95, it wouldn't run on most existing PCs, those who tried to upgrade found it to be a waste of money. Far more memory was needed (from 4 meg to 16 meg), more drive was needed (30 megabyte to 90 megabytes minimum), and plug-and-play only worked with the PCI bus.
This meant that a LOT of companies had a LOT of PCs in the back alleys of Manhattan, in the back of the buildings, or being shipped off to "recyclers", who couldn't GIVE them away as Windows machines. But 12-16 year old kids were more than happy to pick up a "free computer" and install a "free Operating System" - Linux, to come up with a PC that would not only work as a web browser, but could also work like a web server.
By the time Windows 95B was in full swing, there were so many surplus computers that Dell, IBM, HP, and many others were offering "Lease Return service" and shipping the PCs off to India, Africa, and South America, where Red Hat had trained high school and college students in how to install Linux on the machines. By the end of 1997, an NGO was shipping 10 million PCs per year, which were converted to Linux.
As was shown in the Combs vs Microsoft exhibits, as well as exhibits in the DOJ vs Microsoft case, Microsoft responded with a "Win at all cost" strategy. In some cases, they offered free Windows, and even memory upgrades - in hopes of keeping customers on Windows. But in many of these countries, where the Linux PCs were being given away, Microsoft didn't see the threat, or didn't care, because they didn't see cash money in it.
OF course, a few years later, these kids who had been using Linux all these years were looking to upgrade. They had skills, they had knowledge, and they were willing to pay for used computers, or pay for desktop machines using motherboards made by ASUS and Acer. And if it didn't run Linux easily, they sent it back. Linux hostile machines would rot on the shelves, while Linux ready machines were selling as fast as they could be obtained.
In many of these countries, these smart kids had moved into government and business, and were recommending Linux rather than Windows. They were recommending Linux servers and Linux desktops.
Microsoft's tactics are well documented, in court cases, in print media, and in court transcripts. Microsoft has faced trial and sanctions in the United States, Europe, South Korea, Brazil, Africa, Asia, and Eastern Europe.
What has made Linux the remarkable competitor is that even with Microsoft using around $10 billion a year in legal fees, marketing, promotions, licensing restrictions, advertising control, and so on, Linux has not only survived, but has thrived, opening up new markets, new devices, new technologies, and new industries.
Microsoft has spent 30 years bleeding corporations, OEMs, and software vendors dry. They soak up the lion's share of the profits, and retaliate brutally when challenged.
Meanwhile, Linux has created companies like Google, Amazon, Facebook, and so on. Linux has given us the Internet, Digital Television, DVRs, and given us high speed connections to internet and digital media, and taken a token payment, usually a tenth of a percent, in exchange for service and support from non-profit organizations which are immune from corporate take-overs and incapable of extortion tactics such as those used by Microsoft.
Linux doesn't rely on vapor-ware, in fact, often, new products are released with little or no marketing and are discovered and popularized by the user base, who often promote it via blogs, e-mail, newsgroups, FaceBook, and word of mouth.
Give the extraordinary amount of resources Microsoft has deployed to "Kill Linux", and the remarkable success of Linux in so many markets, Linux must be doing something right.
The only question is when the Laptop makers decide that they are tired of paying Microsoft's extortion money, tired of being locked into a Single Platform solution, tired of selling products at a loss, hoping for a few "table scraps" in the form of marketing gimmicks.
These days, a laptop sells for about $429 at Best Buy. An Android tablet sells for about $329. The Android does more, has thousands of applications available for free or very low cost, most under $15, many under $5, hundreds of nice tools and applications for under $2.
With the Windows Laptop, I have to buy AntiVirus - $70. MS-Office - $150 for personal edition, but since I'm a consultant - more like $400. Extended Warranty - voided if I install Linux - $300, replacement plan - voided unless returned with Windows - $300. Visio - $300, MS-Project $500, Adobe Acrobat $100, Photoshop CS 5 - $1500. Visual Studio $300....
It wouldn't be so bad if Microsoft just "nickled and dimed" me to death, but with Windows, it's "Buy a TV, or MS-Office", "fix my car, or MS-Project", New refrigerator, or Visio...
How much longer will consumers stay with Windows on Laptops?