标签: open

So what's the problem with “free” as in “free beer”? I'm a sucker for “buy one, get one free” offers. There are a variety of costs, sometimes hidden, in free (as in beer) software.   Free (as in beer) software makes it difficult to recover the cost of developing software. (The Free Software Foundation (FSF) has a confusing article about how to sell free software  on their website, which skirts the question of how to be paid for developing software, but does include a pitch for donating money to the FSF.) There are ways to make money developing software, but the most effective of these is to either sell support services (like RedHat), build the software into a product (like Cisco), or to use the software internally to provide a service (like Google).   If you happen to have a great idea for a program, something new and novel, FOSS makes it difficult to create a company around your idea. Venture capitalists will not fund development that can be easily undercut simply by copying sources off a website. One can count the major software companies on a few fingers: Microsoft, Oracle, Adobe, Intuit, and a few more, all founded before the rise of FOSS. Of course, if you happen to develop a proprietary program that runs on a smart phone, no matter how pointless  it may seem, the VCs come running.   People confuse cost with value, and this extends to the areas of accounting and taxes. In 2008, the Linux Foundation estimated the cost to develop the Linux kernel at $1.4B, with the cost of a complete distribution about $10.8B. But for the users, the cost is zero. The value of the Open Source software that they use on a company's balance sheet is the cost of acquisition, namely zero. Where development of a new proprietary software product may qualify for an RD tax credit, developers of Open Source software that will not be sold or leased may have a more difficult time justifying the credit.   Few people invest time, money, or energy on things that have little or no cost. You're likely to skimp on maintenance on the $500 clunker you bought to haul junk while sparing no expense on the new SUV you bought for $50K. There is also the Tragedy of the Commons, where a shared resource used by all tends to be over-used to the detriment of all. We don't have to worry about overgrazing the Town Commons, but there is a parallel with Open Source Software, where programs used by many are often supported by very few. The Heartbleed security flaw brought to public attention that a widely used and critical software library, OpenSSL, was maintained by one full-time employee and ten volunteers on a shoe-string budget. Despite the fact that this library was used by many major companies and was an integral part of their products and services, it was Open Source. Where similar software from a proprietary software company would have generated a revenue stream which would fund support and development, the OpenSSL Project depended on donations. There are a number of similar Open Source Software projects, such as the Network Time Protocol (NTP), which are critical to the Internet, but which have minimal support. (The Linux Foundation, in response to the lack of support for OpenSSL and other projects, created the Core Infrastructure Initiative, CII, funded by a number of major tech companies.  CII will fund two full-time core OpenSSL developers)   We think of donations as charity, not as paying for development and maintenance of critical software packages. Indeed, a corporation might have to go through its “corporate giving” department to contribute to one of the FOSS projects, where the application for a donation might be weighed against donations to the local theater company or food bank. Voluntary donations have allowed a number of projects to stave off starvation, but that doesn't result in them thriving. Many companies that use free software almost never think about the cost of development, the value of the software to their products and internal development, or the risk that they are exposed to if the software is not maintained.   Finally, the last hidden cost I see in Free and Open Source Software is minimizing the value of time, energy, and creativity expended by the people who develop software. Few people would say that everyone who writes a book, composes a song, or paints a picture should give their creative work away for free. But many in the FOSS community have a different standard for software developers. People and companies who sell software rather than give it away are characterized as if they are stealing something that belongs to everyone — an opinion widely expressed by the Free Software Foundation and Software Freedom Conservancy.   Science fiction writer Robert Heinlein popularized the saying “There ain't no such thing as a free lunch” in his 1966 novel, The Moon is a Harsh Mistress . While Free and Open Source Software has many positive aspects, including high quality, transparency, and creativity, let's not forget that there are hidden costs. These include a far smaller software development economy (except for games) compared with our vibrant hardware development economy, higher risks from insufficiently maintained software, and undervaluing software development talent (again, except in the area of games). These costs may not be easy to quantify, but they should not be forgotten.   Michael Eager is principal consultant at Eager Consulting in Palo Alto, Calif. He has over four decades experience developing compilers, debuggers, and simulators for a wide range of processor architectures used in embedded systems. His current and former clients include major semiconductor companies and systems developers. Michael has been a member of the ISO C++ Standard Committee and ABI Committees for several processor architectures. He is chair of the Debugging Standards Committee for DWARF, a widely used debug data format. He is active in the open-source and Linux communities.  

