Interviews: Jon "maddog" Hall Answers Your Questions 31
Last week you had a chance to ask Jon "maddog" Hall about his work on Project Caua and FOSS in general. Below you'll find his answers to those questions.
On Project Cauã by eldavojohn
First I praise you for your work and your goals -- they are refreshing compared to "please investors." But one of the keywords in your goals statements for Project Cauã is "capitalistic" as in "do all of this in a capitalistic, sustainable way, with little or no money coming from government." This mildly confuses me. I don't see FOSS as directly contradictory to capitalism but your goal of "triple or quadruple the number of FOSS developers in the world" seems, well, a little more public domain oriented than private industry, ownership and other tenants of capitalism. To put my question bluntly, why even pay petty lip service to capitalism when your goals of reducing electronic landfills, free-of-charge wireless and increasing user security are just not monetarily rewarded by the free market? These goals are about empowering people and protecting our future environment, how precisely does that align with capitalism? I understand how your job creation might benefit the economy but I don't understand how you're going to actually create these jobs. What companies are you talking to that have positions for these jobs? Most countries can't even pay to create jobs -- I'm sure several leaders would gladly put down billions of dollars if it meant magically creating productive and sustainable jobs, what is Project Cauã doing differently?
mad dog: First of all, my comment about “capitalistic” is aimed broadly to differentiate this from a government-sponsored project. Unlike a lot of people today, I do think that some governments do good things, but I also recognize that budgets shift and projects are unfunded or underfunded under governments. I have seen many projects started by government, only to fail over time.
My observation is that when you create a situation where people can make a good living, people do tend to fill the need. Project Caua aims to create millions of one-person businesses that supply computing services to end users. If each of these people creating these businesses were to do all the research of setting up the business, making the business arrangements, dealing with the government, there would be billions of person-days lost, so Project Caua is attempting to put together business plans for each of these people to do have their business with the least possible cost and effort, and we want to do this using as little government money as possible, instead relying on capitalistic means to attain a sustainable model.
This is not to say that we will not go after and accept government grants. We have tried doing this several times, only to go 99.99% of the way through the process to find a blocking point. While several government organizations have looked at the plan and found it “interesting”, many recognize the plan is complex, and depends on a fundamental principles that “Geeks can sell things and geeks can run their own companies”.
So we are creating a series of pilots to prove these fundamental principles and then we hope that some grants will start occurring. These grants, however, are not fundamental to keeping Project Caua going, they are just “accelerators” of various parts of Project Caua. Therefore Project Caua itself will be “self-sustaining in a capitalistic way”.
As to the goals of “empowering people and protecting our future environment”, we believe that these are a natural part of our proposition to customers, not a side effect that we have to create.
So in summary, “capitalism” really means setting up an environment where a single-person business can be sustained providing the person with a good living at the same time sustaining an infrastructure to provide that support without relying on public-sector funding.
Thin client server
by countach
My question is, why do we need thin client/server in an age where a decent computer surely costs about the same as some kind of thin client anyway? I can see benefits in a specialized scenario where you need access to vast computing power, but for every day people in an apartment complex, web browsing and reading mail, why is it necessary, and doesn't it in fact add a lot of complexity for little gain, not to mention administrative problems and a central point of failure. People are using $30 Android tablets for their computing needs without that complexity.
maddog: The term “thin client” in this case simply means a computer system that has no local storage. The storage, including both data and programs, is kept in a central place where it is easy for a systems administrator to do backup, filter spam, fight viruses, etc. When you put local storage into a client, then you have to worry about retrieving that data when the client breaks.
The issue of “central point of failure” is covered by having properly designed redundant servers, which spread the load around any single point of failure.
People using $30 Android tablets for their computing needs belies the fact that they either use cloud services to store their data, or they back up their data themselves, either to a local server or a cloud service. Most tablets typically do little or no “computing”, but is more like a display server. Also, while I have watched movies on a 7” tablet, I would rather watch them on a much larger screen.
The thin client that we are currently using is the Raspberry Pi, which typically lists for 35 USD, and attaches to an existing TV or LCD panel through HDMI.