We often hear the word “freedom” in connection with Free Software and Open Source Software (FOSS). Actually, most use of the word comes from the   Free Software Foundation (FSF) and its founder, Richard M. Stallman, repeating the term over and over , until the word is almost devoid of meaning . The four freedoms expressed by Stallman are the freedom to run a program as you wish, freedom to study or modify a program, freedom to redistribute a program, and freedom to distribute modified copies of a program.    Now, there might be reasons to debate whether Stallman's use of “freedom” has the same meaning as in “freedom of speech” or “freedom of religion” but that's a topic for a different article. When asked about “free software”, Stallman and others will throw out the quip “free as in speech, not free as in beer”. Apparently, someone, somewhere is giving away free beer. I want to find that person.   Anyone can download the source for the many thousands of FOSS programs, modify it, and redistribute a new version. But in practice, only a very small number of people ever download and build a FOSS program from source. I do on occasion, when a program isn't available in my distribution's repository. Of the few people who download the source of a program, a much smaller number will actually study it to see how it works and an even smaller number will make changes.    I use FOSS software. I use GNU tools, LibreOffice, Firefox, GIMP, and more. Like the great majority of Linux users, I use whichever version of a package is included in my Linux distribution. I'm very unlikely to ever modify any of these programs, even if I have the “freedom” to do so. The reasons are simple: Large programs like these are complex and take considerable time to comprehend how they work.   I recently had an online exchange with someone who was upset that his Linux distribution did not provide support for a peripheral he just bought. He said that he had to purchase a driver to get it to work and that this was the first time in many years that he had paid for any kind of software. He wasn't concerned about the freedom to modify the software, or the freedom to understand how it works, or freedom to redistribute it, both because he lacked the technical background and, more important, lacked the desire to do any of these. The price for the driver was very modest, but he thought that the Linux community should have software support for any hardware on the market available without cost.   In practice, the operative word in FOSS is not “freedom” but “free” as in without charge.   I use Adobe Acroread on Linux, which is free software. That is, free as in free ware, no charge but no source, and with a click-to-accept license. I use Chrome which is built out of a large number of FOSS packages, but which has a license agreement no less restrictive than that included with any proprietary program, prohibiting reverse engineering, creating a derivative work, or copying Chrome. (The rationale, as I understand it, is that Chrome is inextricably related to the proprietary services which Google provides, such as collecting your browsing habits and selling it to advertisers.) Chrome is now the most popular browser, with over 50% of the market, beating both FOSS browser Firefox and proprietary Internet Explorer. Like everyone else, I have dozens of apps installed on my Android smartphone. Almost all are free. That is, there is no charge to install them. Some apps are full function, others crippled or adware to encourage me to purchase a fully functional version. Almost all are closed source, although no one seems concerned.   While there is a small, impassioned, and very vocal community which is concerned about Free Software, the great majority are mostly interested in Free.   Free as in free beer, not free as in freedom.   Notes: The word “freedom” appears 43 times in Stallman's 3-page article “Why Open Source misses the point of Free Software,” Communications of the ACM, June, 2009, vol. 52(6). pp. 31–33, https://www.gnu.org/philosophy/open-source-misses-the-point.html . This is called “semantic satiation”. See Wikipedia .     Michael Eager is principal consultant at Eager Consulting in Palo Alto, Calif. He has over four decades experience developing compilers, debuggers, and simulators for a wide range of processor architectures used in embedded systems. His current and former clients include major semiconductor companies and systems developers. Michael has been a member of the ISO C++ Standard Committee and ABI Committees for several processor architectures. He is chair of the Debugging Standards Committee for DWARF, a widely used debug data format. He is active in the open-source and Linux communities.

We usually hear the word “freedom” in relation to Free Software and Open Source Software (FOSS). Actually, most use of the word comes from the   Free Software Foundation (FSF) and its founder, Richard M. Stallman, repeating the term over and over , until the word is almost devoid of meaning . The four freedoms expressed by Stallman are the freedom to run a program as you wish, freedom to study or modify a program, freedom to redistribute a program, and freedom to distribute modified copies of a program.    Now, there might be reasons to debate whether Stallman's use of “freedom” has the same meaning as in “freedom of speech” or “freedom of religion” but that's a topic for a different article. When asked about “free software”, Stallman and others will throw out the quip “free as in speech, not free as in beer”. Apparently, someone, somewhere is giving away free beer. I want to find that person.   Anyone can download the source for the many thousands of FOSS programs, modify it, and redistribute a new version. But in practice, only a very small number of people ever download and build a FOSS program from source. I do on occasion, when a program isn't available in my distribution's repository. Of the few people who download the source of a program, a much smaller number will actually study it to see how it works and an even smaller number will make changes.    I use FOSS software. I use GNU tools, LibreOffice, Firefox, GIMP, and more. Like the great majority of Linux users, I use whichever version of a package is included in my Linux distribution. I'm very unlikely to ever modify any of these programs, even if I have the “freedom” to do so. The reasons are simple: Large programs like these are complex and take considerable time to comprehend how they work.   I recently had an online exchange with someone who was upset that his Linux distribution did not provide support for a peripheral he just bought. He said that he had to purchase a driver to get it to work and that this was the first time in many years that he had paid for any kind of software. He wasn't concerned about the freedom to modify the software, or the freedom to understand how it works, or freedom to redistribute it, both because he lacked the technical background and, more important, lacked the desire to do any of these. The price for the driver was very modest, but he thought that the Linux community should have software support for any hardware on the market available without cost.   In practice, the operative word in FOSS is not “freedom” but “free” as in without charge.   I use Adobe Acroread on Linux, which is free software. That is, free as in free ware, no charge but no source, and with a click-to-accept license. I use Chrome which is built out of a large number of FOSS packages, but which has a license agreement no less restrictive than that included with any proprietary program, prohibiting reverse engineering, creating a derivative work, or copying Chrome. (The rationale, as I understand it, is that Chrome is inextricably related to the proprietary services which Google provides, such as collecting your browsing habits and selling it to advertisers.) Chrome is now the most popular browser, with over 50% of the market, beating both FOSS browser Firefox and proprietary Internet Explorer. Like everyone else, I have dozens of apps installed on my Android smartphone. Almost all are free. That is, there is no charge to install them. Some apps are full function, others crippled or adware to encourage me to purchase a fully functional version. Almost all are closed source, although no one seems concerned.   While there is a small, impassioned, and very vocal community which is concerned about Free Software, the great majority are mostly interested in Free.   Free as in free beer, not free as in freedom.   Notes: The word “freedom” appears 43 times in Stallman's 3-page article “Why Open Source misses the point of Free Software,” Communications of the ACM, June, 2009, vol. 52(6). pp. 31–33, https://www.gnu.org/philosophy/open-source-misses-the-point.html . This is called “semantic satiation”. See Wikipedia .

In the computer world there are a few subjects sure to ignite a flame war. What’s the best editor? Where should one put the braces in C code? How many spaces should we indent in C? Agile or plan-driven?   One that seems to have tamped down, at least somewhat, is Microsoft sneering. You know what I mean: Computer pros are expected to expect little of value from that company. Linux is the gold standard, Windows the vulture pickings. Office is junk, and any knowledgeable person uses an open-source alternative.   And yet.   I have been running Windows 8.1 on some of the machines here for quite a while. It is big. It is complex. It has annoying quirks. My biggest complaint is when it tells me I don’t have permission to do something, even when running as administrator. Hey, who owns this machine? It’s annoying to have to reboot to complete an installation. It seems silly to have to edit the registry to change program properties.   But it has never crashed. Not once. Nor have any of the Microsoft Office 15 or 365 programs. (We run multiple versions of Office here on different computers).   Some of my Office documents are enormous and very complex with video and other demanding resources. It’s not unusual for me to create a 300 page Word file. And Word just works. It’s not quirky. Huge Excel files grind through their computations without fuss. Powerpoint has never crashed in any of the seminars I do. Not once.   Video processing used to be a nightmare. The programs crashed often. When they didn’t, you had to shut down all other applications to ensure they’d render the video and audio in sync. And even then it was a crapshoot. Today I use PowerDirector, which can consume every available CPU cycle on every core in the machine. But it plays nicely, even when working on gigabyte-long files and coexists with the plethora of other open programs. Unexpected stuff just doesn’t happen.   It wasn’t always thus. Once upon a time these applications, and the operating system, crashed frequently. It was expected. We learned to save files often in anticipation of a crash. But unexpected shutdowns happened so often that clicking on FILE:SAVE consumed too much time. We learned to left-hand control-s while still typing. For me it became a nervous twitch; I probably hit those keys after writing each sentence. I probably do it in my sleep. Today, though the applications don’t crash, the control-s impulse is so ingrained it’s still part of my every-few-seconds routine.   I wonder if younger people, those for whom PCs just work, have the habit? Do they know the paranoia of potentially losing a file from a crash?   We developed the control-s habit from bitter experience. Is that experience as irrelevant today as knowing how to develop code on a paper-tape machine is? Or programming a Nova minicomputer?   All of us over 50 suffered from many years of unreliable software from Redmond. On the other hand, since the advent of the PC we’ve gotten a ton of work done that would have been impossible or hugely expensive without those machines, the vast majority of which were running Windows.   Here we run Windows, OS-X, iOS, and Linux. Each has its place. At the risk of being flamed, I have to admit to being very pleased with my current PCs and their software.