As to the added complexity, I have know thin client installations where 4000 logins were supported on 2200 thin clients, using 63 servers and only four system administrators. While it is true that one thin client is more complex to administrate than one tablet, when you have ten or fifteen tablets, laptops, desktops and other computing units in your environment it is harder to keep them all synchronized and up-to-date, filter spam and get rid of viruses.
We believe that a properly configured LTSP system could support 300-600 thin clients on a highly available server utilizing about ten hours a week for the administrator, yet generate a “livable wage”. This would leave that person 30 hours a week to do other tasks that would bring in more money, and Project Caua will help to facilitate those tasks too. So we anticipate that the Project Caua Systems Administrator/Entrepreneur will be able to make a very good living.
What is your opinion on GPLv3?
by Anonymous Coward
What is your opinion on GPLv3?
maddog: I think it is another Free Software license, and if it meets the needs of the copyright holder, then great!
I would have preferred that the GPLv2.x license also be changed to fix some of the issues with wording, etc. without adding the other provisions that created GPLv3, but other people decided not to do that.
Making money off FOSS
by unixisc
Traditionally, the 3 ways of making money off FOSS have been:
1. Selling hardware
2. Selling support services
3. Donations
Long term, do you see any other ways in which one can make money on FOSS?
maddog: I realized some time ago that people do not really want hardware, software and service. They want a solution, and they buy hardware, software and service to obtain that solution. Therefore if a person can identify a high-value solution for a customer they can try to provide that solution using closed-source, proprietary software, and perhaps very costly closed hardware, or they can chose “commodity” hardware and FOSS. Solving a problem that could cost two million dollars a year with hardware and software that is both open, flexible and can often solve the problem faster and better is what can win the lucrative solution contract.
The transition to SoC-based “content devices
by Adekyn
The “Apple iPhone” and “Amazon Kindle” were release in 2007; and the “Apple iPad” followed just 3 years later in 2010. Now, in mid-2013, the combination of smart-phone and tablet devices has eroded the PC market - with projections of tablets out-selling PCs by 1 million units by 2017. It has been estimated that, presently ~70% of these devices are running Linux (in the form of Android) and soon, Canonical will be throwing Ubuntu/Unity into the mix. Ironically, while it is fantastic that Linux has been to be proliferated to the masses, it has done so in a very “closed” way. These are marketed as self-contained content devices _not computers_. To develop software for these products, one (for the most part) cannot simply code with tools/languages of your choosing – you have to conform to the tools and delivery methodologies of the device manufactures. How do you see this trend of abandoning Personal Computers for SoC-based content devices affecting the future development of Linux or, for that matter, the future of programming in general?
maddog: I do not believe that the personal computer will be “abandoned”. As long as we cannot interface with machines through a direct mind-link, I believe there will still be a wide range of devices from cell phones (and things smaller than cell phones) to tablets, laptops, desktops, servers and supercomputers. The mix will change, but the devices and interfaces will still exist.
There are about 2 billion “desktop/laptop” systems in the world, and 7.3 billion people, so there is plenty of room for expansion of all types.
There is also a drive towards more openness, both from the FOSS community and from other communities. The recent issues over longevity of solutions and spyware will be driving a market for FOSS firmware (coreboot is a good example of this) as well as eliminating lots of current day firmware blobs.
There is a huge market for truly open devices, and once market drivers realize this, someone will step in to fill it. One of the goals of Project Caua is to develop a completely “Open” system.
Thoughts on Alpha?
by idunham
I'm wondering what you think looking back at the whole Alpha scene.
-were there any major failings?
-what were the nicest features?
-while the hardware is now abandoned and slow, do you think it could have remained competitive?
-favorite story related to Alpha or Linux/Alpha?
-are you still interested in Alpha, or have you moved on?
maddog: Well, there was obviously a major failing, since the Alpha Architecture is no more. In its day the Alpha maintained its reputation as the “World's Fastest Microprocessor” in the Guinness Book of World Records for a number of years. It was also one of the early 64-bit microprocessors, and had a huge data bus for allowing atomic operations over large data structures. It was a RISC system, and from that I gained my real appreciation for RISC architecture and what it can do. My admiration for CISC dropped a lot, and I never thought the Itanium would ever amount to much. From the first time I heard about Itanium, I kept saying to whoever would listen “haven't we tried this enough”?