In a Washington Post article, Google got it wrong. The open-office trend is destroying the workplace , the author explained how an ad agency moved her from a private office to an open space environment. She now sits at a table with 11 other writers. Every cough, sniffle, telephone call, and business or casual conversation interrupts a dozen workers. Open offices are hip, cool and modern. People lounge around on beanbag chairs. Bright colors and Warhol wall garnishes abound. The article states: “While employees feel like they’re part of a laid-back, innovative enterprise, the environment ultimately damages workers’ attention spans, productivity, creative thinking, and satisfaction. Furthermore, a sense of privacy boosts job performance, while the opposite can cause feelings of helplessness.” Engineering is an intensively creative enterprise. People need undistracted time to think, to focus, to mentally assemble a complex bit of code. The model falls apart after any interruption. In fact, interruptions are the biggest productivity killer for software engineers. Tom DeMarco and Tim Lister documented this well in their seminal Peopleware: Productive Projects and Teams . They found it takes, on average, 15 minutes to assemble that mental model. Yet, the average engineer is interrupted every 11 minutes (Mark, Gonzalez, Harris, 2005, "No Task Left Behind?: Examining the Nature of Fragmented Work." Proceedings of the SIGCHI Conference on Human Factors in Computer Systems). Peopleware showed that eliminating these interruptions gives almost a 300% boost in software engineering productivity. Let’s see: a 3X productivity improvement for engineers, some of the most costly workers in an enterprise. Or, we can save a few bucks and crowd everyone together. So what happens? First cubicles pushed workers into prison-like cells. Then inmates were added; two, three, then four to a cube. Now those annoying walls go and everyone is crammed together in one big, unhappy room. That worker’s pungent perfume or the inevitable result of last night’s bean soup fills the room. Chatter is non-stop. You really want to tune out the discussion of Joe’s looming divorce, but it’s human nature to be curious, to listen closely for the juiciest gossip. Does that sound productive?   (In the too-good-to-be-true department, Thesaurus.com lists “cubicle” as a synonym for “cell.”) Facebook is seating 2800 engineers in what is called the world’s largest open space . A single room houses ten acres of bodies, computers, interruptions, discussions, bells ringing, phones tweeting, and, one supposes, constant Facebook updating. Maybe a little work gets done, too. In The Moral Life of Cubicles the author states that in 2000 the average office worker had 250 square feet of space. That was down to 190 five years later. Facebook’s 10 acres works out to 150 square feet per person, assuming there are no WCs, break rooms, or hallways. Extrapolating, there will be zero square feet per person in under 20 years. A 2011 study ( Scandinavian Journal of Work, Environment Health ) showed that those sentenced to two-person cubicles have 50% more sick days than those in one-person offices. Open spaces increase that to 62%. The upside is that so many people are out sick that the noise level goes down, I suppose. Robert Propst invented the “action office,” the cubicle’s predecessor. Shortly before he died in 2000, he lashed out at cubes, calling them “monolithic insanity.” Do you work in a cube or open office? What’s your take on it?