I was amazed by some of the things the later Alphas could do, particularly out-of-order instruction execution.
Your comment about how the 'hardware is now abandoned and slow” is interesting. I remember thinking how the PDP-8 (also a “RISC” processor in a lot of ways) was blindingly fast at 333,000 instructions per second. Now the hardware is “abandoned and slow”, perhaps because no one has done any research/engineering on it....
Would the Alpha be able to sustain its lead indefinitely? One of the issues is the cost of the fabrication plant and how much it cost to keep shrinking the dies needed to make faster Alphas, and how much the fabrication plant cost to keep it up to date. Digital was not selling enough Alphas to keep the plant at full production, so we also made StrongARMs and even (at one time) SPARC CPUs for Sun Microsystems, our greatest Unix competitor. I do not know if it would be practical for DEC to work as a “fabless” cpu producer of Alpha, like Sun, MIPS, and ARM are, relying on other companies FABs to make Alphas.
My favorite story (and stories) about Alpha is (of course) the Alpha/Linux port. How I met Linus, saw Linux for the first time, “convinced” Linus to port to Alpha (he had already been trying to get one from the DEC office in Helsinki), found the Alpha system in DEC for Linus and bullied it to end up at his house, then helped build the team and community to finish the port in nine months. And of course the “stories” never end, because they are still going on.....
Am I still interested in Alpha, or have I “moved on”? I have moved on from the Alpha the same way I have moved on from the PDP-8. Each contributed to my life, but they are only machines. I am glad that technologies from the Alpha live on in Intel, AMD and ARM architectures where engineers from the Digital semiconductor group all went when that branch was sold to Intel.
The same thing with Digital Unix. Many of the engineers that worked on that went to Red Hat and other Linux companies, and a lot of the technology lives on in GNU/Linux.
I have a couple of Alpha machines in my house, and have not turned them on for ten years. One is still new in the box.....
No UK VAT FOSS bookkeeping program
by eionmac
My question: While developing many other facets of Gnu/Linux so that most small companies can use FOSS to totally run their business, the lack of a FOSS UK or EU VAT system of bookkeeping keeps small start ups 'locked' into Windows as there they have the relevant accounting/bookkeeping programs. (GnuCash is OK for personal accounts, useless for UK or EU VAT systems). Is any effort being made to solve this as it would allow start ups to be independent of Windows. There are some paid Linux bookkeeping systems but that is off putting to start ups.
maddog: I agree this is a need. Sounds like a great project for you to start. There are a couple of issues here, and one of them is that the software itself does not seem to be the issue. From what I have been told, the issue is coding all of the data for the tax code into the software and having it certified.
New Hampshire License Plate
by Ian.Waring
Who had the "Live Free or Die - UNIX" license plate at Spit Brook first; you or Armando Stettner? And do you still have it??
maddog: Bill Shannon had the plate first. When he moved to California, Armando Stettner took it over. After Armando moved to California and relinquished it, I obtained it in 1988, and have had it ever since.
great quote (Score:5, Insightful)
I realized some time ago that people do not really want hardware, software and service. They want a solution, and they buy hardware, software and service to obtain that solution.
Important to remember that non-programmers don't get caught up in the specific religious wars. They don't care if it's big-endian, little-endian, Python, or Visual Basic; they just want their problem solved.
Re: (Score:3)
Indeed, and not just non-programmers. That's another beauty of open source; rather than having to write a custom module from scratch, find something that's almost and modify it to your needs.
Ten years ago or so when I was still programming, my boss' boss gave my boss a programming assignment. After working on it for almost a year without any success (wrong choice of language, I guess, she was more mainframe-oriented) she dumped it in my lap two days before the deadline. I hacked something together in time u
Re: (Score:2)
Re: (Score:2)
I just used HTML and javascript, she probably used SAS. Back then I was mostly programming in dBase (FoxPro and Clipper) on the PC and Nomad on the mainframe.
Back in the early eighties I wrote games for the TS-1000 using Z-80 assembly, hand assembled because there was no assembler.
Re: (Score:2)
They don't care if it's big-endian, little-endian, Python, or Visual Basic; they just want their problem solved.
Sure, same thing applies when people go to a mechanic, or a plumber, etc. They go to a mechanic and the mechanic tells them that their car has broken a belt and that it's probably happening because of some belt-driven component failing and it will take some work to fix. The person demands that the problem just be fixed and that the belt be replaced. Mechanic replaces belt. Car works. Smug car owner congratulates themselves on their intelligence. A month later, their water pump fails completely and they crac
Re: (Score:2)
A month later, their water pump fails completely and they crack their engine block, destroying the engine.
That isn't exactly solving the problem......
Re: (Score:2)
It's solving the problem that the customer demands be solved. It's the problem right in front of them, and they don't care about the details, so when the mechanic says: "We could just replace the belt but..." they don't listen to anything after the "but".
Re: (Score:2)
Re: (Score:2)
Well, it's a car analogy, so it's not an exact fit to the question at hand. Still, how about this: because their car breaks, the can't drive to work and get fired, and then their kids end up homeless and starving. Maybe I don't want that to happen. Also, there's such a thing as professional pride.
Re: (Score:2)
Re: (Score:2)
There are ways to solve this problem. One of them is to simply refer such people to someone else. If we step outside that car analogy, that's not as easy when the client is your nominal boss. Sure, you can look for another job, but that's fairly drastic. As far as I can tell, the main way to get around this is to lie. If you're asked if it can be done quick and dirty to get something that works, you simply say no.
Re: (Score:2)
Re:great quote (Score:4, Insightful)
"People scream: "Just make it work", and don't care about the details."
This is why I have been going to the same car mechanic for the past thirty years. They tell me what is wrong with the car, what will be the minimal fix, and why I should or should not get something else fixed. I know them and trust them, as I know a little about cars, but not enough to fix them, nor do I have the tools to fix them.
There have been times that they did work and the wrong thing was fixed or not fixed. "S**t Happens", and they did the best to fix the issue when I went back, sometimes at "no additional charge".
We expect that the Caua Entrepreneurs (as we call them) will have a set of customers that they know very well, and those customers know them. A bond of trust will be formed. If something goes wrong, eventually it will be fixed.
"That doesn't make the details unimportant."
The water pump about to break was not a "detail".....the make of the replacement belt is a detail.
Warmest regards,
maddog (a.k.a. "real_maddog")
Re: (Score:1)
The water pump about to break was not only a detail, it was a detail explained to the customer and then ignored by the customer. You seem to be implying that, because it was important, it wasn't a 'detail'. There is a reason why the phrase "unimportant detail" is not redundant.
Re: (Score:3, Insightful)
The water pump about to break was not only a detail, it was a detail explained to the customer and then ignored by the customer. You seem to be implying that, because it was important, it wasn't a 'detail'. There is a reason why the phrase "unimportant detail" is not redundant.
I am not only implying that "because it was important, it wasn't a detail", I am stating that fact very loudly. Things that are important are not details...they are important. The customer may ignore them, or not understand them, but if you to not tell them the important things, then you leave yourself open to their criticism later on. You are the expert.
I have known car mechanics to put their "important" opinions onto the work orders, and have the customer sign a waiver that they reject the additional w
Re: (Score:2)
If there's a real bond of trust between those doing the implementing and those they're implementing for such that explanations and warnings are not ignored, that's a different story. In my life I've run into too many management types who think that the important details aren't important. I've been made to listen to listen to recordings from business self-help gurus by some of these people who listen to that nonsense religiously and want to proselytize to others. It doesn't seem to occur to them that technic
Re: (Score:2)
Interesting to read his comments on the Alpha. (Score:1)
Re: (Score:2)
Re: (Score:1)
Re: (Score:1)
Far ahead? Not even close. Oh, I know what you are going to say "Android FTW!!111!!" but in reality, Android is one thing. Most FLOSS projects stagnate and die before they ever reach reasonable usability because the sexy code gets written then everyone leaves to find the next hot bit of code to write instead of cleaning up and improving what they were working on. Even the flagship FLOSS project, GNU/Linux, is based on 40 year old paradigms, uses a bolt-on GUI with arcane configuration and which sends too mu
Re: (Score:2)
And, here we have yet another example why FLOSS falls behind. People want software with specific features but they want it for free.
I was going to say he could probably pay xTuple to add the feature, but it appears somebody else already did [xtuple.com].